Chapter II
Rules for the Observation of Social Facts
The first and most basic rule is to consider social facts as things. I
At the moment when a new order of phenomena becomes the object of a science they are already represented in the mind, not only through sense perceptions, but also by some kind of crudely formed concepts. Before the first rudiments of physics and chemistry were known, men already possessed notions about physical and chemical phenomena which went beyond pure perception alone. Such, for example, are those to be found intermingled with all religions. This is because reflective thought precedes science, which merely employs it more methodically. Man cannot live among things without forming ideas about them according to which he regulates his behaviour. But, because these notions are closer to us and more within our mental grasp than the realities to which they correspond, we naturally tend to substitute them for the realities, concentrating our speculations upon them. Instead of observing, describing and comparing things, we are content to reflect upon our ideas, analysing and combining them. Instead of a science which deals with realities, we carry out no more than an ideological analysis. Certainly this analysis does not rule out all observation. We can appeal to the facts to corroborate these notions or the conclusions drawn from them. But then the facts intervene only secondarily, as examples or confirmatory proof. Thus they are not the subject matter of the science, which therefore proceeds from ideas to things, and not from things to ideas.
60
Rules for the Observation of Social Facts   61
It is clear that this method cannot yield objective results. These notions or concepts - however they are designated - are of course not legitimate surrogates for things. The products of common experience, their main purpose is to attune our actions to the surrounding world; they are formed by and for experience. Now a representation can effectively perform this function even if it is theoretically false. Several centuries ago Copernicus dispelled the illusions our senses experienced concerning the movements of the heavenly bodies, and yet it is still according to these illusions that we commonly regulate the distribution of our time. For an idea to stimulate the reaction that the nature of a thing demands, it need not faithfully express that nature. It is sufficient for it to make us perceive what is useful or disadvantageous about the thing, and in what ways it can render us service or disservice. But notions formed in this way can only present a roughly appropriate practicality, and then only in the general run of cases. How often are they both dangerous and inadequate! It is therefore not by elaborating upon them, however one treats them, that we will ever succeed in discovering the laws of reality. On the contrary, they are as a veil interposed between the things and ourselves, concealing them from us even more effectively because we believe to be more transparent.
Such a science can only be a stunted one, for it lacks the subject matter on which to feed. It has hardly come into existence, one might say, before it vanishes, transmuted into an art. Allegedly its notions contain all that is essential to reality, but this is because they are confused with the reality itself. From then onwards they appear to contain all that is needful for us not only to understand what is, but also to prescribe what should be done and the means of implementation, for what is good is in conformity with the nature of things. What goes against nature is bad, and the means of attaining the good and eluding the bad both derive from that same nature. Thus if we have already comprehended the reality from the first, to study it has no longer any practical interest. Since it is this interest which is the reason for our study, there is henceforth no purpose to it. Our reflective thought is thus induced to turn away from what is the true subject matter of the science, namely the present and the past, and in one fell swoop to proceed to the future. Instead of seeking to understand the facts already discovered and acquired, it immediately undertakes to reveal new ones, more in accord with the ends that men pursue. If men think
62   The Rules of Sociological Method
they know what is the essence of matter, they immediately embark on the quest for the philosopher's stone. This encroachment of art upon science, which hinders the latter's development, is made easy also by the very circumstances which determine the awakening of scientific reflection. For, since this reflection comes into being only to satisfy vital needs, it is quite naturally directed towards practical matters. The needs which it is called upon to assuage are always pressing ones, and consequently urge it to arrive at conclusions. Remedies, not explanations, are required.
This procedure is so much in accordance with the natural inclination of our mind that it is even to be found in the beginnings of the physical sciences. It is what characterises alchemy as distinct from chemistry, and astrology from astronomy. It is how Bacon characterises the method followed by the scholars of his day - one which he fought against. Indeed the notions just discussed are those notiones vulgares, or praenotiones,1 which he points out as being at the basis of all the sciences,2 in which they take the place of facts.3 It is these idola which, resembling ghost-like creatures, distort the true appearance of things, but which we nevertheless mistake for the things themselves. It is because this imagined world offers no resistance that the mind, feeling completely unchecked, gives rein to limitless ambitions, believing it possible to construct - or rather reconstruct - the world through its own power and according to its wishes.
If this has been true for the natural sciences, -how much more had it to be true for sociology. Men did not wait on the coming of social science to have ideas about law, morality, the family, the state or society itself, for such ideas were indispensable to their lives. It is above all in sociology that these preconceptions, to employ again Bacon's expression, are capable of holding sway over the mind, substituting themselves for things. Indeed, social things are only realised by men: they are the product of human activity. Thus they appear to be nothing save the operationalising of ideas, which may or may not be innate but which we carry within us, and their application to the various circumstances surrounding men's relationships with one another. The organisation of the family, of contracts, or repression, of the state and of society seems therefore to be a simple development of the ideas we have about society, the state, justice, etc. Consequently these and similar facts seem to lack any reality save in and through the ideas
Rules for the Observation of Social Facts 63
which engender them and which, from then on, become the subject matter proper of sociology.
The apparent justification for this view derives from the fact that since the details of social life swamp the consciousness from all sides, it has not a sufficiently strong perception of the details to feel the reality behind them. Lacking ties that are firm enough or close enough to us, this all produces the impression upon us that it is clinging to nothing and floating in a vacuum, consisting of matter half unreal and infinitely malleable. This is why so many thinkers have seen in the social organisation mere combinations which are artificial and to some degree arbitrary. But if the details and the special concrete forms elude us, at least we represent to ourselves in a rough, approximate way the most general aspects of collective existence. It is precisely these schematic, summary representations which constitute the prenotions that we employ in our. normal way of life. Thus we cannot visualise their existence being called into, question, since we see it at the same time as we see our own. Not only are they within us, but since they are the product of repeated experiences, they are invested with a kind of ascendancy and authority, by dint of repetition and the habit which results from it. We feel their resistance when we seek to free ourselves from (hem, and we cannot fail to regard* as real something which pits itself against us. Thus everything conspires to make us see in them the true social reality.
