Human Studies 7.'343-362 (1984). Martinus Nijhoff Publishers, Dordrecht. Printed in the Netherlands.

FUNCTIONAL METHOD AND PHENOMENOLOGY: THE VIEW OF NIKLAS LUHMANN

JOHN BEDNARZ Jr. Seton Hall University

Since its introduction as a method of social scientific investigation 1 functional analysis has been exposed to a variety of criticisms. Among these are the reproaches that its basic concept - the function - is vague and imprecise, that it involves value judgments or that it is insensitive to problems of social change. One of the most serious criticisms, however, comes from Kingsley Davis ~ who calls into question the very independence of the functional method itself. Of course this does not mean that, as Davis (1959, p. 379) recognizes, "...the work done under the functional label is poor and unscientific..." It merely means that the method employed to obtain these results is actually in no way (Davis, 1959, p. 379), "...distinct from other sociological modes of analysis." Whereas it may have been useful - Davis uses the word "strategic" - in the past to speak of a functional mode of analysis as separate and distinct from other modes of sociological analysis, Davis believes that any such talk is essentially meaningless today. The reason for this is essentially historical. In trying to establish itself as a science of equal rank with the other sciences sociology had to overcome what Davis refers to as "obstacles". On the one hand there was the encyclopedism and reductionism of Comte and Spencer and on the other hand the ethics and social reform of Le Play. The method introduced by Smith and Durkheim viewed itself as distinct from those employed by Comte, Spencer and Le Play precisely because it was not infected with their particular drawbacks. In any event, the gradual overcoming of the various obstacles to the establishment of a science of sociology was accompanied by the development of a unified sociological method, according to Davis. 3 To speak of a separate functional method, today, different from

344 the one which developed historically would be, in Davis' view, to return to the initial sociological situation. The functional method introduced by Smith and Durkheim and passed on to Parsons and Merton through Malinowski and Radcliffe-Brown is no longer confronted with the obstacles faced by its initiators. The obstacles have been overcome and the functional method has been incorporated into the unified causalscientific method. And only because of this incorporation can sociology claim to meet the requirements of science: the explanation and prediction of (social) phenomena. 4 Included in this view of the sociological method is the subordination of the concept of function to that of causal relation which is the source of serious methodological difficulties which Niklas Luhmann tries to overcome by revaluating this relation.

II By incorporating functional method within the unified scientific method sociologists were unavoidably led, says Luhmann, to view functions as the effects of performances (causes) which contributed to the maintenance of complexly structured unities (systems). When formulations like "contribution to the maintenance of a system", "solution of system problems" or "furthering the integration or adaption of a system" indicate simple causal relations, functional explanations assume the form of fixed relations between causes and effects. With this interpretation, however, comes the problem of accounting for - i.e. explaining - causes (the occurrence of system-maintaining performances) on the basis of their effects. If strict scientific m e t h o d o l o g y requires the explanation and prediction of empirical data through the determination of invariant relations between particular causes and particular effects than this is precisely what functional explanation cannot accomplish. "For," as Luhmann (1970, p. 9) recognizes, "ever since the causal relation received a univocal temporal sense (which it possessed neither for Greek nor medieval thinkers), effects o f any kind could no longer explain the occurrence of causes. ''s Of course, sociologists were acquainted with this serious problem and tried to remedy it by qualifying their effects with what Luhmann refers to as "an auxiliary causal construction" (eine kausale Hilfskonstruktion) wbAch would, in Luhmann's (1970, p. 11 ) words, "...expand the effects into an acceptable explanatory foundation. ''6 He mentiones three of the "classical" devices contrived by sociologists in order to overcome

345 their serious methodological problem. The first, which Luhmann attributes to Malinowski, was to refer functional explanations to "needs" which were equated with the causes (motives) of actions performed to satisfy these needs. This device, however, leads inevitably to the equation of imagined effect and the cause of its production. The result is a tautology which vitiates the explanatory power of the function. 7 But if one does not equate need and motive then problems arise of their separate empirical determination, the logical relation between them and the empirical verification of this relation. 8 In either of these cases the result is unsatisfactory. A second device employed by sociologists in order to overcome the problem of functional explanation was the famous theory of equilibrium which was - not accidentally - borrowed from biology where it proved to be very fruitful. In this case functional explanation was formulated in terms of systems which maintained themselves "in equilibrium" against a changing environment which produced "disturbances" within the system. The disturbances released or triggered built-in structures and processes (causes) which brought the system back into a state of equilibrium. 9 Here again the concept of function is interpreted as an effect, viz. of structures and processes of equilibrium. But while the theory of equilibrium may have worked quite well in biology where a clear-cut criterion of equilibrium - viz. death existed, no such criterion could be found for the social dimension. "A social system is not fixed according to type like an organism. An ass cannot become a snake even if such a development would be necessary for its survival. A social order however can experience profound structural changes without surrendering its identity and continued existence. ''1 In this sense the transferral of the model of equilibrium from biology to the social sciences remains "stuck in vague analogies and metaphors" as Luhmann says. This too militates against the validity of functional explanations. The third auxiliary construction Luhmann mentions is the concept of functional reciprocity? I This concept is unsatisfactory for the same reasons as in the cases of the theories of needs and equilibrium because it makes the same presuppositions: the equation of needs and causes and the possession by systems of mechanisms for the preservation of equilibrium. In each of these cases (needs, equilibrium, functional reciprocity) the additional qualification of functions in order to meet the standards of scientific methodology, Luhmann believes, have essentially failed. The reason for this is to be found, he thinks, in the above mentioned

