University of Uppsala

The field of political science traditionally includes both politics per se - actual politics - and man's perceptions of it. Within actual politics are found, inter alia, such phenomena as political institutions, structures, processes, occurrences, and behaviour. Human perceptions of these phenomena are usually described in words such as political opinions, ideologies, doctrines, programmes or opinions. In this context, legal codes, rules of political ethics, and similar norm systems, as well as the concepts, models, and theories of political science, must necessarily be placed within the general category of perceptions about actual politics.

1. The Subject Matter of the Study of Political Ideas

Perceptions about actual politics accordingly encompass individual concepts of the type 'government' and 'political system', cognitive convictions as to what is real, possible or necessary, as well as emotive convictions as to what is desirable, forbidden or commanded. What we have said, however, is complicated by the fact that phenomena on the level of perceptions or ideas do not necessarily refer to politics per se. Both concepts and convictions may refer to other perceptions or ideas.

At this point in our argument, it is necessary to distinguish between such concepts as are of a political nature and such as are judged to be scientific. In ordinary political texts authors not only often state their own perceptions of the configuration of actual politics, but also their opinions as to the views which their opponents have on the same actual politics. In certain contexts, it is furthermore not unlikely that the latter - wishing to enter into polemics - form an opinion of the manner in which the former perceive their own viewpoints. In this latter

* This is a revised version of a part of my paper 'Innehållslig och functionell idéanalysis' in
Statsvetenskaplig Tidskrift, 1974, pp. 1-20. I wish to express my gratitude to the Siamon
foundation which offered me financial support for my work.

case there occur four different levels: one for politics per se - the object level - and three for convictions about politics per se - the object-language, metalanguage, and metametalanguage levels. Within political language any number of further levels can also be imagined extending beyond this last level.

The expression 'political ideas' may be used as a collective designation for all these perceptions. It is broad enough, in contrast to, e.g., the expression 'political ideologies', to cover all forms of political ideas on all conceivable linguistic levels, from individual concepts to elaborate arguments, from emotively loaded statements to cognitive observations, and from simple descriptions of immediately observable phenomena to general theories of the nature of the state and the course of history. In this context 'political ideas' are not merely to be construed as concepts or more or less coherent systems of convictions, explicitly expressed by the person who embraces them; the expression also embraces implicit assumptions which can be deduced from the individual's verbalized convictions.

At this point it would be simplest to allow 'analysis' or 'study of political ideas' to designate the scholarly examination of 'political ideas'. This would, however, cause far too sharp a clash with accepted linguistic usage, which distinguishes between such political ideas as have been brought to light through interviews or have been accessible from public statistics, on the one hand, and, on the other, such as may be distilled from extant written documents. I shall follow this convention, using the term 'analysis of political ideas' for both such investigations of political ideas as are based on extant written sources and such scientific activities or processes as aim at producing such inquiries. As is the case with 'political science' and 'history', 'analysis of political ideas' will thus designate both a process and the product of that process.

The situation becomes somewhat more difficult to control when one leaves political perceptions and moves on to scientific perceptions about actual politics. These, as well, may of course occur at a number of different language levels. Neither substantive nor linguistic reasons, however, would justify allowing the expression 'study of political ideas' to embrace all these levels. It would seem necessary to establish certain demarcations.

When a researcher creates ideals for theoretical, methodological, or technical aspects of scientific investigation of actual politics, neither his activities nor their results should be considered as 'analysis of political ideas'. Nor is it reasonable to classify a penetration into actual politics with the primary purpose of improving the construction of theories about it as 'analysis of political ideas'. Between these two extremes are found those studies of scientific conceptions on various levels which I wish to designate as involving analysis of ideas. In order to differentiate them from investigations of political ideologies or political discussions and controversies, one might perhaps designate them as theory studies, but, as the methodology in both cases is the same, it seems more practical to place them both within the common category of 'analysis of political ideas'.

Particular emphasis should also be placed upon the fact that it is not possible
to draw a sharp line between the study of political ideas and the investigation of
politics per se. Even if actual politics is not included as a primary goal for the

analysis of political ideas, it must be considered to the extent that it is in some respects relevant for knowledge of political ideas and theories. As we shall see, this will occur in a number of different situations, inter alia, in testing the tenability of statements of fact.

2. Types of Analysis of Political Ideas

In a chapter concerning theory and method in my work, The Dissolution of the Union in 1905, I called attention to the fact that the study of political ideas may be pursued in two, fundamentally different, ways. One way may be called - for lack of a better name - 'content-oriented' or 'content-directed analysis of ideas'. The other way, in a similar fashion, may be designated 'function-oriented', 'function-directed', or perhaps merely 'functional'. Of course, various rewordings with 'content' and 'function' may and will also be employed.¹

In the content-oriented study of ideas, the object is treated either as theses
around which a discussion revolves or as conceivable direct or indirect arguments
for and against these theses. It starts from the premise that the chief task of

language should be to convey knowledge between human beings. It does not, however, ignore the fact that language also has the task of expressing emotions, influencing attitudes or inciting to action, but rather considers such contributions as having no place whatsoever in a disciplined discussion. The discussions and controversies which are studied are thus seen as an ordered conversation between people who wish to increase - or more fittingly arrange - their knowledge; the discussion is assumed to be a systematic attempt to elucidate a more or less abstruse subject.

It is assumed that the participants in the discussion espouse the norm of adhering to the actual point at issue, even if it should harm their own interests. It is also assumed they accept the rule that one ought to avoid such ambiguous and vague words as might run the risk of audience misunderstandings; one should keep accounts of the theses and arguments of opponents neutral with respect to other viewpoints - in particular, one's own; when giving an account of the viewpoints of an opponent, one should avoid omitting certain things and emphasizing others in such a manner as to give uninitiated individuals a biased or even an out-andout erroneous picture of the presentation of an opponent; one should not pursue implications of the actual utterances of an opponent without noting that the opponent would probably protest, or without stating the reasons one has for ascribing viewpoints to him which he himself disclaims; one should strive to establish a true comprehension of reality; and one should refrain from abusing the context hi a tendentious manner, e.g., through the use of insults and other simpler verbal reactions which cannot reasonably be interpreted as arguments, or by using music, serving food and drink, or by threatening violence.²

Analysis of the content of ideas may perhaps be best understood by comparing it to a disciplined conversation between an 'analyst of ideas' and an author, concerning a contribution the author has made to the discussion. The analyst of ideas enters the exchange of opinions as one who, in principle, is afforded equal rank among the debaters and who, therefore, is expected to adhere strictly to the regulations governing a disciplined exchange of opinions among human beings. Should he choose to study political philosophers from earlier times, he is expected to treat them as if they were contemporary thinkers well worth listening to. The contents of their teachings are to be considered as serious contributions within a disciplined discussion concerning a problem complex which still is relevant.³ He does not, however, go beyond the bounds of the argument and does not raise as an objection to the viewpoints of another debater that they gratify a certain psychological need he has, or that they are connected with his social status in society, the race he belongs to, or with his geographical surroundings.