And indeed up to now sociology has dealt more or less exclusively not with things, but with concepts. It is true that Comte proclaimed that social phenomena are natural facts, subject to natural laws. In so doing he implicitly recognised their character as things, for in nature there are only things. Yet when, leaving behind these general philosophical statements, he tries to apply his principle and deduce from it the science it contained, it is ideas which he too takes as the object of his study. Indeed, what constitutes the principal subject matter of hi^ sociology is the progress over time of humanity. His starting point is the idea that the continuous evolution of the human species consists of an ever-growing perfection of human nature. The problem with which he deals is how to discover the sequence of this evolution. Yet, even supposing this evolution exists, its reality can only be established when the science has been worked out. Thus the evolution cannot be made the subject of research unless it is
64   The Rules of Sociological Method
postulated as a conception of the mind, and not a thing. In fact, so much is this a wholly subjective idea, this progress of humanity does not exist. What do exist, and what alone are presented to us for observation, are particular societies which are born, develop and die independently of one another. If indeed the most recent societies were a continuation of those which had preceded them, each superior type might be considered merely as the repetition of the type at the level immediately below it, with some addition. They could all then be placed end-on, so to speak, assimilating together all those at the same stage of development; the series thus formed might be considered representative of humanity. But the facts do not present themselves with such extreme simplicity. A people which takes the place of another is not merely a prolongation of the latter with some new features added. It is different, gaining some extra properties, but having lost others. It constitutes a new individuality, and all such distinct individualities, being heterogeneous, cannot be absorbed into the same continuous series, and above all not into one single series. The succession of societies cannot be represented by a geometrical line; on the contrary, it resembles a tree whose branches grow in divergent directions. Briefly, in his consideration of historical development, Comte has taken his own notion of it, which is one that does not differ greatly from that commonly held. It is true that, viewed from a distance, history does take on somewhat neatly this simple aspect of a series. One perceives only a succession of individuals all moving in the same direction, because they have the same human nature. Moreover, since it is inconceivable that social evolution can be anything other than the development of some human idea, it appears entirely natural to define it by the conception that men have of it. But if one proceeds down this path one not only remains in the realm of ideology, but assigns to sociology as its object a concept which has nothing peculiarly sociological about it.
Spencer discards this concept, but replaces it with another which is none the less formed in the same way. He makes societies, and not humanity, the object of his study, but immediately gives to societies a definition which causes the thing of which he speaks to disappear and puts in its place the preconception he has of them. In fact he states as a self-evident proposition that 'a society is formed only when, besides juxtaposition, there is co-operation'; it
Rules for the Observation of Social Facts 65
is solely in this way that the'union of individuals becomes a society proper.4 Then, starting from this principle, that co-operation is the essence of social life, he divides societies into two classes according to the nature of the predominant mode of co-operation. There is', he states, 'a spontaneous co-operation which grows up without thought during the pursuit of private ends; and there is a cooperation which, consciously devised, implies distinct recognition of public ends',5 The first category he dubs industrial societies, the latter military societies. One may say of this distinction that it is the seminal idea for his sociology.
But this initial definition enunciates as a thing what is only a mental viewpoint. It is presented as the expression of a fact that is immediately apparent, one sufficiently ascertained by observation, since it is formulated from the very beginning of the science as an axiom. Yet from mere inspection it is impossible to know whether co-operation really is the mainspring of social life. Such an assertion is only scientifically justified if at first all the manifestations of collective life have been reviewed and it has been demonstrated that they are all various forms of co-operation. Thus once again a certain conception of social reality is substituted for that reality.6 What is defined in this way is not society but Spencer's idea of it. If he fe'els no scruples in proceeding in this fashion it is because for him also society is only, and can be only, the realisation of an idea, namely that very idea of co-operation by which he defines society.7 It would be easy to show, in each of the particular problems that he tackles, that his method remains the same. Also, although he has an air of proceeding empirically, because the facts accumulated in his sociology are used to illustrate analyses of notions rather than to describe and explain things, they seem indeed to be there to serve as arguments. All that is really essential in his doctrine can be directly deduced from his definition of society and the different forms of co-operation. For if we have only the choice between co-operation tyranically imposed and one that is free and spontaneous, it is plainly the latter which is the ideal towards which humanity does and ought to strive.
These common notions are not to be encountered only at the basis of the sciences, but are also to be found constantly as the arguments unravel. In our present state of knowledge we do not know exactly what the state is, nor sovereignty, political freedom, democracy, socialism, communism, etc. Thus our method should
66   The Rules of Sociological Method
make us forswear any use of these concepts so long as they have not been scientifically worked out. Yet the words that express them recur continually in the discussions of sociologists. They are commonly used with assurance, as if they corresponded to things well known and well defined, while in fact they evoke in us only confused notions, an amalgam of vague impressions, prejudices and passions. Today we mock at the strange ratiocinations that the doctors of the Middle Ages constructed from their notions of heat and cold, humidity and dryness, etc. Yet we do not perceive that we continue to apply the selfsame method to an order of phenomena which is even less appropriate for it than any other, on account of its extreme complexity.
In the specialised branches of sociology this ideological character is even more marked.
It is particularly so in the case of ethics. It may in fact be asserted that there is not a single system which does not represent it as the simple development of an initial idea which enshrines it potentially in its entirety. Some believe that men possess this idea complete at birth; on the other hand, others believe that it has grown up at a varying rate in the course of history. But for both empiricists and rationalists this is all that is truly real about morality. As for detailed legal and moral rules, these would have, in a manner of speaking, no existence per se, being merely applications of the basic notion to the particular circumstances of living, and varying according to different cases. Hence the subject matter of morality cannot be this unreal system of precepts, but the idea from which the precepts derive and which is interpreted differently according to cases. Thus all the questions that ethics normally raises relate not to things but to ideas. We must know what constitutes the ideas of law and morality and not what is the nature of morality and law considered in their own right. Moralists have not yet even grasped the simple truth that, just as our representations of things perceived by the senses spring from those things themselves and express them more or less accurately, our representation of morality springs from observing the rules that function before our very eyes and perceives them systematically. Consequently it is these rules and not the cursory view we have of them which constitute the subject matter of science, just as the subject matter of physics consists of actual physical bodies and not the idea that ordinary people have of it. The outcome is that the
Rules for the Observation of Social Facts 67
basis of morality is taken* to be what is only its superstructure, namely, the way in which it extends itself to the individual consciousness and makes its impact upon it. Nor is it only for the more general problems of science that this method is followed; it is not modified even for more specialised questions. From the essential ideas that he studies at the outset the moralist passes on to the examination of second-order ideas, such as family, country, responsibility, charity and justice - but it is always to ideas that his thinking is applied.