346 subordination of the concept of function to that of causal relation and the consequent attempt to establish invariant relations between particular causes and particular effects. When this is done one not only places himself before the alternative of choosing between mechanical explanation via causes or teleological explanation via effects but he also makes a crucial ontological presupposition which shares with an important philosophical tradition the belief that the unity of any complex structure is substantial. 12 In other words, the permanence of a complexly structured unity (system) is to be found in its essence which was, "...defined as an object .... which contained in itself only the properties common to all objects belonging to a class... ''13 Any method which makes this presupposition proceeds by excluding whatever is variable, i.e. in our case, it seeks to establish constant (invariant) relations between particular causes (performances) and particular effects (functions). By fixing these invariant relations it hopes to establish laws which will meet the rigorous scientific standards of explanation and prediction. Luhmann wants to indicate that the historical functional method inherited this ontological presupposition from a philosophical tradition that interpreted permanence in terms of an underlying, unchanging substance. 14 Despite these criticisms - with which he fully agrees - Luhmann believes that functional method can, nonetheless, accomplish precisely what Davis said it cannot: establish itself as an independent method. But in order to do this it must, as it were, undo what sociological history has done, i.e. it must disincorporate itself from the unified scientific method. This, then, introduces a radical change not only in functional methodology but also in the concept of the function itself) s As long as functional analysis viewed its task as the establishment of causal laws the concept of function was interpreted as an effect to be brought about by system maintaining performances, i.e. the functional relation itself was an invariant one between a particular cause and a particular effect. Some of the serious problems involved with this method have been indicated above. Rather than trying to circumvent these problems by a further qualification of the concept of function as an effect Luhmann believes that the real drawback to this method is the very attempt to discover functional laws (invariant relations) between a particular cause and a particular effect. For him the task or goal of functional method is not the discovery of, "...an invariant or more or less probable relation between particular causes and particular effects.... rather (it is) the establishment o f the functional equivalence

347

o f several possible causes from the viewpoint o f a problematic effect. ''~ A change in the concept of function is thereby introduced. It is no longer interpreted as an, "...effect to be caused, (but) rather (as) a regulative meaning-schema that organizes a domain of comparison of equivalent performances. ' ' 7 Thereby the functional method becomes one of comparison, a comparison of alternative possibilities that have nothing in common except that they provide a solution to a problem. 18 The advantage that the comparative method enjoys results from the fact that since it does not try to establish invariant relations between causes (performances) and effects (system-maintenance) it thereby does not fall victim to all the criticisms of the causal-scientific version. As far as Luhrnann is concerned, the most important change resulting from this reinterpretation of the concept of function itself is the reversal that occurs in its relationship with the concept of the causal relation. If functions are interpreted as effects of system performances then this means that they are species of causal relations? 9 But when they are rather understood as regulative meaning schemas which organize domains of comparison this relationship is reversed. In this case then, "...the function is not a specific kind of causal relation, rather the causal relation is a case o f application o f functional ordering. "'2 This can be explained in the following manner. Ever since the rise of m o d e m science the concept of causality has been infected with a problem which Luhmann likes to call the "problematic of infinity," one that classical and medieval philosophers were able to avoid only insofar as they tried to reduce the chain of causal connections to a singular ground of being, wether it was water (Thales), the idea of the Good (Plato), the unmoved first mover (Aristotle) or God (Middle Ages). In doing this a finitude is introduced into the concept of causality because there was an end (of some kind) to the series of causal connections. Modem science, on the other hand, recognizes that every effect has infinitely many causes and that every cause has infinitely many effects. This understanding of causality is the one which ultimately forms the basis of the difficulties that social sciences have had in trying to establish causal relations. Of course, the ceteris-paribus phrase can always be invoked to circumvent this problem. But, as Luhmann (1970, p. 16) notes, then, "...these statements (the ones including the ceteris-paribus phrase J.B.) do not possess any empirical value if the exclusion of all other causal factors cannot factually be carried out. ''2~ Functional method, however, recognizes the "problematic of infinity" and attempts a more modest approach of (arbitrarily)

348 holding constant only one cause or only one effect and using this as the starting point for analysis. The cause or effect held constant in this way can be used as a reference point for the investigation of functional equivalent effects or causes as the case m a y be. The functional relation then, in distinction to the causal, does not hold between individual cause and individual effect but rather among causes - from the p o i n t o f view o f an effect (arbitrarily) held constant - or among effects from the point of view of a cause (arbitrarily) held constant. In this case then, a causal relation expresses the limit case o f functional ordering, i.e. the relation between one cause and an effect (arbitrarily) held constant or between one effect and a cause (arbitrarily) held constant. As can be seen from this, functional relations, n o w u n d e r s t o o d as regulative meaning schemas, are established independently o f causal relations because functions are no longer understood here exclusively as particular effects of particular causes. :2 The m e t h o d e m p l o y e d therefore changes to a comparative one which seeks to include rather than exclude further equivalent possibilities. :3

III All comparison requires a point o f reference from which alternative possibilities can be compared. According to Luhmann, a function is the relation between this reference point and the alternative possibilities. ~4 One o f the major difficulties for functional m e t h o d then b e c o m e s that of defining this reference-point because, as Luhrnann (1970, p. 56) admits, "Viewpoints for comparison can be chosen quite arbitrarily. One can compare actions from the point o f view o f their duration or their calorie consumption or from the point o f view o f the n u m b e r o f their on-lookers without influencing the truth o f the result o f the comparison b y the choice of the viewpoint. ''2s The functional - i.e. the comparative - m e t h o d would operate, as it were, in a theoretical vacuum if it was not supplemented b y a theory o f social systems which would "concretize", "...the class o f functionally equivalent alternatives...so that explanations or predictions become possible. ''26 It can do this because it interprets reference-points as problems and all social systems are, as Luhmann - relying u p o n Blau - notices, "plagued" with problems. 27 Functions are fulffflled n o t insofar as a determinate cause is found for a determinate effect b u t insofar as different possible solutions are found to be equivalent in their ability to solve a particular system problem. In the language o f cause and