Karl Popper's criticism of Plato, Hegel, and Marx, in The Open Society and its
Enemies, is a well-known example of such a content-oriented procedure in a
study of the political philosophy of previous generations.4

Function-oriented analysis of ideas, on the other hand, calls for an entirely different scientific endeavour. This type of inquiry gives rise to a series of questions which are not logically linked to the contents and validity of ideas. As the expression indicates, our attention is now directed towards the functions of ideas

in a larger context - whether psychological, sociological, or doctrinal. But functional analysis of ideas also includes the study of causal connections, statistical connections, and even weaker, non-logical connections between political ideas and the surroundings which evoke or receive them. In analysis of the function of ideas, one attempts to determine which groups embrace these ideas and how large these groups are, where one may find the origin of ideas, how their origin and development are to be explained, how they cause repercussions in the world around them.

It should particularly be noted that analysis of the function of ideas includes not only the effects that surroundings have upon ideas, upon which the sociology of knowledge and historical materialism have tended to concentrate their efforts, but also the opposite perspective, one less noted in the social sciences: the effects of ideas upon their surroundings. Analysis of the function of ideas may be defined as a systematic study of the origin, dispersion, and consequences of political ideas.

One way of dividing functional analysis of ideas into subdivisions is to do so with reference to what the political ideas that are studied are or should be related to. In this fashion, a number of different forms are obtained, which in many cases can be quite difficult to keep conceptually separated.

A method which commonly occurs among political scientists oriented towards the history of ideas is to relate the ideas being studied to other ideas. One is then pursuing genetic analysis of ideas, either by placing ideas into the larger tradition of ideas which they belong to, or by investigating how they were influenced by other ideas. Another common approach is the sociological, which implies that one attempts to establish relationships between ideas and a type of society or individual social phenomena, such as class, occupation, generation, sex, etc. A third consists of finding psychological connections between a personality type and ideas. The existential determinants mentioned here may well be the ones most commonly found in modern political science, but others can of course also be imagined. When the Swedish political philosopher Rudolf Kjellén wrote The State as an Organism, he was preoccupied with the influence of geographic factors upon the development of states and the conceptual world of man. Montesquieu's theory of climates is an example of a theory concerning the connection between climate and thought. The German Nazis, in their theory of races, placed great emphasis upon the socalled materialistic relations between thinking and biological structure.⁵

The various functional approaches are seen in Fig. 1. Naturally, not all possible
approaches have been shown in the diagram.

Furthermore, it should be noted that whatever may be assumed to be in one of the outer circles can of course be affected by the contents of other outer circles. Thus, nature, for example, can affect society, which in turn can affect personality. A complicated network of arrows might therefore be drawn connecting the outer circles of the diagram. The reason this has not been done is, of course, the fact that the investigation of these relationships cannot reasonably be considered analysis of political ideas. Analysis of ideas does not come into play until the personality in the above example is seen in connection with a structure of ideas.

3. The Places Allotted the Two Approaches within Political Science

Content-oriented and function-oriented analyses of ideas are traditionally found in political science, and in my opinion should continue to be so. A political idea, after all, fits into both an intellectual and a non-intellectual - e.g., social, psychological and historical - context. Political ideas play an important role when politicians are to choose, as rationally as possible, means and ends as a basis for their decision-making. But political ideas may also simultaneously serve as a cover-up for hidden private aims, class interests, or goals of party strategy. A political scientist thus finds it important to study ideas in both perspectives, that of a disciplined discussion and that of a socio-psychologico-historical context.

The discipline of philosophy faces an entirely different situation. Here, there can be no question that it is the content of the ideas that is at stake. The professional contribution of a philosopher consists of, e.g., a further logical development of Wittgenstein's, Carnap's or Popper's ideas, not the discovery of historical factors which explain their origin or dissemination.

The discipline of the history of ideas and sciences is in yet another position. There, it is surely not the content of ideas which primarily engages scholars, but rather the connections these ideas have with historical circumstances. The functional aspect thus tends to be dominant in this case.

However, within political science both methods of studying ideas exist side by side. Certain behaviouristically oriented scholars nevertheless wish to discard content-directed study, while still other researchers who find themselves tending towards the classical value-objectivistic direction firmly reject functional analysis. It would be enticing to examine these oppositions in greater detail, but that must remain beyond the bounds of this study. I shall accordingly assume that both analyses are legitimate - that of contents as well as that of function. This conclusion must, however, be qualified in at least two respects.

In the first place, these are in an analytical or logical sense two distinctly separate enterprises. An issue concerning contents can only be settled by means of a content-oriented analysis. In a similar manner, a question of function can only be answered by means of a functional examination. One can never answer a question of content by means of an analysis of function, and vice versa.

In the second place, both must be pursued in a specific manner if they are to
be worthy of the name 'science'. In sections 7-10 I shall attempt to show what
this means in greater detail.

4. A Metametalinguistic Distinction

This section concerning the meaning of the concepts 'content-oriented analysis
of ideas' and 'function-oriented analysis of ideas' will be concluded with an
examination of two borderline cases.

First, the conceptual distinction will be related to the discussion of the various
levels of language, as already presented in the introduction. We then achieve the
arrangement seen in Table I.

If the participants' actual act of voting is claimed to be the 'object', then 'object language' becomes the designation for the statements which the participants themselves make in the discussion being studied. 'Metalanguage' designates the terms in which the analyst of political ideas attempts to capture the object language of the participants, while the 'metametalanguage' is used for the terms, concepts, and sentences which are used to describe the metalanguage of the researchers.

In the light of this explication, it is obvious that the twin concepts 'analysis of the contents'/'analysis of the functions' lie on the metametalanguage level. The distinction is intended to simplify description and analysis of the language used by political scientists, historians of ideas and sciences, or sociologists of knowledge in pursuing their studies.

This clarification has a greater practical importance than many might be willing to believe; namely, that it now becomes obvious that the distinction between content-function on the level of metametalanguage may not be confused with an entirely different question on the level of object language, the problem of the nature of those ideas which compose the raw material for the various forms of scholarly studies of ideas. Various theories on the relationships between ideas and individual elements in their surroundings can of course be subjected to analysis of the content of ideas, although the theories in question may perhaps be statements about causal or statistical relationships between ideas and their surroundings. In that case, however, it is the ideas concerning the relationships that are primarily the object language and subjects of study; as long as these ideas concerning the relationships are not in turn brought into relation with their surroundings, the investigation is one of content.

An illustration will suffice to elucidate our argument. An examination of, for example, the beliefs of Nazi theoreticians concerning the relationship between belonging to the Jewish race and a certain way of thinking will of course not become functional merely because the raw material happens to treat the relationship between intellectual perceptions and their surroundings - in this case, socalled materialistic relationships. It will not become functional until the author of the study begins to be interested in how this Nazi theory of relationships in turn

is related to its surroundings. At this point, the analysis may be based on what

ever thoughts the Nazis may possibly have had on the subject as well as on hypotheses which the scholar himself formulates. The relationships between the Nazi theory of races and its surroundings may in turn be taken as a starting-point for a pure study of contents, which in turn can become functional if the theories concerning the relationships of racial theories to their surroundings are also related to their surroundings. One can, in principle, continue in like manner ad injinitum.