The same applies to political economy. John Stuart Mill states that its subject matter is the social facts which arise principally or exclusively with a view to the acquisition of wealth.8 But, in order for the facts defined in this way to be submitted to the scrutiny of the scientist as things, at the very least it should be possible to indicate the means whereby those which satisfy this condition can be recognised. With a new science one is no position to affirm that the facts exist, and even less to know what they are. In any kind of investigation it is only when the explanation of the facts is fairly well advanced that it is possible to establish that they have a goal and what that goal is. There is no problem more complex or less likely to be resolved at the very beginning. We therefore lack any prior assurance that a spheTe of social activity exists where the desire for wealth really plays this predominant role. Consequently the subject matter of economics so conceived is made up not of realities which may be precisely pointed to, but merely of possible ones, pure conceptions of the mind. They are facts which the economist conceives of as relating to the purpose under consideration, and facts as he conceives them. If, for example, he embarks on a study of what he terms production, he believes it possible immediately to spell out and review the principal agencies which assist it. This means therefore that* he has not ascertained their existence by studying on what conditions depends the thing that he is studying. If he had, he would have begun, by setting out the operations from which he drew that conclusion. If, in summary terms, at the beginning of his researches he proceeds to make such a classification, it is because he has arrived at it by mere logical analysis. He starts from the idea of production and as he dissects it he finds that it logically entails ideas of natural forces, of work, of tools or capital and he then goes on to treat in the same way these ideas which he has derived.9
68  The Rules of Sociological Method
The most basic economic theory of all, that of value, has clearly been built up according to the same method. If value were studied as a fact having reality should be, the economist would show how the thing so designated could be identified; he would then classify its various kinds, testing by methodical inductions how these vary according to different causes, and finally comparing the various results in order to arrive at a general formulation. A theory could therefore only emerge when the science was fairly well advanced. Instead it is met with at the very beginning. To do this the economist contents himself with his own reflective thinking, evoking his idea of value, namely that of an object capable of being exchanged. He finds that this implies the ideas of utility and scarcity, etc., and it is from these fruits of his analysis that he constructs his definition. He doubtless backs it up with a few examples. But, reflecting on the countless facts which such a theory must explain, how can one concede the slightest validity of proof to the necessarily very few facts which are cited at random as they suggest themselves to him?
Thus in political economy, as in ethics, the role of scientific investigation is extremely limited, and that of art is preponderant. The theoretical part of ethics is reduced to a few discussions on the ideas of duty, goodness and right. But such abstract speculations do not strictly speaking constitute a science, since their purpose is not to determine what is, in fact, supreme moral law, but what ought to be. Likewise, what economists dwell on most in their researches is the problem of knowing, for example, whether society should be organised on individualistic or socialist lines; whether it is better for the state to intervene in industrial and commercial relations or abandon them entirely to private initiative; whether the monetary system should be based on monometallism of bimetallism, etc. Laws properly so called are very few; even those which by custom we call laws do not generally merit the term, but are merely maxims for action, or in reality practical precepts. For example, the celebrated law of supply and demand has never been established inductively as an expression of economic reality. Never has any experiment or methodical comparison been instituted to establish whether, in fact, it is according to this law that economic relations are regulated. All that could be done, and has been done, has been to demonstrate by dialectical argument that individuals should act in this way if they perceive
Rules for the Observation of Social Facts 69
what is in their best interest; any other course of action would be harmful to them, and if they followed it would indeed constitute an error of logic. It is logical that the most productive industries should be the most prized, and that those who hold goods most in demand and most scarce should sell them at the highest price. But this entirely logical necessity in no way resembles the one that the true laws of nature reveal. These express the relationships whereby facts are linked together in reality, and not the way in which it would be good for them to be linked.
What we state about this law can be repeated for all those that the orthodox school of economists term 'natural' and which, moreover, are scarcely more than special cases of this first law. They may be said to be natural in the sense that they enunciate the means which are, or may appear to be, natural to employ in order to reach some assumed goal. But they should not be termed so if by a natural law is understood any inductively verified mode of existence of nature. All in all, they are mere counsels of practical wisdom. If it has been possible to present them to a more or less plausible extent as a clear expression of reality, it is because, rightly or wrongly, the assumption has been that these counsels were effectively those followed by most men and in the majority of cases.
Yet social phenomena are things and should be treated as such. To demonstrate this proposition one does not need to philosophise about their nature or to discuss the analogies they present with phenomena of a lower order of existence. Suffice to say that they are the sole datum afforded the sociologist. A thing is in effect all that is given, all that is offered, or rather forces itself upon our observation. To treat phenomena as things is to treat them as data, and this constitutes the starting point for science. Social phenomena unquestionably display this characteristic. What is given is not the idea that men conceive of value, because that is unattainable; rather is it the values actually exchanged in economic transactions. It is also not some conception or other of the moral ideal; it is the sum total of rules that in effect determine behaviour. It is not the idea of utility or wealth; it is all the details of economic organisation. Social life may possibly be merely the development of certain notions, but even if this is assumed to be the case, these notions are not revealed to us immediately. They cannot therefore be attained directly, but only through the real phenomena that
70   The Rules of Sociological Method
express them. We do not know a priori what ideas give rise to the various currents into which social life divides, nor whether they exist. It is only after we have traced the currents back to their source that we will know from where they spring.
Social phenomena must therefore be considered in themselves, detached from the conscious beings who form their own mental representations of them. They must be studied from the outside, as external things, because it is in this guise that they present themselves to us. If this quality of externality proves to be only apparent, the illusion will be dissipated as the science progresses and we will see, so to speak, the external merge with the internal. But the outcome cannot be anticipated, and even if in the end social phenomena may not have all the features intrinsic to things, they must at first be dealt with as if they had. This rule is therefore applicable to the whole of social reality and there is no reason for any exceptions to be made. Even those phenomena which give the greatest appearance of being artificial in their arrangement should be considered from this viewpoint. The conventional character of a practice or an institution should never be assumed in advance. If, moreover, we are allowed to invoke personal experience, we believe we can state with confidence that by following this procedure one will often have the satisfaction of seeing the apparently most arbitrary facts, after more attentive observation, display features of constancy and regularity symptomatic of their objectivity.