349 effect this means that a function is not fulffiled when one discovers that A causes B but when one discovers that A,C,D, etc. are equivalent in their ability to cause B. :8 It can be argued, however, that this m e t h o d still does not meet the methodological requirements of science: explanation and prediction. L u h m a n n recognizes this. He admits (1970, p. 16) that, considered abstractly, "The reference problems obviously do not "explain" the factual occurrence of determinate performances. They have precisely the opposite meaning: to refer to other possibilities. ''29 This is also the reason that he can say (1970, p. 18) that, "The validity o f functional analysis does not depend on whether in the individual case the problem is solved, the effect occurs, the system continues to exist. ''3 Nevertheless, explanation and prediction are possible within the context o f comparative functional m e t h o d with t h e realization that, "...a system must solve several problems. ''31 This realization comes about, however, only after the introduction of a theory which will meet the methodological needs of functional analysis. According to Luhmann this is a systems-theory for which causes and effects, as functional viewpoints, become problems, n It then becomes the task o f functional analysis to try to find different possible system performances which can serve as solutions to the problem - i.e. reference point of comparison - in question. Luhmann (1970, p. 41), however, is quick to admonish, "...that the experienced problems...are not simply identical with the functional reference-problems. ''33 The reason for this is that the selection o f any o f the possible solutions indicated by the reference problem brings with it unforeseen consequences which appear as new problems on different system levels. In other words, the solution of one problem - or better yet, on one system level - creates further problems on other levels. All system problems can never definitively be solved. Luhmann refers to this state of affairs as the permanence of problems. The latter, however, should not be understood as a defect within systems nor as an embarrassment for systems-theory. It is, in fact, very important thereoticaUy, and for more than just one reason. On the one hand it means that the permanence o f a system, 3a "...is not suitable as reference point for functionalist analysis, ''3s for with the uncovering of ever new system problems new system performances are connected together into different structures permitting the system to undergo profound changes without surrendering its i d e n t i t y ? 6 And on the other hand, the permanence of problems - on different levels - makes explanation and prediction possible because not all functionally equivalent problem solutions are eufunctional with regard

350 to problems on other system levels. Indeed only very few - eventually only one - will have the fewest dysfunctional consequences. These are the ones which are the best or most likely to occur and in terms of which explanation and prediction occur. Of course, all of this requires concrete analysis for, "Which values belong to such a functional class or variable is a matter of empirical knowledge and by no means arises out of the formulation o f the reference point. ''37 The reference point (problem) however does guide the investigation for these values and provides a criterion for their inclusion in the class. Again, analysis reveals the subsequent problems arising from the insertion of one value (solution) and thereby contributes to the narrowing of the members of this class to those - or the one - which are (is) least dysfunctional. Explanation and prediction then become possible within the context of functional analysis. Luhmann illustrates what is meant here with the case of role conflicts in social systems. 38 Role conflicts, considered abstractly, can be solved in any number of ways. The different ways are the alternative possible solutions to the reference problem. Yet since the alternative possibilities are functionally equivalent it would seem that the results this method could yield would be, at best, ambiguous because, simply from the point of view of the problem (role conflicts), there would be no way to choose between the various possible solutions. But when a particular concrete system is analysed by means of the comparative method the flexibility in the selection of the various possible alternative solutions of the reference problem is greatly reduced. Take the case of the solution of role conflicts in a large bureaucracy and in a small family. Although the possible solutions, considered abstractly -- i.e. without regard to any particular system - remain the same for both large bureaucracy and small family, the other problems these respective systems have to solve restricts the choice in the solution to the problem of role conflicts. The institutionalization of precedence claims and the separation of persons and situations which might occasion conflicting behavior are among the functionally equivalent solutions to the problem o f role conflict in social systems. But while the separation of persons and situations which might occasion conflicting behavior is a genuine possibility in the case of large bureaucracies this solution comes into conflict with another problem which exists only for intimate groups: communicative openness without any restrictions concerning what can be discussed. According to Luhmann, then, small families will tend to solve their problems of role conflicts by means of precedence

351 claims, especially in the form of vocational roles. Otherwise the problems that arise through the separation of persons and situations - the neglect of communicative openness and its subsequent personality strain - would have to be compensated in some other way. The restriction of choices among possible alternative solutions to the reference problem - conditioned by the other problems a social system must solve - permits a prediction of which alternative the system will select. Only because of this preditability does Luhmann (1970, p. 18) believe that, "...a problem can function as the basis of an explanation and as the valid ground of an analysis. ''39 Functional analysis can, therefore, meet the methodological requirements of science as long as it is not considered abstractly.

IV What has been said so far has been in preparation for a presentation of aspects of Luhmann's interpretation of functional method which come very close to certain phenomenological investigations of the later Hussefl and Schutz. The similarity is most clearly exhibited in the concepts of "problem" and "alternative possibility". We have already seen that an integral - perhaps the essential - concept in the comparative functional method is the reference-point from which alternative possibilities can be compared. Without this there can be no comparison and therefore no comparative functional method. We have also seen that, for Luhmann, this reference-point assumes the form of a problem for which the alternative possibilities are, then, solutions. Given this presentation we can see in Husserl's later writings - particularly Erfahrung und Urteil - a close4 similarity in his thinking between the concepts of "problem" and "alternative possibilities". The similarity comes out most clearly in Erfahrung und Urteil section 21 where Husserl explains the difference between problematic and open possibility. 41 In investigating the origin of different modes of predicative judgments (certainty, possibility, probability) in pre-predicative experience Husserl makes use of the distinction between open and problematic possibilities. These are two entirely different modes of possibility for which the modalizing consciousness, in each case, has a basically different origin: in the case of open possibility naive certainty and in the case of problematic possibility doubt. It is part of the intentional structure of consciousness according to