By reasoning as I do above, I wish to emphasize the fact that content-oriented analysis and function-oriented analysis are to be distinguished on a metametalinguistic level, created in order to furnish a language with which the various ways political researchers analyse ideas may be classified and discussed. The concepts have, in principle, nothing to do with the characteristics of the object language.

5. Genetic Analysis of Ideas — a Subtype of Functional Analysis of Ideas

The second borderline case which will be treated here concerns the question of where the so-called genetic analysis of ideas is to be placed. In The Dissolution of the Union in 1905, it is grouped under the functional, along with the psychological and social. In a review of the work, Professor Leif Lewin apparently wishes to question whether this step can be accepted: he seems to prefer placing it within the category of content-oriented analysis.⁶

In his subsequently published work on the study of the contents and functions of political ideologies, the reasoning of his review is expounded even more clearly. Lewin there claims that 'much of the interest in political ideologies has been historical, i.e., attempts have been made to demonstrate how ideological concepts may have received a new content and how the political belief systems have changed in meaning with the passing of time. The scholar wishes to inform a subsequent generation about this development of ideas and in a critical spirit demonstrate how a thought tradition or a view of society has changed.' This 'clarification of a thought tradition' means, according to Lewin, an application of 'the genetic method', which is seen as an approach within research concerning 'the content of ideologies'.⁷

From both the general wording and the example, it may be seen that Lewin to all appearances includes within the expression 'the genetic method' a task of a semantic nature. To establish the contents of concepts and the purport of belief systems at various points in time, i.e., to demonstrate how a tradition of ideas changes - without taking up the question of why the tradition changes - is, after all, according to current philosophic terminology, a 'semantic' activity. The scholar is then merely investigating the relationship between terms and concepts or sentences and their meaning. Lewin is of course correct in that this is an activity concerning contents. But to furthermore designate it as 'genetic' is from a linguistic point of view misleading. The term 'genetic' is primarily associated with 'genesis' or 'origin', and the expression 'the genetic method' leads one to

think of an activity which aims at establishing the origin, source or genesis of a phenomenon. Without special qualifications, it should not be used to designate an activity which confines itself to examining the relationship between terms and their conceptual content, even if this is 'historical' in the simple sense that any changes of meaning that may occur are followed during a long period of time. To use the term 'history of ideas' is also slightly misleading, since this particular term usually designates the relativization of ideas in some respect, e.g., with respect to the historical situation or the social structure.

I therefore maintain that the analytical boundary between the approaches dealing with content and with function must be drawn in such a manner as to prohibit the 'genetic' approach from belonging to the former approach. A precise definition of the concept 'genetic analysis of ideas' is, however, quite difficult and requires in any case a relatively comprehensive treatment.⁸

It is thus manifest that 'genetic analysis of ideas' must be distinguished from 'semantic analysis of ideas', but it is also important to distinguish it from a comparative analysis of ideas. Even a routine simple classification of political ideas implies that the researcher places idea structures beside one another - comparing them with reference to logical construction and semantic contents - in order to find similarities and dissimilarities. The comparison may completely ignore the dimension of time and restrict itself to ideas which were expressed contemporaneously. But it may also consider the temporal aspect, in the very simplistic sense that ideas which have been expressed at quite different times may be compared with one another (e.g., Plato's convictions with those of Hegel and Marx). If the ideas are intended to be included in a serious discussion - and the comparison is a step towards grouping them according to theses and arguments - then this is evidently a question of the semantic stage in the analysis of the content of ideas, irrespective of whether the ideas are totally contemporaneous or whether they are expressed centuries apart. Even when one places two idea structures opposite one another hi order to determine whether the one belongs to the same tradition of thought as the other, one is performing a comparison on the semantic and logical level, so that both the action and the result must be designated as analysis of the content of ideas.

One does not change over to functional analysis of the genetic type until the comparison is intended to create a basis for claims, hypotheses, and theories that one school of thought has been exposed to influences originating from the other. This presupposes that the semantic analysis is completed, but implies that a further step be taken. We are now interested in non-logical relationships between ideas, e.g., of causal, functional (in the sense of functionalism), statistical, or even a weaker type. The approach used in such an historical analysis of influences will be described more closely here.

The first step consists of investigating whether the texts at hand satisfy what the Swedish political scholar Gunnar Heckscher once called the internal criteria of influence. One first looks for similarities in questions of terminology and expression. Now and then, the texts may perhaps agree word for word. It then becomes impossible to imagine that the one document would have been written in that manner unless its author had the other as a model or both authors had used

a common model. But the similarity of expression may occur in a more watereddown form. The same sequence of events or problem complex may be clothed in linguistic guises which only partially resemble one another. Of course, a demonstration of verbal similarities does not prove an influence. The same words and expressions may actually occur to two authors when they attempt to describe the same sequence of events or analyse the same problem complex. But the more complicated the relationships under consideration and the greater the similarities which can be demonstrated in the choice of language, the stronger the evidence of influence becomes. One thus reaches, in the limiting case, the unquestioned copy.

There may, however, exist reasons for suspecting influences, even where verbal similarities are absent. This might take the form of a common selection of material. Reality is so incredibly complicated that at any rate more notable agreements in questions of events, presentation of problems, and solutions suffice for a dawning suspicion that there may exist a genetic relationship. A third approach for establishing an influence, via internal criteria, is to demonstrate mistakes - occurring in the work one is exarnining - mistakes which cannot be explained in any other way than that the author had a particular, stipulated second text as a model. The model may contain an obscure reference which the copyist misunderstood in some characteristic fashion. The copyist may also have made an easily explained jump in the text of his model, a jump with disastrous consequences for the contents of the copy.⁹

In many cases, by applying these techniques, one can come a long way towards a conclusive proof of influences. In order to express an even more certain opinion concerning influences, one must nevertheless demand that the object examined satisfy external criteria of influence as well. One must be able to demonstrate the likelihood or give proof that the person whose ideas are the centre of investigation actually came into contact with the other individual in some fashion, be it through listening to his lectures, conversing with him, or reading his works.¹⁰

Such investigations of various sorts of non-logical relationships between political
ideas and other ideas I wish to designate as 'genetic analysis of ideas'.

6. Terminological Remarks

The two scholarly activities which I have here described in terms of contentoriented and function-oriented analysis of ideas are occasionally in the habit of turning up, with greater or lesser divergences, in other linguistic guises in the international social science literature. The following may serve as examples:

I shall not in this connection attempt to compare the concepts I have used and the concepts listed here. Suffice it to say that manifest similarities - as well as certain differences - are probably to be found in the majority of cases. Instead, I shall discuss the benefits and drawbacks of my terminology.