In general, moreover, what has been previously stated about the distinctive features of the social fact gives us sufficient reassurance about the nature of this objectivity to demonstrate that it is not illusory. A thing is principally recognisable by virtue of not being capable of modification through a mere act of the will. This is not because it is intractable to all modification. But to effect change the will is not sufficient; it needs a degree of arduous effort because of the strength of the resistance it offers, which even then cannot always be overcome. We have seen that social facts possess this property of resistance. Far from their being a product of our will, they determine it from without. They are like moulds into which we are forced to cast our actions. The necessity is often ineluctable. But even when we succeed in triumphing, the opposition we have encountered suffices to alert us that we are faced with something independent of ourselves. Thus in considering facts as
Rules for the Observation of Social Facts 71
things we shall be merely conforming to their nature.
In the end, the reform that must be introduced into sociology is identical in every respect to that which has transformed psychology over the last thirty years. Just as Comte and Spencer declare that social facts are facts of nature, but nevertheless refuse to treat them as things, the different empirical schools had long recognised the natural character of psychological phenomena, while continuing to apply to them a purely ideological method. Indeed the empiricists, no less than their opponents, proceeded exclusively by introspection. But the facts observable in ourselves are too few, too fleeting and malleable, to be able to impose themselves upon the corresponding notions that habit has rooted in us and to prevail over them. Thus when these notions are not subject to some other control, no countervailing force exists; consequently they take the place of facts and constitute the subject matter of the science. Thus neither Locke nor Condillac considered physical phenomena objectively. It is not sensation they study, but a certain idea of it. This is why, although in certain respects they were its forerunners, scientific psychology arose only much later. It arose after it had been finally established that states of consciousness can and must be studied externally and not from the perspective of the individual consciousness which experiences them. This is the great revolution that has been accomplished in this field of study. All the special procedures and new methods which have enriched this science are only various expedients for realising more fully this basic idea. Such an advance remains to be accomplished in sociology, which must pass from the subjective stage, beyond which it has hardly progressed, to the objective stage.
This transition, moreover, is less difficult to accomplish in sociology than in psychology. Psychical facts naturally appertain to states of the individual, from whom they do not even appear to be separable. Internal by definition, such states cannot seemingly be treated as external save by doing violence to their nature. Not only is an effort of abstraction necessary, but a whole gamut of procedures and artifices as well, for them to be considered successfully from the external viewpoint. Social facts, on the other hand, display much more naturally and immediately all the characteristics of a thing. Law is enshrined in legal codes, the events of daily life are registered in statistical figures and historical
72 The Rules of Sociological Method
monuments, fashions are preserved in dress, taste in works of art. By their very nature social facts tend to form outside the consciousnesses of individuals, since they dominate them. To perceive them in their capacity as things it is therefore not necessary to engage in an ingenious distortion. From this viewpoint sociology has significant advantages over psychology which have hitherto not been perceived, and this should accelerate its development. Its facts are perhaps more difficult to interpret because they are more complex, but they are more readily accessible. Psychology, on the other hand, has not only difficulty in specifying its facts, but also in comprehending them. Thus one may legitimately believe that as soon as this principle of sociological method has been universally acknowledged and is put into practice, sociology will be seen to progress at a speed that its present slow rate of development would scarcely allow one to suppose, even making up the lead of psychology, which it owes solely to its prior historical place10.
II
But our predecessors' experience has shown us that, in order to realise in practice the truth just established, it is not enough to demonstrate it theoretically or even to absorb it oneself. The mind has such a natural disposition to fail to recognise it that inevitably we will relapse into past errors unless we submit ourselves to a rigorous discipline. We shall formulate the principal rules for this discipline, all of which are corollaries of the previous rule. (1) The first of these corollaries is: One must systematically discard all preconceptions. Special proof of this rule is unnecessary: it follows from all that we have stated above. Moreover, it is the basis of all scientific method. Descartes' method of doubt is in essence only an application of it. If at the very moment of the foundation of science Descartes prescribed a rule for himself to question all the ideas he had previously accepted, it is because he wished to use only concepts which had been scientifically worked out, that is, constructed according to the method that he devised. All those of another origin had therefore to be rejected, at least for the time being. We have seen that Bacon's theory of the idols has the same significance. The two great doctrines, so often placed in contradiction to each other, agree on this essential point. Thus
Rules for the Observation of Social Facts 73
the sociologist, either when he decides upon the object of his research or in the course of his investigations, must resolutely deny himself the use of those concepts formed outside science and for needs entirely unscientific. He must free himself from those fallacious notions which hold sway over the mind of the ordinary person, shaking off, once and for all, the yoke of those empirical categories that long habit often makes tyrannical. If necessity sometimes forces him to resort to them, let him at least do so in full cognisance of the little value they possess, so as not to assign to them in the investigation a role which they are unfit to play.