352 Husserl that in the perceptual field objects present themselves only perspectivally. Each one of these perspectives (perceptual noernata), however, carries with it a horizon of co-intended perspectives (anticipations) of those sides which are not actually perceived. Moreover, each one of these perspectives - such as the back of the sphere at which I am now looking - is co-intended emptily, i.e. indeterminately but, at the same time, with a naive certainty. I anticipate, with naive certainty, that the object which I am now perceiving will evince a back-side somewhat in conformity with the front-side and possessing some color. For instance, if the object which I am perceiving is a sphere I anticipate, with naive certainty, that the back-side, perhaps, will be like the frontside in shape and size and that it will be colored. Since I have never seen this object before I do not k n o w what the back-side of the object is like. 42 It may be spherical like the front which I am actually perceiving or it may be deformed in some way. It may be of the same color as the front or of a different color. My naive anticipations, which are given in imaginative presentification, therefore, c a n be disappointed for instance, when I go around the sphere and see that it is in fact deformed. It is certain, however, that the naively anticipated back-side of the sphere will be of s o m e shape and s o m e color which is otherwise completely unspecified by the requirements of the intentional structure of consciousness. Any possible shape and any possible color can meet these requirements. I can imagine the back-side as uniformly spherical or of any irregular shape. Correlatively I can imagine the back-side as of the same color as the front or of any other color. These possibilities of shape and color - are, therefore, referred to by Husserl as " o p e n " , i.e. completely indeterminate within a delimited horizon. This notion of open possibility can be transformed into what Husserl calls problematic possibility when the naive certainty which ordinarily accompanies perceptual experience gives way to doubt. Doubt and naive certainty are completely different modalizations of consciousness in which possibilities are given. But because o f the radical difference between them the possibilities they present are o f an entirely different kind. We have just seen how the possibilities given in naive certainty are "open", i.e. completely indeterminate. Doubt, however, changes this situation significantly. In effect, it removes the indeterminacy from the open possibilities by imposing a structure u p o n them, unifying them as opposing alternatives. Normal perception operates in the mode of naive certainty until the open possibility which is imaginatively presentified is disappointed in actual experience. Perhaps I go around the object and see t h a t the
-

353 back-side is not as I had anticipated. In this case m y anticipation - m y imaginatively presentified open possibility - is cancelled, "crossedo u t " as Husserl says, by the actual perception of the back-side which differs from m y anticipation in some way. It is out o f this cancellation that the concept o f negation arises, according to Husserl. The object is not uniformly colored as I had anticipated. A new perceptual meaning supersedes the one that I had anticipated. According to Husserl, naive certainty and negation present themselves as two extreme modalizations o f perceptual experience between which falls the transitional mode o f doubt. *3 In the case o f doubt perception does not proceed with naive certainty but neither is m y anticipation completely cancelled. Perception, as it were, is caught between these two and cannot, for the moment, decide. Husserl - and later Schutz 44 -- illustrates what he means in this case with the example o f an observer standing before a department store window. Inside the window is a display o f a fgure. At first, perception proceeds normally - i.e. in the mode o f naive certainty - and the observer sees a h u m a n being in the window, no doubt an employee o f the store. But then a doubt arises for the observer. Is this really a h u m a n being in the window or merely a mannequin? Normal perception undergoes a change o f modalization at this point, according to Husserl. My anticipations no longer proceed with naive certainty. But they are not simply cancelled either. Rather, as Husserl says, a new content or interpretation (Gehalt) is superimposed upon the sense data 4s which remain identical for both the perception o f the human being and the percepo f the mannequin. The problem for consciousness resides here in the presence o f two competing interpretations. Of course, both the perception o f the human being and that o f the mannequin do not co-exist at the same time. Only one perception is given at any instant. But as long as the doubt is not resolved the one interpretation is merely "surpressed" by the other and may surface again and surpress its opposite. Neither o f these two interpretations is cancelled while the doubt persists. 4~ However, perception no longer proceeds with naive certainty. 47 The ego for whom the doubt exists vacillates between the interpretation, h u m a n being or mannequin. It is attracted to one interpretation and then the other. Eventually one o f these will prevail because in the continued process o f observation more "speaks for" it, as Husserl says, than for its opposite. 48 In any event, a concept o f possibility has its origin in this conflict o f inclinations to believe, says Husserl, because the observer is not certain that the figure before him is a human being or a mannequin. 49

354 He is, therefore, confronted with mere possibilities, not certainties (knowledge in the strong sense). But the possibilities presented to the observer who experiences doubt are quite different in kind from those presented to an observer who perceives in the mode of naive certainty. As indicated above, the latter observer anticipates an open (indeterminate) range of possibilities, all of which share equal status. This is not the case for the observer caught in doubt. His possibilities present themselves as alternatives. The figure is either a human being or a mannequin. He cannot, as it were, interpret this figure in any way, as a horse for instance, because there is nothing presented to him which "speaks for" this interpretation in this case. The kind of possibilities presented to an observer who experiences doubt are called "questionable" (fragliche) or, more properly, "problematic" possibilities by Husserl because they arise out of a problematic situation. In this case the possibilities present themselves as conflicting. This conflict (problem), however, places a reciprocal restriction on the possibilities in a way which is totally absent in the mode of naive certainty. It unifies them as true alternatives. The concept of the alternative, as can now be seen, can only arise in a situation where naive certainty no longer exists but knowledge - in the form of affirmation or negation - has not yet occurred. The alternatives, therefore, appear as possibilities (non-certainties) determined as conflicting interpretations. In other words, a problem occurs for which two or more definite possibilities (of solution) present themselves. Husserl refers to these conflicting, definite possibilities as alternatives.