The designation 'functional analysis of ideas' is not, in my opinion, an entirely satisfactory solution. The positive factors are its relatively broad distribution and that it at least partially leads to quite reasonable associations relating to the subject. The negative factors are primarily the notorious ambivalence and vagueness of the expression. 'Analysis of the function of ideas' naturally implies an attempt to establish whether a certain structure of ideas fills a function as a balancing or stabilizing factor in a larger psychological, social or other context, e.g., for society as a whole. The expression also implies an attempt to establish functional relationships in the sense of covariations between ideas and their surroundings. In both cases, the expression is quite apt. But analysis of ideas not primarily directed towards the contents of perceptions also implies a search for relationships which are stronger than those named above - i.e., causal connections, and relationships of a non-logical nature, which are weaker than those named above. For both these types of relationship, the term 'functional' is not really apt. All this suggests the use of another word as a principal term, letting the expressions causal and functional analyses of ideas designate something subordinate.¹¹

I have also considered the designation 'relational' analysis. In the first place, this expression causes one to think of words which are firmly established in scientific and technical prose, viz., 'relations' and 'to relate', and in relational analysis of ideas, one relates ideas to their surroundings, one investigates their relationships with their surroundings. In the second place, such terms as 'related to', relationship' and 'relations' are very often used when describing what, e.g., the sociology of knowledge is concerned with. Thus, Werner Stark contrasts 'the content of thought' with 'its relationship to the wider setting within which it has been conceived' ¹² Ami Merton emphasizes that the sociology of knowledge 'is primarily concerned with the relations between knowledge and other existential factors.¹³

In the third place, the term 'relational' has also been adopted into the scientific literature, even if it hardly can be considered as generally accepted usage. In Ideology and Utopia, Karl Mannheim makes a distinction which was much discussed and which was denoted by the concept-pair 'relativism-relationism'. The approach used by the sociology of knowledge is designated by Mannheim as relational' and, if we ignore certain epistemological consequences, he apparently means in principle approximately that which I intend by 'relational'.¹⁴ From a linguistic point of view, 'relational' refers to broader relationships in general and not, as 'functional', primarily to only a relatively restricted group of relationships. But this wider range of applications at the same time implies a decisive disadvantage. Putting a structure of ideas into a relationship with its surrounding does not always necessarily imply that one is occupied with that which I wish to have contained in 'relational analysis of ideas'; after all, bringing statements of fact into relation with reality in order to determine their truth value is, e.g., a very important task for content-oriented analysis.

Furthermore, philosophers speak of logical relationships between statements, but the very investigation of such relationships is itself a content-oriented task. The term 'relational' thus seems to have far too wide a scope and the term 'functional' one far too narrow. As the latter has nevertheless become more accepted, I have in this connection preferred it, although not without a certain hesitation.

But political research does not lack a firmly established language convention merely when designating its subordinate units, content-oriented and functionoriented analyses of ideas. Oddly enough, no appropriate collective designation has existed for the entire research area at issue here. Because it has distinct associations with the historization of ideas, i.e., with a judgment of ideas based on their historical surroundings, the traditional term 'history of political ideas' is clearly inappropriate as a comprehensive designation of the study of contents and functions of political ideas. Analysis of ideas from the point of view of their contents, after all, presupposes that ideas be understood as quantities worth serious consideration for the sake of their contents, irrespective of their origin in or effects upon the historic environment which surrounds them. The history of ideas', in a narrow sense, does not designate such a programmatically ahistorical procedure. The expression is thus used for something which strongly hints of the analysis of the function of ideas, and in fact is a very particular form of such analysis, namely, one which amounts to finding distinctive but non-general and non-immutable relationships between ideas and their surroundings. It may perhaps have been reasonable to speak of 'historians of ideas', when an older, historically-oriented school dominated the analyses of political ideas; but it becomes misleading when representatives for the newer, content-oriented and non-chronological investigations of scientific theories, political ideologies and political debates are meant.

Expressions containing the word 'ideology', e.g., the Continental metaphor 'ideology critique' or 'criticism', or 'the analysis of political ideologies', suffer from the limitation that the subject of analysis or criticism must be 'ideologies'. In my opinion, however, even normal contributions to discussions, as well as scientific theories and conceptual frameworks, should be within the province of analysis of ideas. The same technique is used, irrespective of whether one studies an ideology, various contributions to a debate, or a scientific general theory. The term 'ideology critique' is furthermore intimately connected to various Marxist ideological doctrines. When Marxists conduct 'ideology critique', they demonstrate that the bourgeois consciousness is false, but that it fills the function of justifying and preserving the existing capitalistic mode of production and social order; their own doctrines are, on the other hand, assumed to be scientific and above all ideological suspicion.₁₅

Finally, a decision must be made concerning the sociologists' term 'the sociology of knowledge'. It is completely unsuitable for the patently obvious reason that it roughly refers to analysis of the function of ideas, or perhaps more precisely, merely to the sociological analysis of ideas. Content-oriented analysis of ideas is automatically excluded, as well as such functional efforts as the genetic and psychological,

not to mention others of less interest for political scientists. Furthermore, the knowledge aspect of the sociology of knowledge is often defined more broadly, since it includes not only political ideas, but also social and economic notions, art, music - in fact, all forms of thought and all manifestations of culture whatsoever.

After thus considering matters, I finally decided, in The Dissolution of the Union in 1905, to use the designation 'analysis of political ideas'. It is broad enough to cover all conceivable variants of idea analysis, both the contentoriented and the function-oriented, as well as the older historical and the newer systematic analysis. It also covers both analysis of discussions, ideologies, scientific theories, and concepts.

7. Content-Oriented Analysis of Ideas: Semantic and Logical Methods

Within the bounds of an approach via contents, it is clear that, against this background,
ideas may in every case be treated with four different methods:

1) semantic method
2) logical method

3) analysis of the tenability of statements of fact
4) analysis of the tenability of statements of value.

It is necessary to apply semantic techniques. Before progressing to a logical or empirical testing of political ideas, their meaning must be established. The semantic question of the meaning of words must, then, be distinguished from the logical question of consistency and the empirical question of truth. The logical consistency or empirical validity of a sentence cannot be examined until one is satisfied as to the meaning of that sentence.