What makes emancipation from such notions peculiarly difficult in sociology is that sentiment so often intervenes. We enthuse over our political and religious beliefs and moral practices very differently from the way we do over the objects of the physical world. Consequently this emotional quality is transmitted to the way in which we conceive and explain our beliefs." The ideas that we form about them are deeply felt, just as are their purposes, thereby taking on such authority that they brook no contradiction. Any opinion which is embarrassing is treated as hostile. For example, a proposition may not accord with our view of patriotism or personal dignity. It is therefore denied, whatever may be the proofs advanced. We cannot allow It to be true. It is rejected, and our strong emotions, seeking a justification for so doing, have no difficulty in suggesting reasons which we find readily conclusive. These notions may even be so prestigious that they will not tolerate scientific examination. The mere fact of subjecting them, as well as the phenomena they express, to cold, dry analysis is repugnant to certain minds. The sociologist who undertakes to study morality objectively as an external reality seems to such sensitive souls bereft of moral sense, just as the vivisectionist seems to the ordinary person devoid of normal feelings. Far from admitting that these sentiments are subject to science, it is believed that it is to them one should address onself in order to construct the science of things to which they relate. 'Woe', writes an eloquent historian of religions, 'Woe to the scientist who approaches the things of God without having in the depths of his consciousness, in the innermost indestructible parts of his being, in which sleep the souls of his ancestors, an unknown sanctuary from which at times there arises the fragrance of incense, a verse of a psalm, a cry of sorrow or triumph that as a child, following his
74   The Rules of Sociological Method
brothers' example, he raised to heaven, and which suddenly joins him once again in communion with the prophets of yore!'11
One cannot protest too strongly against this mystical doctrine which - like all mysticism, moreover - is in essence only a disguised empiricism, the negation of all science. Feelings relating to social things enjoy no pride of place over other sentiments, for they have no different origin. They too have been shaped through history. They are a product of human experience, albeit one confused and unorganised. They are not due to some transcendental precognition of reality, but are the result of all kinds of disordered impressions and emotions accumulated through chance circumstance, lacking systematic interpretation. Far from bringing enlightenment of a higher order than the rational, they are composed exclusively of states of mind which, it is true, are strong but also confused. To grant them such a predominant role is to ascribe to the lower faculties of the intelligence supremacy over superior ones and to condemn oneself more or less to a rhetorical logomachy. A science constituted in this way can only satisfy those minds who prefer to think with their sensibility rather than their understanding, who prefer the immediate and confused syntheses of sensation to the, patient, illuminating analyses of the reason. Feeling is an object for scientific study, not the criterion of scientific truth. But there is no science which at its beginnings has not encountered similar resistances. There was a time when those feelings relating to the things of the physical world, since they also possessed a religious or moral character, opposed no less violently the establishment of the physical sciences. Thus one can believe that, rooted out from one science after another, this prejudice will finally disappear from sociology as well, its last refuge, and leave the field clear for the scientist.
(2) But the above rule is entirely negative. It teaches the sociologist to escape from the dominance of commonly held notions and to direct his attention to the facts, but does not state how he is to grasp the facts in order to study them objectively.
Every scientific investigation concerns "a specific group of phenomena which are subsumed under the same definition. The sociologist's first step must therefore be to define the things he treats, so that we may know - he as well - exactly what his subject matter is. This is the prime and absolutely indispensable condition of any proof or verification. A theory can only be checked if we
Rules for the Observation of Social Facts 75
know how to recognise the facts for which it* must account. Moreover, since this initial definition determines the subject matter itself of the science, that subject matter will either consist of a thing or not, according to how this definition is formulated.
To be objective the definition clearly must express the phenomena as a function, not of an idea of the mind, but of their inherent properties. It must characterise them according to some integrating element in their nature and not according to whether they conform to some more or less ideal notion. When research is only just beginning and the facts have not yet been submitted to any analysis, their sole ascertainablecharacteristicsare those sufficiently external to be immediately apparent. Those less apparent are doubtless more essential. Their explanatory value is greater, but they remain unknown at this stage of scientific knowledge and cannot be visualised save by substituting for reality some conception of the mind. Thus it is among the first group of visible characteristics that must be sought the elements for this basic definition. Yet it is clear that the definition will have to include, without exception or distinction, all the phenomena which equally manifest these same characteristics, for we have no reason nor the means to discriminate between them. These properties, then, are all that we know of reality. Consequently they must determine absolutely how the facts should be classified. We possess no other criterion which can even partially invalidate the effect of this rule. Hence the following rule: The subject matter of research must only include a group of phenomena defined beforehand by certain common external characteristics and all phenomena which correspond to this definition must be so included. For example, we observe that certain actions exist which all possess the one external characteristic that, once they have taken place, they provoke on the part of society that special reaction known as punishment. We constitute them as a group sui generis and classify them under a single heading: any action that is punished is termed a crime and we make crime, so defined, the subject matter of a special science of criminology. Likewise we observe within all known societies the existence of a smaller society outwardly recognisable because it is formed for the most part of individuals linked by a blood relationship and joined to each other by legal ties. From the relevant facts we constitute a special group to which we assign a distinctive name: phenomena of domestic life. We term every aggregate of
76   The Rules of Sociological Method
this kind a family and make the family, so defined, the subject matter of a specific investigation which has not yet received a special designation in sociological terminology. When we later pass on from the family in general to the different types of family, the same rule should be applied. For example, embarking upon a study of the clan, or the maternal or patriarchal family, we should begin by defining them according to the same method. The subject matter of each topic, whether general or specialised, should be constituted according to the same principle.
By proceeding in this way from the outset the sociologist is immediately grounded firmly in reality. Indeed, how the facts are classified does not depend on him, or on his own particular cast of mind, but on the nature of things. The criterion which determines whether they are to be grouped in a particular category can be demonstrated and generally accepted by everybody, and the observer's statements can be verified by others. It is true that a notion built up in this way does not always chime - or does not generally even chime at all - with the notion commonly held. For example, it is evident that acts relating to freedom of thought or lapses in etiquette which are so regularly and severely punished in many societies, from the viewpoint of common sense are not regarded as crimes when people consider those societies. In the same way a clan is not a family in the usual sense of the word. But this is of no consequence, for it is not simply a question of how we can discover with a fair degree of accuracy the facts to which the words of common parlance refer and the ideas that they convey. What has to be done is to form fresh concepts de novo, ones appropriate to the needs of science and expressed by the use of a special terminology. It is certainly not true that the commonly held concept is useless to the scientist. It serves as a benchmark, indicating to him that somewhere there exists a cluster of phenomena bearing the same name and which consequently are likely to possess common characteristics. Moreover, since the common concept is never without some relationship to the phenomena, it occasionally points to the approximate direction in which they are to be discovered. But as the concept is only crudely formulated, it is quite natural for it not to coincide exactly with the scientific concept which it has been instrumental in instituting.12
However obvious and important this rule is, it is scarcely observed at present in sociology. Precisely because sociology deals
Rules for the Observation of Social Facts 77
with things which are constantly on our lips, such as the family, property, crime, etc., very often it appears useless to the sociologist initially to ascribe a rigorous definition to them. We are so accustomed to using these words, which recur constantly in the course of conversation, that it seems futile to delimit the meaning being given to them. We simply refer to the common notion of them, but this is very often ambiguous. This ambiguity causes us to classify under the same heading and with the same explanation things which are in reality very different. From this there arises endless confusion. Thus, for example, there are two kinds of monogamous unions: the ones that exist in fact, and those that exist legally. In the first kind the husband has only one wife, although legally he may have several; in the second kind polygamy is legally prohibited. Monogamy is met with de facto in several animal species and certain societies at a lower stage of development, not sporadically, but indeed with the same degree of generality as if it had been imposed by law. When a tribe is scattered over a wide area the social bond is very loose and consequently individuals live isolated from each other: Hence every man naturally seeks a female mate, but only one, because in his isolated state it is difficult for him to secure several. Compulsory monogamy, on the other hand, is only observed in societies at the highest stage of development. These two kinds of conjugal union have therefore very different significance, and yet the same word serves to described them both. We commonly say that certain animals are monogamous, although in their case there is nothing remotely resembling a legal tie. Spencer, embarking on his study of marriage, uses the term monogamy, without defining it, in its usual and equivocal sense. Consequently for him the development of marriage appears to present an incomprehensible anomaly, since he thinks he can observe the higher form of sexual union from the very earliest stages of historical development, while it apparently tends to disappear in the intermediate period, only to reappear again later. He concludes from this that there is no consistent relationship between social progress in general and the progressive advance towards a perfect type of family life. A definition at the appropriate time would have obviated this
13
error.