V The similarities between Luhmann's presentation of the comparative functional method and Husserl's presentation of the concept o f problematic possibilities now becomes clear. Both interpret alternatives as definite possibilities of solution to a problem. The subject matter, however, is different in each case. Luhmann applies the comparative functional method to human actions. Husserl discusses problematic possibilities in terms of perception. The bridge between these t w o can be found, I think, in the writings of Alfred Schutz (1962), especially his essays entitled, "Choosing among Projects of Action" and Reflections on the Problem of Relevance, chapter Two. In each o f these Schutz - in a manner strikingly similar to Luhmann - interprets

355 human action in terms of choice or selection. "Man," says Schutz (1962, p. 83), "acting in the social world among and upon his fellowmen finds that the preconstituted social world imposes upon him at any m o m e n t several alternatives among which he has to choose. According to modern sociology, the actor has 'to define the situation'. By doing so he transforms his social environment of 'open possibilities' (of action J.B.) into a unified field of 'problematic possibilities' within which choice and decision - especially so-called 'rational' choice and decision -- become possible." The transition from open to problematic possibilities, which is so important to the making of a rational decision (choice, selection), is, as we have attempted to show, one from naive certainty to doubt. "Defining the situation", therefore, becomes a process of providing clear contours to the particular problem at hand. In this way definiteness is imposed upon the overwhelmingly indefinite complexity of our social environment and rational action made possible. According to Schutz (1970, pp. 2 6 - 4 5 ) , this process requires systems of relevances, both topical and interpretative. The f'trst (topical) system of relevances governs those aspects of conscious life which divide consciousness into theme and field. It consists of memory and what Schutz refers to as one's "stock-of-knowledge-at-hand". In our department store window example this system operates when our attention is drawn to the figure in the window. Aspects of our memory and stockof-knowledge-at-hand single out this figure as of interest to us. It becomes of topical relevance to us, a theme of consciousness. This is the first step in transforming the social environment of open possibilities into a unified field of problematic possibilities. The figure (the theme) has become a problem (of interpretation) for consciousness, viz. what is this figure? Interpretation, according to Schutz, employs a system of interpretative relevances. It subsumes the theme (problem) under a type which is already part of one's stock-of-knowledge-at-hand. In our example, the figure is subsumed under the type "human being" and then under the type "mannequin". These interpretations become what Schutz and Husserl call problematic possibilities because they constitute a unified field of choice. They comprise the domain of possible alternative interpretations of the theme (problem) of consciousness. The importance of problematic possibilities, therefore, resides not in the actual choice made between or among them but rather in the establishment of the very possibility of choice itself. For m o d e m sociology, which interprets rational action as dependent upon a process of choice, the absence of such possibilities would render choice impos-

356 sible, sl T h e p r o b l e m a t i c possibilities are, therefore, i n d i s p e n s a b l e to a c o n c e p t o f c o m p a r a t i v e f u n c t i o n a l m e t h o d such as L u h m a n n ' s b e c a u s e his m e t h o d aims, as we have seen, n o t at the e s t a b l i s h m e n t o f an invariant relation b e t w e e n causes and effects b u t at the e s t a b l i s h m e n t o f a relation o f f u n c t i o n a l equivalence a m o n g alternatives. I n this sense, t h e n , we can say with L u h m a n n ( 1 9 7 0 , p. 52), " . . . t h a t t h e conc e p t i o n o f a r a t i o n a l i t y based u p o n c o m p a r i s o n w o u l d be a p p r o p r i a t e t o bring p h e n o m e n o l o g y a n d f u n c t i o n a l i s m closer t o g e t h e r . ''s2

NOTES 1. Functional method, as it is used here, refers to a procedure of scientific investigation first employed by social scientists such as WiUiam Robertson Smith and Emile Durkheim and then by Bronislaw Malinowski and A. RadcliffeBrown. The term "functionalism" is introduced by Malinowski in Encyclopedia Britannica (I 3th ed. 1926). 2. Kingsley Davis, "The Myth of Functional Analysis as a Special Method in Sociology and Anthropology" in N.J. Demerath and R.A. Peterson, System, Change and Conflict, pp. 379-402. This essay was first published in American Sociological Review, (24), 757-772, (I 959). 3. Indeed, one should say that the establishment of a science of sociology was due to the development of a unified (causal) sociological method. 4. There are numerous presentations of what is referred to here as the "unified scientific method". Of present concern, however, is the relation between explanation and prediction within this method. Hempel (1965) presents the orthodox interpretation of the structural identity of the two. This interpretation, however, is, as Hempel himself remarks, by no means universally accepted. Some, such as Scheffler, Scriven, Barker and Toulmin question it. Common to their criticisms, nevertheless, is the problematic status of predictions. In other words, as far as the latter are concerned, for various reasons every explanation is not necessarily a potential prediction. It is necessary to mention the problematic status of prediction for those who are otherwise adherents of the "unified scientific method" because Luhmann himself seems to represent the orthodox interpretation. This, of course, does not mean that he is either an advocate or a critic of the "unified scientific method". He, rather, intends to show that if the concept of function is subordinated to that of causal relation then (insuperable?) methodological problems arise for functional "explanations". 5. "Denn seitdem die Kausalbeziehung einen eindeutigen zeitlichen Richtungssinn erhalten hat (den sic weder fuer griechische noch fuer mittelalterliche Denker besass), koennen Wirkungen irgendwelcher Art das Vorkommen von Ursachen nicht mehr erklaeren." 6. "...die Wirkungen zu einer tragfaehigen Erklaerungsgrundlage ausbauen.'" 7. Cf. Malinowski, (1944, p. 169), "This type of functional analysis is easily exposed to the accusation of tautology and platitude, as well as to the criticism that it implies a logical circle, for, obviously, if we define function as the