In this respect, my opinion may perhaps be at variance with the viewpoint which is developed in the finest Swedish contribution to the discussion of principles in the analysis of political ideas since the contributions of Tingsten and Heckscher at the end of the 1930s and beginning of the 19405: Leif Lewin's study, published last year, called 'On the study of the contents and function of political ideologies.' To wit, Lewin does not in his work perceive any particular semantic method. He hardly discusses the semantic problem at all.¹⁶ In my opinion, the first step in analysis of the contents of ideas must be an investigation of the structure of argument. We must ascertain what the issues being argued are and what the arguments for and against these issues or theses are. As political language in general constitutes an incredibly vague and ambiguous tool, the issueexpressions and arguments must be interpreted and made more precise. In Communication and Argument, Arne Naess has developed a theory for interpreting words and making them more precise which may be useful as a guideline for this activity. Naess's main concepts are 'interpretation', 'reasonable interpretation', and 'precization'. His concept structure is thus based - for very good reasons -

on the conviction that it is not possible to arrive at a uniquely correct interpretation of the material which the analyst of political ideas is faced with. The next step in the semantic analysis is to more or less systematically order the discussion into arguments for and against each thesis in question. Within the group of arguments pro and con, the reasons must be treated in some sort of succession, which presupposes that they are arranged according to some principle of order.¹⁷

Problems of interpretation and precization are not, however, brought into the picture solely when a scholar wishes to describe linguistically formulated conceptions of certain discussants. They arise at all conceivable linguistic levels. Whenever the scholar wishes to make his own conceptual framework as clear, precise and rich in nuances as is necessary for the purposes of his investigation, he is forced to grapple with semantic problems. This holds true whether the conceptual frame is to be used to investigate politics per se or beliefs about politics per se. It holds even for someone who, like the present author, is attempting to construct conceptual tools in order to be capable of describing and dissecting the analytical language of scholars. The same problem complex is found on a much lower linguistic level when the problem is posed as to whether the interpretations which the participants themselves have given of the standpoints and arguments of other participants, or of constitutions, conventions and the like are reasonable or whether they are tendentiously distorted. Semantic analysis is thus not merely reserved for those who dedicate themselves to so-called debate or theory studies. Elements of semantic analysis of ideas are part of all serious scholarly endeavours.

The empirical-semantic step within the particular form of analysis of contents of ideas which is focused on political controversies can nevertheless at first glance seem quite uncomplicated. The exchange of opinions we have agreed to study already exists, and all that remains for the scholar to do is, as accurately as possible, to render the actual statements of the participants. This is, however, a fundamental mistake which I think every scholar with practical experience of such studies of controversies must vigourously disclaim.

The researcher is often faced with difficult interpretational tasks. He is invariably confronted with the problem of bringing scattered and relatively independent statements together into a whole. In many cases, the actors lack that comprehensive view which is a prerequisite for determining to what extent their arguments agree or clash with those of others. The scholar is forced to construct a debate in which the actors themselves are perhaps only in a very vague way aware that they are discussing a problem concerning which many others also have viewpoints. In order to solve this problem, it is necessary to have a conceptual framework which enables one to trace interesting theses and arguments, as well as order them into groups. It is naive to believe that this theoretical search pattern is to be found ready-made in the raw material; it must be conceived and constructed by the describer. For a scholar to claim that it is possible to pursue an analysis of the contents of a political controversy without having his own theoretical concepts and ordering terms would be to take a completely unreasonable standpoint.

Once the empirical-semantic clean-up work has been accomplished according to certain theoretical principles, content-oriented analysis of ideas may be advanced with the help of the arsenal of methods at the disposal of philosophical semantics. It now becomes a question of accounting for those interpretations, precizations, and characterizations of issue-expressions and arguments which may be of importance for further scrutiny. It may at times be necessary to divide the arguments, according to theoretical meaning, into atheoretical and theoretical statements. The latter can in turn be divided up into analytic and synthetic. It is occasionally useful to classify statements according to their semantic function, which implies sorting them into informative, prescriptive, expressive, performative, and interrogative statements. These various operations at the philosophicsemantic level are performed in order to be capable of deciding which rules are to be used when determining the validity of the statements.

When the labour of interpreting, specifying, structuring, and describing the material has been completed, we may consider the semantic stage of the contentdirected analysis as having been completed. It would surely be possible to hunt up examples of studies which were stopped at that stage, but they normally progress to an actual testing of the intellectual force of the argument. One component of this testing consists of logical analysis, in which the logical construction of the sentences and their mutual logical relationships are at the centre of interest. We attempt to decide how well an argument of a participant, including the implications which may be deduced from it, fulfils the demand of logical consistency. We attempt to show that statements are contrary or contradictory, that there are inconsistencies concerning the valuation of various circumstances, or incompatibilities between recommendations.

An examination of the relevance of the argument for the thesis in question may also be undertaken in connection with the logical method. A relevant argument should augment the correctness or desirability of the thesis, and it must be decided whether the reasons here presented in any way speak for or against the issue being argued.¹⁸

8. Content-Oriented Analysis of Ideas: Examining the Tenability of Statements of Fact

After this, we may progress to a third step in the content-directed study, the tenability test, where an attempt is made to decide whether the assertions concerning reality, as expressed in various theoretical sentences, are true or false, or probably true or false, on the basis of certain established criteria of verification or falsification. The tenability test concerns so-called synthetic statements and implies that we establish either that these statements are actually true or false, or that it is more or less probable that they are true or false. For the sake of clarity, it should be noted that, in my opinion, a statement is true if and only if it claims that certain circumstances exist and it can be shown that these circumstances actually do exist.¹⁹ At this point the analysis of political ideas must therefore in

a quite decided way take actual politics into consideration. It progresses to what is,
somewhat improperly, usually called 'empirical research'.

There exist no insurmountable hindrances to establishing the truth value of a statement of fact, as long as it deals with relatively concrete and specific descriptions. The claim that the Liberal leader, Karl Staff, according to the Official Records of the Lower House of the Swedish Parliament, on 6 February 1914, made a contribution to the defence debate, may easily be falsified by examining the Official Records for that date. But descriptive political texts also teem with statements richer in contents and hence more difficult to prove. One often deals with far-reaching generalizations of simpler assertions about reality, as well as explanations, on various levels of abstraction, of simpler observations of reality.²⁰ The scholar is faced with a series of methodological and practical difficulties. In order to carry out an empirical test, the statements and concepts must be operationalized, whereby difficult problems of validity can of course occur. Even if these barriers should be dismantled in a somewhat acceptable fashion, there still often remains a quite comprehensive task, that of gathering and reworking the material so that one may be able to express some relatively assertive answers.

In many instances a notion, which in principle is of such a nature that it may be examined for tenability, cannot be expressed in statements which may be directly tested. This applies to statements as to how events in the past would have developed, given this and that presupposition. 'If the Western powers had not bowed to Hitler's Germany at Munich in 1938, the Second World War would have been avoided' is an example of such a statement that may well be discussed, but which cannot be tested in the true meaning of the word. Examples of other statements whose tenability cannot be directly examined are predictions about the flow of events in the future. One may choose the strategy of investigating whether the premises upon which the prognoses are based agree with the available facts, and in this manner indirectly test the tenability of the predictions.

Fortunately, it is not always so difficult to pursue criticism of tenability. Many tunes, one can make considerable progress by showing that participants have advanced groundless or incomplete or insufficiently grounded assertions about reality. It is accordingly not necessary to achieve a complete verification or falsification of the assertions about reality; it is sufficient to bring their tenability into question.