In other cases great care is taken to define the subject matter of the research but instead of including in the definition and grouping
if
78   The Rules of Sociological Method
under the same heading all phenomena possessing the same external properties, a selection is made. Certain phenomena, a kind of elite, are chosen as those considered to have the sole right to possess these characteristics. The others are held to have usurped these distinctive features and are disregarded. It is easy to envisage that, using this procedure, only a subjective and partial notion can be obtained. Such a process of elimination can in fact only be made according to a preconceived idea, since at the beginnings of a science no research would have been able to establish whether such a usurpation was real, even assuming it to be possible. The phenomena selected can only have been chosen because, more than the others, they conformed to the ideal conception that had already been formed of that kind of reality. For example, Garofalo, at the beginning of his Criminologie, demonstrates extremely well that the point of departure for that science should be 'the sociological notion of crime'.14 Yet, in order to build up this notion, he does not compare indiscriminately all the actions which in different types of society have been repressed by regular punishment, but only certain of them, namely those which offend the normal and unchangeable elements in the moral sense. As for those moral sentiments which have disappeared as a result of evolution, for him they were apparently not grounded in the nature of things for the simple reason that they did not succeed in surviving. Consequently the acts which have been deemed criminal because they violated those sentiments seemed to him to have merited this label only through chance circumstances of a more or less pathological kind. But he proceeds to make this elimination by virtue of a very personal conception of morality. He starts from the idea that moral evolution, considered at the source or its close proximity, carries along with it all sorts of deposits and impurities which it then progressively eliminates; only today has it succeeded in ridding itself of all the extraneous elements which at the beginning troubled its course. But this principle is neither a self-evident axiom nor a demonstrated truth: it is only a hypothesis, which indeed nothing justifies. The variable elements of the moral sense are no less founded in the nature of things than those that are immutable; the variations through which the former elements have passed evidence the fact that the things themselves have varied. In zoology those forms peculiar to the lower species are not considered any less natural than those which recur at all
Rules for the Observation of Social Facts 79
levels on the scale of animal development. Similarly, those actions condemned as crimes by primitive societies, but which have since lost that label, are really criminal in relation to those societies just as much as those, we continue to repress today. The former crimes correspond to the changing conditions of social life, the latter to unchanging conditions, but the first are no more artificial than the rest.
More can be added to this: even if these acts had wrongly assumed a criminal character, they neverthless should not be drastically separated from the others. The pathological forms of a phenomenon are no different in nature from the normal ones, and consequently it is necessary to observe both kinds in order to determine what that nature is. Sickness is not opposed to health; they are two varieties of the same species and each throws light on the other. This is a rule long recognised and practised both in biology and psychology, and one which the sociologist is no less under an obligation to respect. Unless one allows that the same phenomenon can be due first to one cause and then to another -which is to deny the principle of causality - the causes which imprint upon an action, albeit abnormally, the distinctive mark of a crime, cannot differ in kind from those which normally produce the same effect. They are distinguishable only in degree, or because they are not operating in the same set of circumstances. The abnormal crime therefore continues to be a crime and must consequently enter into the definition of crime. But what happens? Thus Garofalo takes for the genus what is only the species or merely a simple variation. The facts to which his formulation of criminality are applicable represent only a tiny minority among those which should be included. His formulation does not fit religious crimes, or crimes against etiquette, ceremonial or tradition, etc., which, although they have disappeared from our modern legal codes, on the contrary almost entirely fill the penal law of past societies.
The same error of method causes certain observers to deny to savages any kind of morality.15 They start from the idea that our morality is the morality. But it is either clearly unknown among primitive peoples or exists only in a rudimentary state, so that this definition is an arbitrary one. If we apply our rule all is changed. To decide whether a precept is a moral one or not we must investigate whether it presents the external mark of morality. This
80   The Rules of Sociological Method
mark consists of a widespread, repressive sanction, that is to say a condemnation by public opinion which consists of avenging any violation of the precept. Whenever we are confronted with a fact that presents this characteristic we have no right to deny its moral character, for this is proof that it is of the same nature as other moral facts. Not only are rules of this kind encountered in more primitive forms of society, but in them they are more numerous than among civilised peoples. A large number of acts which today are left to the discretion of individuals were then imposed compul-sorily. We perceive into what errors we may fall if we omit to define, or define incorrectly.
But, it will be claimed, to define phenomena by their visible characteristics, is this not to attribute to superficial properties a kind of preponderance over more fundamental qualities? Is this not to turn the logical order upside down, to ground things upon their apex and not their base? Thus when crime is defined by punishment almost inevitably one runs the risk of being accused of wanting to derive crime from punishment, or, to cite a well known quotation, to see the source of shame in the scaffold rather than in the crime to be expiated. But the reproach is based upon a confusion. Since the definition, the rule for which we have just enunciated, is made at the beginnings of the science its purpose could not be to express the essence of reality; rather is it intended to equip us in order to arrive at this essence later. Its sole function is to establish the contact with things, and since these cannot be reached by the mind save from the outside, it is by externalities that it expresses them. But it does not thereby explain them; it supplies only an initial framework necessary for our explanations. It is not of course punishment that causes crime, but it is through punishment that crime, in its external aspects, is revealed to us. And it is therefore punishment that must be our starting point if we wish to understand crime.