357 satisfaction of a need, it is easy to suspect that the need to be satisfied has been introduced in order to satisfy the need of satisfying a function." 8. Cf. Harry C. Bredemeier (1955), esp. pp. 1 7 6 - 1 8 0 . 9. In biology the concept of "equilibrium" was expressed by the concept "homeostasis". In this case living organisms possess built4n structures and processes which come into play in order, for instance, to maintain constant body temperature or heal wounds etc. when something in the organism's environment disturbs it. 10. "Ein soziales System ist nicht, wie ein Organismus, typenfest fixiert. Aus einem Esel kann keine Schlange werden, selbst wenn eine solche Entwicldung zum Ueberleben notwendig waere. Eine Sozialordnung kann dagegen tiefgreifende struktureUe Aenderungen erfahren, ohne ihre Identitaet und ihren kontinuierlichen Bestand aufzugeben." 11. Luhmann attributes the concept to Gouldner. But he also mentions that it is subsequently to be found in Parsons and Homans. 12. Cf. Luhmann (1970, p. 36), "The classical school philosophy conversely had expected information about the genuine, unchangeable (the word "umwandelbare" appears at this place in the text. I have taken this for a misprint because it would destroy the sense of what Luhmann is saying. I read "unwandelbare" for "umwandelbare") being of a thing seemed to appear in qualities which were always and everywhere the same. It (the classical school philosophy) took sameness of the phenomenon for signs of the true being that excludes its nonbeing, and thereby all other possibilities, and, in this sense, is itself (substance)." "Die klassische Schulphilosophie hatte gerade umgekehrt yon der Betrachtung des Gleichen Aufschluss ueber das engentliche, umwandelbare (sic.) Sein des Seiende erwartet. In immer und ueberall gieichen Qualitaeten schien ihr das Wesen einer Sache zum Vorschein zu kommen. Sie nahm Gleichheiten tier Erscheinung als Zeichen fuer das wahrhaft Seiende, das sein Nichtsein und damit alle anderen Moeglichkeiten ausschliesst und in diesem Sinne es selbst (Substanz) ist." 13. R. Ingarden (1925, p. 187), "...als einen Gegenstand definiert .... tier nur die gemeinsamen Eigenschaften aller zu einer Klasse gehoerenden Gegenstaende in sich enthaelt." Luhmann cites Ingarden as referring to this specific ontological presupposition in the above work. 14. Mention of this ontological presupposition is significant in distinguishing between causal-scientific and the comparative functional methods. According to this presupposition, for which the permanence of any complex structure (system) is determined by several constant attributes, a principle of exclusion predominates. What is not common to all members of a group is excluded from its essence. If the essence of any complex structure is what confers unity (permanence) upon it, then this can only occur through a process of excluding what does not belong to its essence. The causal-scientific functional method makes this presupposition according to Luhmann. It interprets systems as wholes which are maintained causally by several constant performances. The task, then, of this method is to discover these performances through a process of exclusion of the performances that do not contribute causally to the maintenance of the system. By a process of exclusion it beheves that it can establish causal functional laws.

358 The comparative functional method proceeds not by exclusion but by inclusion because it does not interpret the permanence of a system in t e r m s of constant (invariant) relations between causes and effects. Its task is r a t h e r to discover performances which are equivalent in their function of solving particular system problems. In this case the performances need have n o t h i n g in common except, of course, their function of solving system-problems. The situation is quite different in the case of the causal-scientific functional method. 15. The two main consequences of this change are (1) the goal of functional method is no longer the explanation and prediction of specific events and ( 2 ) the concept of function is no longer a specific difference of the concept of causal relation. 16. Luhmann, (1970, p. 14), "...eine gesetzmaessige oder mehr oder weniger wahrscheinliche Beziehung zwischen bestimmten Ursachen und b e s t i m m t e n Wirkungen..., sondern...die Feststellungen der funktionalen Aequivalenz mehrerer

moeglicher Ursachen unter dem Gesichtspunkt einer problematischen Wirkung." 17. Ibid., p. 14, "...zu bewirkende Wirkung, sondern ein regulatives Sinnschema,
das einen Vergleichsbereich aequivalenter Leistungen organisiert. This is the task of the functional method to perform: to discover functional equivalences. 18. Cf. ibid., p. 27, "All functionalistic analyses are ultimately conducted i n reference to problems of stabilization." "Alle funktionalistische Analysen werden letzlich in Bezug auf Stabilisierungsprobleme als Leitfaden gefuehrt." p. 35, "Both in thinking and acting problem solving requires orientation towards alternatives. The problematic of thinking arises out of a c o m p e t i t i o n of different possibilities which structure them as alternatives. The p r o b l e m is meaningful if a comparison of the alternatives makes a solution of t h e problem possible." "Problemloesung erfordert im Denken und Handeln gleichermassen Orientierung an Alternativen. Die Problematik des Denkens besteht aus e i n e r Konkurrenz verschiedener Moeglichkeiten, einer Konkurrenz, welche die Moeglichkeiten als Alternativen strukturiert. Das Problem ist sinnvoU, wenn e i n Vergleich der Alternativen zur Loesung des Problems befaehigt," p. 13, "Thus the sense of functionalist analysis lies in the o p e n i n g of a (limited) domain of comparison." "Der Sinn funktionalistischer Analysen liegt mithin in der Eroeffnung eines (begrenzten) Vergleichsbereich." 19. The concept of causality is understood here, of course, to contain b o t h the concept of cause and the concept of effect. The latter is every bit as m u c h a causal category as the former. 20. Luhmann, (1970, p. 16), "...Die Funktion ist nicht eine Sonderart der Kausalbeziehung, sondern die Kausalbeziehung ist eine Anwendungsfall funktz'onaler