9. Content-Oriented Analysis: Testing the Validity of Statements of Value

Because man's conceptions of politics consist in great part of assumptions about the actual nature of reality, testing the tenability of cognitive convictions must be given a leading position in analysis of the content of ideas. To a certain extent, however, the political notions of man also consist of convictions which in one way or another are an answer to the question as to what constitutes a good and just

government. The tenability of such statements of value cannot be established in the same manner as for cognitive convictions. This forces the analyst of ideas to bring up one of the most discussed problems in modern science: how can one, using scientific methods, decide the question of the validity of statements of value? This conundrum, which may be called the fourth methodological stage in an analysis of the content of political ideas, has in recent years attracted the interest of, among others, Swedish political scientists. Their endeavours have, however, had to meet criticism of, e.g., their confusion of various meta-ethical theories through unclear use of such key terms as 'value relativism'.²¹

A renewed review of the demands which must be satisfied by any argument concerning questions of value which claims to be scientifically binding may therefore be appropriate for this account of content-oriented methods. It is of the greatest importance that we differentiate between valuations in the scholarly activity and valuations in the results of this scholarly activity. To the former category belong questions concerning the effects the valuations have upon the scholar's choice of:

(a) research problems

(b) concepts, hypotheses and theories
(c) methods

(d) ways of presenting his results
(e) ways of distributing his results.

It is assuredly reasonable to claim that all these decisions must be made through choices, and that these choices must be guided by valuations of what interesting problems, good methods, elegant publications or effective publishers consist of. According to a method which Gunnar Myrdal recommends, the scholar should strive to become conscious of the valuations which govern the selection. This is, however, a research-psychology or research-sociology (sociology of knowledge) aspect of the valuation problem which, while it may be quite interesting, will be disregarded here.²² The claim is also occasionally raised that the scholar should consciously investigate which interests or valuations the findings of political science may further, as well as see to it that at any rate his own publications do not help the wrong parties. This question, as well, will be left untreated. The problem of the validity of valuations, not the problem of their origins or consequences, is to be treated here.²³

In this connection, 'valuation', 'value', or 'value judgment' mean the thought
behind sentences of the type 'this is good', 'this is pretty', 'this is correct', 'this
is a duty', as well as sentences of obligation.24

Before beginning a discussion of the various theories of value, it will be helpful to distinguish between ontological theories of value and theories of value sentences. An ontological claim is a statement concerning what objects, characteristics, and relations exist and do not exist. The ontological theories of value contain state

ments concerning whether value characteristics actually exist in reality or not. Theories of sentences of value are, on the other hand, theories concerning the interpretation and precization of value sentences in such a manner as to allow a clearer picture than in the original expression as to just what sort of thoughts are at issue.₂₅

In the light of this conceptual review, the classical meta-ethical value theories
may be interpreted as varying combinations of various sorts of theories of value
sentences and value ontologies. We obtain the following arrangement:26

The traditional objectivistic theories of value are thus considered as a combination of an objectivistic theory of value sentences and a positive value ontology.²⁷ A property which the group of objectivistic theories of value sentences has in common is that value sentences of the type 'the democratic form of government is good' are interpreted as assertions about reality, and the object in question - in this case, the democratic form of government - is felt to be equipped with a particular objective characteristic, namely 'goodness. Because these are assertions, they can be tested against reality and, in principle, be either true or false.²⁸ The positive objectivistic value ontology states that all value sentences are statements of the type noted above concerning objective characteristics, and that there furthermore exist, or at any rate in principle can exist, objects which, along with all their empirically testable natural qualities, also have a collection of value qualities which may be comprehended either through experience or through some form of immediate experience or intuition.²⁹

Value objectivism in this classic sense, however, has been subject to harsh criticism
and must be considered as largely abandoned.30 It has been shouldered aside

by two other groups of theories, which are usually called value-relativistic (subjectivistic)
and non-cognitivistic (emotive) theories.

The normal meta-ethical value-relativism has a subjectivistic theory of value sentences, which implies that sentences of the type 'the democratic form of government is good' are to be interpreted as if they contained the assertion that the person who utters the sentence experiences a certain feeling or adopts a certain attitude towards the object of which he is speaking. They are not to be interpreted as statements concerning the object - in this case, the democratic form of government - but rather concerning the subject or the person speaking. Hence the designation 'subjectivistic' theory of value sentences.³¹ Meta-ethical value relativism further implies a negative value ontology. It accordingly specifically denies the existence of objects in the real world which, alongside all their empirically testable qualities, would also be equipped with a collection of value qualities which could be reached by human beings, whether through the senses or directly through intuition.

Value relativism shares this final criterion with the third classical value theory, the non-cognitivistic theory of value - or as it is called in the Nordic countries, value nihilism. In contrast to value relativism, however, value nihilism has a noncognitive value theory, which implies that value sentences are not seen as expressing any assertions whatsoever about reality, not even about the person speaking. Value sentences are conceived as being consistently atheoretical, so that criteria of truth or falsehood have no relevance whatsoever for them. Value sentences directly and immediately express a feeling, an order, advice, or a recommendation. They are not to be interpreted as if they described how the person speaking is experiencing a feeling, giving an order, giving advice, or making a recommendation. Value sentences fill expressive or prescriptive functions, but not informative ones.³²

An extensive philosophical discussion has taken place concerning the advantages
of value nihilism over value relativism. In my opinion, there is considerable
evidence pointing towards the superiority of the value nihilism interpretation.33

What does this value nihilism standpoint, then, imply for the analysis of the contents of value judgments? If value sentences do not express any assertions whatsoever, but merely emotions, commands or advice, then it is obviously not a particularly meaningful scientific task to express value sentences. This conclusion must nevertheless be modified in a few very important respects.

One must distinguish between making assertions about valuations and making utterances in value questions. Nothing prevents science from making assertions about valuations, since such statements may be, in principle, verified or falsified. It is possible to establish which valuations actually exist in a certain connection at a certain given time. One can attempt to decide what importance people attribute to them. One can investigate why a certain person or group of persons embraces a certain valuation, what functions the retention of a certain valuation fiil for a personality or for the political system, or what causal consequences such a retention has. The above is valid for all forms of valuations, irrespective of, e.g., their level of abstraction and place in an ends-means chain.³⁴

Even in questions of value, however - and this must be strongly emphasized

- one can make utterances and still claim to be scientific. The slogan, used by the older value nihilists, that scientific ethics may only be a doctrine about ethics and not in ethics, must be resolutely abandoned in this, its extreme form. At this point, however, still another important distinction must be introduced, that between valuations as an end in themselves, and valuations as a means to reach an end.³⁵ If the end is established, there are in principle no restrictions to the possibilities science possesses for expressing itself about the validity of the means for reaching the postulated end. One can make semantic investigations of the sentences in which the means are expressed and suggest interpretations and precizations of them. One can investigate their logical consistencies, both in relation to each other and in relation to the end. One can establish how effective they are, by means of a test, using current empirical methods, as to whether they are necessary or sufficient conditions for achieving the end, or with what degree of probability they will lead to the end. In this manner, one may in principle refute or confirm the validity of the valuations as means. This is the mode of procedure adopted within the so-called normative political theory.³⁶