The objection referred to above would be well founded only if these external characteristics were at the* same time merely accidental, that is, if they were not linked to the basic properties of things. In these conditions science, after having pointed out the characteristics, would indeed lack the means of proceeding further. It could not penetrate deeper into reality, since there would be no connection between the surface and the depths. But, unless the principle of causality is only empty words, when clearly deter-
Rules for the Observation of Social Facts   81
mined characteristics are to -be found identically and without exception in all phenomena of a certain order, it is assuredly because they are closely linked to the nature of these phenomena and are joined indissolubly to them. If any given set of actions similarly presents the peculiarity of having a penal sanction attached to it, it is because there exists a close link between the punishment and the attributes constituting those actions. Consequently, however superficial these properties may be, provided they have been methodically observed, they show clearly to the scientist the path that he must follow in order to penetrate more deeply into the things under consideration. They are the prime, indispensable link in the sequence later to be unfolded by science in the course of its explanations.
Since it is through the senses that the external nature of things is Revealed to us, we may therefore sum up as follows: in order to be objective science must start from sense-perceptions and not from concepts that have been formed independently from it. It is from observable data that it should derive directly the elements for its initial definition. Moreover, it is enough to call to mind what the task of scientific work is to understand that science cannot proceed otherwise. It needs concepts which express things adequately, as they are, and not as it is useful in practical living to conceive them. Concepts formed outside the sphere of science do not meet this criterion. It must therefore create new concepts and to do so must lay aside common notions and the words used to express them, returning to observations, the essential basic material for all concepts. It is from sense experience that all general ideas arise, whether they be true or false, scientific or unscientific. The starting point for science or speculative knowledge cannot therefore be different from that for common or practical knowledge. It is only beyond this point, in the way in which this common subject matter is further elaborated, that divergences will begin to appear. (3) But sense experience can easily be subjective. Thus it is a rule in the natural sciences to discard observable data which may be too personal to the observer, retaining exclusively those data which present a sufficient degree of objectivity. Thus the physicist substitutes for the vague impressions produced by temperature or electricity the visual representation afforded by the rise and fall of the thermometer or the voltmeter. The sociologist must needs observe the same precautions. The external characteristics where-
82   The Rules of Sociological Method
by he defines the object of his research must be as objective as possible.
In principle it may be postulated that social facts are more liable to be objectively represented the more completely they are detached from the individual facts by which they are manifested.
An observation is more objective the more stable the object is to which it relates. This is because the condition for any objectivity is the existence of a constant, fixed vantage point to which the representation may be related and which allows all that is variable, hence subjective, to be eliminated. If the sole reference points given are themselves variable, continually fluctuating in relationship to one another, no common measure at all exists and we have no way of distinguishing between the part of those impressions which depends on what is external and that part which is coloured by us. So long as social life has not succeeded in isolating itself from the particular events which embody it, in order that it may constitute itself a separate entity, it is precisely this difficulty which remains. As these events do not take on the same appearance each time nor from one moment to another and as social life is inseparable from them, they communicate to it their own fluctuating character. Thus social life consists of free-ranging forces which are in a constant process of change and which the observer's scrutinising gaze does not succeed in fixing mentally. The consequence is that this approach is not open to the scientist embarking upon a study of social reality. Yet we do know that social reality possesses the property of crystallising without changing its nature. Apart from the individual acts to which they give rise, collective habits are expressed in definite forms such as legal or moral rules, popular sayings, or facts of social structure, etc. As these forms exist permanently and do not change with the various applications which are made of them, they constitute a fixed object, a constant standard which is always to hand for the observer, and which leaves no room for subjective impressions or personal observations. A legal rule is what it is and there are no two ways of perceiving it. Since, from another angle, these practices are no more than social life consolidated, it is legitimate, failing indications to the contrary,16 to study that life through these practices.
Thus when the sociologist undertakes to investigate any order of social facts he must strive to consider them from a viewpoint where
Rules for the Observation of Social Facts 83
they present themselves in isolation from their individual manifestations. It is by virtue of this principle that we have studied elsewhere social solidarity, its various forms and their evolution, through the system of legal rules whereby they are expressed.17 In the same way, if an attempt is made to distinguish and classify the different types of family according to the literary descriptions imparted by travellers and sometimes by historians, we run the risk of confusing the widely differing species and of linking types extremely dissimilar. If, on the other hand, we take as the basis of classification the legal constitution of the family, and more especially the right of succession, we have an objective criterion which, although not infallible, will nevertheless prevent many errors.18 If we aim at a classification of different kinds of crime, the attempt must be made to reconstitute the various modes of living and the 'professional' customs in vogiie in the different worlds of crime. As many criminological types will be identified as there are organisational forms. To penetrate the customs and popular beliefs we will turn to the proverbs and sayings which express them. Doubtless by such a procedure we leave outside science for the time being the concrete data of collective life. Yet, however changeable that life may be, we have no right to postulate a priori its incomprehensibility. But in order to proceed methodically we must establish the prime bases of the science on a solid foundation, and not on shifting sand. We must approach the social domain from those positions where the foothold for scientific investigation is the greatest possible. Only later will it be feasible to carry our research further and by progressive approaches gradually capture that fleeting reality which the human mind will perhaps never grasp completely.
Notes
1. Bacon, Novum Organum, I, p.26.
2. Ibid., I, p.17.
3. Ibid., I, p.36.
4. H. Spencer, The Principles of Sociology, II, p.244 (London, Williams & Norgate,1882).
5. Ibid.,II,p.245.
6. This is moreover a conception which is controversial (cf. Division du travail social, II, 2, ss. 4).
84   The Rules of Sociological Method
1. Spencer, op. cit., II, p.244: 'Cooperation, then, is at once that which cannot exist without a society, and that for which a society exists.'
8. J.S. Mill, A System of Logic, vol. 2, book VI, ch. IX, p.496 (London, Longmans, Green Reader & Dyer, 1872): 'Political economy shows mankind as occupied solely in acquiring and consuming wealth.'
9. This trait emerges from the very expressions used by economists. They continually talk of ideas, of the ideas of utility, savings, investment and cost. (Cf. C. Gide, Principes de Fecanomie politique book III, ch. 1, ss.l; ch. 2, ss.l; ch. 3, ss. 1 [First edition, Paris, 1884].)