Ordnung.'"
Again, as understood here, the "causal relation" is an invariant c o n n e c t i o n between a particular cause and a particular effect. The functional relation, on the other hand, is a connection of several causes from the point of view o f a

359 particular effect or - and this is equally important - a connection of several effects from the point of view of a particular cause. 21. "...diese Aussagen besitzen keinen empirischen Wert, wenn die Ausschaltung aller anderen Kausalfaktoren faktisch nicht durchgefuehrt werden kann." 22. But not - it must be emphasized - independently of the concept of causality! The relation of foundation is not abrogated here, only reversed. The functional relations are established independently of causal relations because they hold between several causes or several effects, not between one cause and one effect. If this latter relation - the causal relation - can in fact be established then it is only as the limit case of comparison (of the functional method). 23. All this, of course, does not mean that causality is rejected as an epistemic category or that Luhmann intends to oppose functional to causal research. His critique of functional method has been performed precisely in order to establish it as an independent method by reversing the basic relation between causal relations and functional ones. 24. Cf. Luhmann, (1970, p. 56), "A function is the relation of a thing to an abstract - and thereby ambiguous - viewpoint, if this viewpoint serves to confront the thing with other, functionally equivalent possibilities." "Eine Funktion ist die Beziehung eines Seienden auf einen abstrakten - und damit mehrdeutigen - Gesichtspunkt, wenn dieser Gesichtspunkt dazu dient, das Seiende mit anderen, funktional aequlvalenten Moeglichkeiten zu konfrontieren." 25. "Vergleichsgesichtspunkten koennen rein logisch beliebig gewaehlt werden. Man kann Handlungen unter dem Gesichtspunkt ihrer Dauer oder ihres Kalorienverbrauchs oder unter dem Gesichtspunkt der Zahl ihrer Zuschauer vergleichen, ohne dass die Wahrheit des Vergleichsresultates durch die Wahl des Vergleichsgesichtspunktes beeinflusst wuerde." 26. Luhmann, (1970, p. 37), "...die Klasse der funktionat aequivalenten Alternativen...so dass Erklaerungen bzw. Voraussagen moeglich werden." 27. Cf. for the definition of reference as problem ibid., p. 18, "The reference point is viewed as (a) problem..." "Die Bezugseinheit wird als Problem gesehen..." p. 35, "The advance which the functional method introduces exists...in the fixing of an abstract view-point, viz. the "problem"..., from which different possibilities of action...can be treated as functionally equivalent." "Der Gewinn, den die funktionale Methode einbringt, besteht,...in der Fixierung eines abstrakten Bezugsgesichtpunktes, naemlich des "problems", von dem aus verschiedene Moeglichkeiten des Handelns,...als funktional aequivalent behandelt werden koennen." For the idea of systems 'plagued' with problems cf. ibid., p. 34, "It (the system) is therefore plagued with problems..." "Sic wird deshalb yon Problemen (organizational dilemmas) geplagt..." In the context in which this quote appears the subject of this phrase refers to "organizations" but it is also clear from the context that organization and system in this place are equivalent for Luhmann. Cf. P. Blau Dynamics o f Bureaucracy and P. Blau & W.R. Scott, Formal Organizations: A Comparative Approach.

360
28. It should be mentioned that Luhmann distinguishes two kinds of functional equivalence: disjunctive and conjunctive. The former is the kind presented in the text. The latter occurs when, for instance, several causes together enter into a relation of equivalence. In this case ABCD o r EFG o r HKM are equivalent in their ability to cause P. As can be seen, whole groups instead of individual causes are functionaUy equivalent here. 29. "Selbstverstaendlich "erklaeren" die Bezugsprobleme darum auch n i c h t das fakfische Vorkommen bestimmter funktionaler Leistungen. Sic haben gerade den entgegengesetzten Sinn: auf andere Moeglichkeiten hinzuweisen." 30. "Die Gueltigkeit funktionalisfischer Analysen haengt nicht davon ab, o b im Einzelfall das Problem geloest wird, die Wirkung eintritt, das System fortbesteht." 31. Luhmann, (1970, p. 38), "...ein System mehrere Probleme loesen muss." 32. Cf. Luhmann, (1970, p. 17). One can then ask in what sense do causes and effects become problems .They do so as events which in some way c o n t r i b u t e to the instability of the system because, according to Luhmann, all s y s t e m problems are essentially problems of stabilization. Cf. footnote 18 above. 33. "... dass die erlebten Probleme...nicht ohne weiteres idenfisch sind m i t den funktionalen Bezugsprobleme." 34. It must be made clear here that what is meant by "permanence of a s y s t e m " is its structure at any particular time. So that what is kept permanent is that particular structure. 35. Luhmann, (1970, p. 19), "...sich nicht als Bezugsgesichtspunkt fuer f u n k t i o nalistische Analysen eignet." 36. Cf. i b i d . , p. 18, "A social order (i.e. system) can, however, experience profound structural changes without giving up its identity and its c o n t i n u e d existence. It can change from an agrarian into an industrial society, a n extended family can become a tribe with transfamilial political order w i t h o u t one being able to decide when a new system exists." "Eine Sozialordnung kann dagegen tiefgreifende strukturelle Aenderungen erfahren, ohne ihre Identitaet und ihren kontinuierlichen Bestand aufzugeben. Sic kann sich aus einer AgrargeseUschaft in eine Industriegesellschaft verwandeln, aus einer Grossfamilie kann ein Stature mit ueberfamiliaerer pollfischer Ordnung werden, ohne dass entscheidbar waere, wann ein neues S y s t e m vorliegt." 37. Luhmann (1970, p. 15), "Welche Einsatzwerte zu einer solchen funktionalen Klasse oder Variable gehoeren, ist dagegen Sache empirischer Erkenntnis und ergibt sich keineswegs schon aus der Formulierung des Bezugsgesichtspunktes." 38. Role-conflicts therefore constitute the reference-problem (reference-point) of functional analysis in this case. 39. "...ein Problem als Erklaerungsbasis und als tragender Grund einer Analyse fungieren kann." 40. I can say "non-accidental" because Luhmann refers to it himself. Cf. (Luhmann, 1970, p. 49), footnote 16 and p. 35. "The problematic of thinking arises out of a competition between different possibilities which structures them as alternatives." "Die Problematik des Denkens besteht aus einer Konkurrenz, welche die Moeglichkeiten als Alternativen strukturiert."