The situation is different when it is a question of how to give the validity of ultimate ends a scientific status. It is not the case that science is completely incapable of dealing with them. It may express its opinion as to their semantic clarity and their logical consistency. Among these proper objectives of science may also be numbered the task of finding out whether the ultimate ends can be realized at all, and in that case, what the degree of likelihood is for their fulfilment. Science may not, however, express itself concerning the correctness or validity of ultimate ends with the same basic certainty, as when it is dealing with assertions about reality - since things in reality do not have any value qualities - and one therefore in logical terms cannot deduce any ultimate value judgments from assertions concerning the objective, actual nature of reality. In this sense, it is impossible to decide upon the validity of ultimate value judgments, using scientific methods. Science cannot, in absolute terms, demonstrate which highest goal is best. Here, we find an area for which science must admit its limitation. Arnold Brecht refers to this when he speaks of a 'painful area of scientific impotence'.³⁷

This review of various value theories has been made in the terminology of philosophical meta-ethics proper. There exists, however, a considerable inceptive political science nomenclature in this area. As it exhibits a striking terminological resemblance to, but conceptual difference from, the established meta-ethical linguistic usage, it indeed has a share in the blame for the confusion of concepts which has arisen in this field. In his major work on the foundations of modern political theory, Arnold Brecht develops a viewpoint which he calls 'scientific value relativism' and which, slightly incompletely, is often shortened to 'value relativism'.. Some political scientists may perhaps have believed that scientific value relativism may out of hand be equated with meta-ethical value relativism. This, however, is absolutely not so. The relationship may be graphically indicated as follows:

The foundations of scientific value relativism consist of the fact that it is absolutely impossible to establish in a scientific manner any positive or negative value ontology. The doctrine deals with what science may constructively accomplish concerning questions of value, having once accepted this thesis. Since scientific value relativism in no respect claims that there do not in reality exist any things with objective value qualities, but is content with the considerably more modest ascertainment that the existence of objective value qualities cannot be scientifically proven, it implies in this respect no unambiguous rejection of the classical meta-ethical value objectivism, which an acceptance of the meta-ethical value relativism automatically means. The most important practical consequence of this doctrine, however, is that in scientific works no claims may be advanced for having achieved a scientifically valid positive value ontology.

The scientific value relativism in its practical consequences thus quite nearly approaches the more basic denial of the positive value ontology advanced by meta-ethical value relativism and meta-ethical non-cognitivistic theory of value.³⁸ The value sentence theory seems to play a subordinate role in Brecht's presentation, and it does not seem possible to discover any directly explicit position he has taken. It seems, however, obvious that he does not distinguish between noncognitive and subjectivistic value sentence theories. He is after all building his presentation, in all important respects, upon the writings of Max Weber, who, as is well known, formulated his viewpoint long before the non-cognitivistic value sentence theories had received any great dissemination. He rather seems to be leaning towards a subjectivistic interpretation. It is, however, of lesser importance to him; the core of his argument consists of showing the absurdity of scientifically proving that there exist objective values. What some political scientists have called 'scientific value relativism' — which has occasionally been shortened to 'value relativism' thus in at least some cases very clearly does not refer to a combination of a value-relativistic value sentence theory and a negative value ontology. It is thus not at all a question of meta-ethical value relativism, but is merely concerned

with - or, rather, if political scientists have correctly understood Brecht,
should be concerned with - a denial that it is scientifically possible to set up a
positive or a negative value ontology.

My adherence in this study to a non-cognitivistic theory of value does not therefore imply any distancing whatsoever from the scientific value relativism of Weber and Brecht. The former should instead be seen as complementing the latter doctrinal construction with an overtly negative value ontology and an overtly non-cognitive value sentence theory.

10. Functional Analysis of Ideas

Functional analysis of ideas can, as we have seen, be defined as a systematic study
of the origin, dissemination and consequences of political ideas. It gives rise to at
least three interesting problems.

One concerns the question of which ideas are to be 'functionalized', or related to their surroundings. In the present case, this may simply be solved so that it must be a question of 'political ideas': it becomes the task of the individual scholar to make a more specific demarcation from case to case, according to the various hypotheses or theories he wishes to examine.₃₉

The second problem concerns the question of what the political ideas should be related to. In other words, this is a question of determining what meaning such expressions as 'surroundings', 'existential basis of thought', or 'Seinsverbundenheit des Wissens' have had and should have. Here, as well, many answers have been given, depending upon which factor has been felt to have the greatest explanatory power. A survey of these answers is to be found in the argument and in Figure 1 above.

The third problem may well be the most central and at the same time the one which seems most difficult to master. After all, function-oriented idea analysis attempts to ascribe to a certain group or class (or whatever) a particular constellation of ideas, in the sense that the ideas in question 'belong to' the group or are 'especially characteristic of the group. Exactly what type of relationship exists, then, between the structure of ideas and the group?

This question, which usually is given the heading 'the problem of imputation', has also been answered in numerous different ways. In a noted chapter in his book on the sociology of knowledge, Social Theory and Social Structure, Robert Merton has drawn up a list of those terms which are normally used to indicate various distinguishing characteristics of the relationships. Merton lists the following within the category of words which are used to express causal or statistical relationships: 'determination, cause, correspondence, necessary condition..., functional interdependence and interaction.' Within the category of symbolic, organic, or meaningful relationships, he includes such expressions as 'consistency, harmony, coherence, unity, congruity, compatibility,. .., Struktur zusammenhang, structural identities, inner connection, stylistic analogies, logico-meaningful integration (and) identity of meaning.' Finally, he works with a third category for

terms which are so ambivalent that they cannot be placed in the other two categories:
'correspondence, reflection, bound up with, in close connection with,
etc.'40

The list, as is evident, is quite comprehensive, but it gives a good picture of the disorder and lack of both precision and a generally accepted terminology which even today characterize this area of research. Considerations of space do not allow any more detailed discussion of those theories of relationships which have been presented by various political philosophers and scholars. One thing, however, must be pointed out. If function-oriented analysis of ideas ever is to achieve more than scattered impressionistic observations, these theories must be systematically mapped out, made precise, modified, operationalized, and subjected to empirical tests.⁴¹

---

NOTES

1. E. Vedung, Unionsdebatten 1905. En jamforelse mellan argumenteringen i Sverige och Norge (Stockholm: Almqvist & Wiksell, 1971), pp. 34 ff.

2. In the main, this refers to the survey of principles of effective discussion which Arne Naess has presented in Communication and Argument. Elements of Applied Semantics (London: Allen & Unwin, 1966), pp. 121 ff.

3. This remark can be found, for instance, in the lecture which Klaus von Beyme delivered when he entered upon his duties as a professor of political science in Tubingen. What I have designated as content-oriented analysis of ideas is a near equivalent to von Beyme's 'Ansatz der politischen Philosophic', which, however, in my opinion is too strongly associated with such value objectivists as Leo Strauss and Eric Voegelin. See K. von Beyme, Politische Ideengeschichte. Probleme eines interdisziplinaren Forschungsbereiches (Tubingen: Mohr, 1969), pp. 13 ff.

4. K. Popper, The Open Society and Its Enemies (London: Routledge & K. P., 1966), I-11, sth rev. ed.

5. For a comprehensive survey of various 'extrinsic approaches', see W. Stark, The Sociology of Knowledge. An Essay in the Deeper Understanding of the History of Ideas (London: Routledge & K. P., 1958), pp. 215 ff. See even F. E. Hartung, 'Problems of the Sociology of Knowledge', Philosophy of Science, 19, 1952, p. 18.

6. L. Lewin, 'Review of Evert Vedung: Unionsdebatten 1905. En jamforelse mellan argumenteringect i Sverige och Norge', Statsvetenskaplig Tidskrift, 1972, p. 286. Cf. p. 297.

7. L. Lewin, 'Om studiet av de politiska ideologiernas innehåll och funktion', Statsvetenskaplig Tidskrift, 1972, pp. 460 and 466.

8. Cf. Vedung, op.cit., pp. 55 f.

9. As to the methods of ascertaining dependence, see R. Torstendahl, Historia som vetenskap (Stockholm: Natur och Kultur, 1966), pp. 98 ff.

10. Concerning internal and external criteria of influence, see G. Heckscher, 'Politisk idéhistoria', Statsvetenskaplig Tidskrift, 1937, p. 433.

11. The ambiguity of the terms 'function' and 'functional' has been substantiated by Ernest Nagel in The Structure of Science. Problems in the Logic of Scientific Explanation (London: Routledge & K. P., 1961), pp. 522 ff. It is also emphasized in E. Meehan, Contemporary Political Thought. A Critical Study (Homewood, 111.: Dorsey, 1967), pp. 112f.

12. Stark, op.cit., p. 213.

13. R. K. Merton, Social Theory and Social Structure (New York: Free Press, 1968), p. 510. See also K. H. Wolff, 'The Sociology of Knowledge: Emphasis on an Empirical Attitude', Philosophy of Science, 10, 1943, p. 107; L. Coser, 'Sociology of Knowledge', in D. L. Sills, (ed.) International Encyclopedia of the Social Sciences, 8, (New York: Macmillan & Free Press, 1968), p. 428.

---

14. K. Mannheim, Ideology and Utopia. An Introduction to the Sociology of Knowledge (London: Routledge & K. P., 1968), p. 253; Cf. E. Laszlo, Collectivism and Political Power. A Relational Analysis of Ideological Confllct (The Hague: Nijhoff, 1963), pp. 4 ff. and 17.

15. See, e.g., E. Lemberg, Ideologie und Gesellschaft. Eine Theorie der ideologischen Syslerne, ihrer Struktur und Funktion (Stuttgart: Kohlhammer, 1971), pp. 31 ff.

16. Lewin, Om studiet av de politiska ideologiernas innehåll och funktion, op.cit., pp. 456 ff.

17. For a general survey of the methods of empirical semantics, see Naess, op.cit.

18. Naess, op.cit., pp. 108 ff.

19. Thus, I am a supporter of the correspondence theory of truth, not the coherence theory, pragmatic theory, or some other theory of truth. See A. R. White, Truth (New York: Anchor, 1970). The definition of truth is taken from L. Bergstrom, Objektivitet. En undersokning av inneborden, mojligheten och onskvdrdheten av objektivitet i samhdllsvetenskapen (Stockholm: Prisma, 1972), p. 92.

20. For a survey, see Meehan, op.cit., pp. 8 ff.

21. J.-E. Lane, 'Leif Lewin, statskunskapen och demokratin', Statsvetenskaplig Tidskrift, 1972, pp. 329 ff.

22. G. Myrdal, Value in Social Theory. A Selection of Essays on Methodology (London: Routledge & K. P., 1958), pp. 50 ff., 134, 157 ff., 233. See also G. Myrdal, Asian Drama. An Inquiry into the Poverty of Nations (London: Allen Lane, 1968), pp. 38, 54, 155, 157 ff.; G. Myrdal, Objectivity in Social Research (London: Duckworth, 1970), pp. 55 ff., 64 ff., 70 ff.

23. G. Hermerén, Viirdering och objektivitet (Lund: Studentlitteratur, 1972), pp. 11 ff. and 183.

24. The psychological process, event or mental situation which is causing a person to express these sentences is not included in the concept of 'valuation', 'value', or 'value judgment'. See Hermerén, ibid., p. 14.

25. M. Moritz, Inledning i vdrdeteori (Lund: Studentlitteratur, 1968), pp. 8 ff. Cf. Hermerén, op.cit., pp. 184 ff.

26. The figure is taken over from Moritz, op.cit., p. 93.

27. By rights, value objectivism can be interpreted as a variety of theories of value sentences, theories of value ontologies, or combinations of both. See Moritz, op.cit., pp. 9 and 92 ff.

28. Moritz, op.cit., p. 11. Cf. Vedung, op.cit., p. 49.

29. Moritz, op.cit., pp. 17 and 40.

30. For a good summary of critical viewpoints, see Moritz, op.cit., pp. 18 ff., 32 ff., and 37 ff.

31. If the subject and the attitude åre defined in diverse ways, we get a great variety of different subjectivistic theories of value sentences. See Moritz, op.cit., pp. 45 ff.

32. Moritz, op.cit., pp. 79 ff., 87, and 90. See also I. Hedenius, Om rått och moral (Stockholm: Wahlstrom & Widstrand, 1963), pp. 16 ff. No non-cognitivistic theory of value sentences denies the faet that sentences such as 'this is right' in principle åre formulated in the same way as sentences expressing assertions. What is disavowed is that this formal similarity also would imply that value sentences would express theoretical assertions. See Hedenius, op.cit., p. 23.

33. See Hedenius, op.cit., pp. 22 ff. Cf. Moritz, op.cit., pp. 51 ff., 82 ff.

34. A. Brecht, Political Theory: The Foundations of Twentieth-Century Political Thought (Princeton: Princeton Univ. Press, 1959), pp. 121 ff. Cf. Hedenius, op.cit., pp. 17, 21, 32 f.

35. Cf. Brecht, op.cit., pp. 121 f.

36. Brecht, op.cit., pp. 121 f.

37. Brecht, op.cit., p. 131. See also 121 ff., 207 ff., 261 ff., and 417 ff. Cf. I. Hedenius, Liv och nytta (Stockholm: Bonnier, 1961), pp. 97 ff.

38. Cf. Brecht, op.cit., pp. 125, 131, 256 ff., 262 ff., 284 ff., and 300 footnote 7.

39. For an account of various approaches to this problem within the sociology of knowledge, see Merton, op.cit., pp. 515, 521 ff. and B. Eriksson, Om den sociologiska analysen av kunskap. Några steg i riktning mot ett kunskapssociologiskt paradigm (Stockholm: Almqvist & Wiksell, 1972).

40. Merton, op.cit., p. 515.

41. For some fresh suggestions, see Stark, op.cit., p. 246 ff; A. Child, The Problem of Imputation in the Sociology of Knowledge', Ethics, 51, 1941, pp. 200 f f., A. Child, 'The Problem of Imputation Resolved', Ethics, 54, 1944, pp. 96 ff.