361
41. The context of investigation in Erfahrung und Urteil is pre-predicative experience, i.e. the level of consciousness at which objectivities are constituted passively, non-thematically, viz. perception. Yet despite the very elementary level of investigation, what is said there refers, in Schutz's words, "... to activities of all kinds..." (Collected Papers I, p. 82), i.e. not just to the level of pre-predicative experience. For this reason what follows about open and problematic possibilities is not meant as a detailed analysis of pre-predicative experience but rather as an indication of the origin of the concept of open and problematic possibility. 42. This is also why Husserl speaks of the naive certainty involved in perceptual experience. In the case of objects I have never seen before I am naively certain that the back-side will be so and so etc. Experience, however, has shown me that my expectations can and often are disappointed. This is why my certainty is merely naive. I nevertheless perceive objects with this kind of certainty even if my anticipations are deceived. 43. Cf. Husserl (1972, p. 99), "Doubt presents a mode of transition towards negation..."

"Der Zweifel steUt einen Uebergangsmodus zur negierenden Aufhebung dar..."

44. The pervasiveness of this analysis of doubt in Husserl's writings can be demonstrated by its appearance in the works of entirely different periods of his career. Cf. (Husserl, 1968), Investigation V, section 27, (Husserl, 1980), section 103, and finally (Husseri, 1972), section 2 l b . Schutz uses the same example as (Husserl, 1972) above in his (1962, p. 8). He develops the concept much more fully in (1970, chapter 2). 45. The theory of hyletic data which Husserl maintained to the end was criticized by later phenomenologists, especially those influenced by Gestalt psychology. Cf. Aron Gurwitsch, "Phenomenology of Thematics and of the Pure Ego: Studies of the Relation between Gestalt Psychology and Phenomenology" section 16, in (Gurwitsch, 1966). Husserl discusses hyletic data in (Husserl, 1980) section 85. 46. Of course, the alternative need not be limited to two. 47. What Husserl means here is that each one of the different interpretations carries with it its own anticipations. The observer cannot, therefore, anticipate with naive certainty - as in the case of normal perception - in the case of doubt because the figure may turn out to be a mannequin rather than a human being and the anticipations of the former are of an entirely different kind than the latter. 48. The resolution of the doubt through continued observation of the figure also gives rise to the concept of probability, according to Husserl, because through the continued process of observation more "speaks for" one possibility than the other. One interpretation becomes "more likely", i.e. it is probably that the figure is a human being. 49. Cf. Husserl, (1972, p. 104), "...a concept of probability has its origin in this conflict of inclinations to believe..." "...in diesem Streit yon Glaubensneigungen...hat ein Begriff yon Moeglichkeit seinen Ursprung." 50. Schutz also makes use of a third system of motivational relevances. But for our

362 purposes the first two systems are all that are necessary. By systems of relevance Schutz means the structured aspects of conscious life which make the mind's selective activity possible. Cf. Schutz, (1962, p. 13), "...the theory concerning the mind's selective activity is simply the title for a set of problems...for the basic phenomenon we suggest calling relevance. 51. Cf. Dewey, (1957, p. 193), "Choice is not the emergence of preference out of indifference. It is the emergence of unified preference out of competing preferences." Schutz refers to Dewey often as a source of some of his ideas concerning choice and rational action. 52. "...dass die Konzeption einer auf Vergleich beruhenden Rationalitaet geeignet waere, Phaenomenologie und Funktionalismus einander naeherzubringen.'"

REFERENCES P. Blau, Dynamics of Bureaucracy, Chicago: University of Chicago Press, 1955. P. Blau & W.R. Scott, Formal Organizan'ons: A Comparative Approach, San Francisco: Chandler Publishing Co., 1962. H.C. Bredemeier, "The Methodology of Functionalism" in American Sociological Review, vol. 20, 1955, pp. 173-80. N.J. Demerath & R.A. Peterson, System, Change and Conflict, New York: Free Press, 1967. J. Dewey, Human Nature and Conduct, New York: Modern Library, 1957. Encyclopedia Britannica, 13th edition, 1926. A. Gurwitsch, Studies in Phenomenology and Psychology, Evanston, IU.: Northwestern University Press, 1966. C.G. Hempel, Aspects of Scientific Explanation, New York: Free Press, 1965. E. Husserl, Logische Untersuchungen, Tuebingen: Max Niemeyer, 1968. E. Husserl, Erfahrung und Urteil, Hamburg, Felix Meiner, 1972. E. Husserl, Ideen I, Tuebingen: Max Niemeyer, 1980. R. Ingarden, "Essentiale Fragen", in Jahrbuch fuer Philosophie und phaenomenologische Forschung, vol. 7, 1925, pp. 125-304. N. Luhmann, Soziologische A ufklaerung 1, Opladen: Westdeutscher Verlag, 1970. B. Malinowski, A Scientzfic Theory of Culture and Other Essays, Chapel Hill, N.C.: University of North Carolina Press 1944. A. Schutz, Collected Papers I, The Hague: Martinus Nijhoff, 1962. A. Schutz, Reflections on the Problem of Relevance, New Haven: Yale University Press, 1970.