10. It it true that the greater complexity of social facts renders the science that relates to them more difficult. But, as compensation, precisely because sociology is the latest arrival on the scene, it is in a position to benefit from the progress realised by the lesser sciences, and to learn from them. This use of previous experience cannot fail to hasten its development.
11. J. Darmsteter, Les Prophetes d Israel (Paris, 1892) p.9.
12. It is in practice always the common concept and the common term which are the point of departure. Among the things that in a confused fashion this term denotes, we seek to discover whether any exist which present common external characteristics. If there are any, and if the concept formed by grouping the facts brought together in this way coincides, if not entirely (which is rare) but at least for the most part, with the common concept, it will be possible to continue to designate the former by the same common term, retaining in the science the expression used in everyday parlance. But if the difference is too considerable, if the common notion mixes up a number of different notions, the creation of new and special terms becomes a necessity.
13. It is the same absence of definition which has sometimes caused it to be stated that democracy occurred both at the beginning and the end of history. The truth is that primitive and present-day democracy are very different from each other.
14. R. Garofalo, Criminologie (Paris, 1888) p.2 (trans, by the author from the Italian).
15. J. Lubbock, Origins of Civilization, ch. VIII. More generally still, it is stated, no less inaccurately, that ancient religions are amoral or immoral The truth is that they have their own morality. [Durkheim may have read Sir John Lubbock's work in translation. It was published in French translation by E. Barbier in 1873. Two further editions in French followed in 1877 and 1881.]
16. For example, one should have grounds to believe that, at a given moment, law no longer expressed the real state of social relationships for this substitution to be invalid.
17. Cf. Division du travail social, L 1 •
18. Cf. Durkheim, 'Introduction a la sociologie de la famile', Annates de la Faculty des Lettres de Bordeaux, 1889.
Chapter III
Rules for the Distinction of the Normal from the Pathological
Observation conducted according to the preceding rules mixes up two orders of facts, very dissimilar in certain respects: those that are entirely appropriate and those that should be different from what they are - normal phenomena and pathological phenomena. We have even seen that it is necessary to include both in the definition with which all research should begin. Yet if, in certain aspects, they are of the same nature, they nevertheless constitute two different varieties between which it is important to distinguish. Does science have the means available to make this distinction?
The question is of the utmost importance, for on its solution depends one's conception of the role that science, and above all the science of man, has to play. According to a theory whose exponents are recruited from the most varied schools of thought, science cannot instruct us in any way about what we ought to desire. It takes cognisance, they say, only of facts which all have the same value and the same utility; it observes, explains, but does not judge them; for it, there are hone that are reprehensible. For science, good and evil do not exist. Whereas it can certainly tell us how causes produce their effects, it cannot tell us what ends should be pursued. To know not what is, but what is desirable, we must resort to the suggestions of the unconscious - sentiment, instinct, vital urge, etc., - by whatever name we call it. Science, says a writer already quoted, can well light up the world, but leaves a darkness in the human heart. The heart must create its own illumination. Thus science is stripped, or nearly, of all practical effectiveness and consequently of any real justification for its existence. For what good is it to strive after a knowledge of reality if the knowledge we acquire cannot serve us in our lives? Can we reply that by revealing to us the causes of phenomena knowledge
85
86   The Rules of Sociological Method
offers us the means of producing the causes at will, and thereby to achieve the ends our will pursues for reasons that go beyond science? But, from one point of view, every means is an end, for to set the means in motion it requires an act of the will, just as it does to achieve the end for which it prepares the way. There are always several paths leading to a given goal, and a choice must therefore be made between them. Now if science cannot assist us in choosing the best goal, how can it indicate the best path to follow to arrive at the goal? Why should it commend to us the swiftest path in preference to the most economical one, the most certain rather than the most simple one, or vice versa? If it cannot guide us in the determination of our highest ends, it is no less powerless to determine those secondary and subordinate ends we call means.
It is true that the ideological method affords an avenue of escape from this mysticism, and indeed the desire to escape from it has in part been responsible for the persistence of this method. Its devotees were certainly too rationalist to agree that human conduct did not require the guidance of reflective thought. Yet they saw in the phenomena, considered by themselves independently of any subjective data, nothing to justify their classifying them according to their practical value. It therefore seemed that the sole means of judging them was to relate them to some overriding concept. Hence-the use of notions to govern the collation of facts, rather than deriving notions from them, became indispensable for any rational sociology. But we know that, in these conditions, although practice has been reflected upon, such reflection is not scientific.
The solution to the problem just posed will nevertheless allow us to lay claim to the rights of reason without falling back into ideology. For societies, as for individuals, health is good and desirable; sickness, on the other hand, is bad and must be avoided. If therefore we find an objective criterion, inherent in the facts themselves, to allow us to distinguish scientifically health from sickness in the various orders of social phenomena, science will be in a position to throw light on practical matters while remaining true to its own method. Since at present science is incapable of directly affecting the individual, it can doubtless only furnish us with general guidelines which cannot be diversified appropriately for the particular individual unless he is approached through the senses. The state known as health, in so far as it is capable of
Rules for the Distinction of the Normal from the Pathological 87
definition, cannot apply exactly to any individual, since it can only be established for the most common circumstances, from which everyone deviates to some extent. None the less it is a valuable reference point to guide our actions. Because it must be adjusted later to fit each individual case, it does not follow that knowledge of it lacks all utility. Indeed, precisely the opposite is true, because it establishes the norm which must serve as a basis for all our practical reasoning. Under these conditions we are no longer justified in stating that thought is useless for action. Between science and art there is no longer a gulf, and one may pass from one to the other without any break in continuity. It is true that science can only concern itself with the facts through the mediation of art, but art is only the extension of science. We may even speculate whether the practical shortcomings of science must not continue to decrease as the laws it is establishing express ever more fully individual reality.
I
Pain is commonly regarded as the index of sickness. It is certain that in general a relationship exists between these two phenomena, although one lacking uniformity and precision. There are serious physical dispositions of a painless nature, whereas'minor ailments of no importance, such as that resulting from a speck of coal-dust in the eye, cause real torment. In certain cases it is even the absence of pain, or indeed the presence of positive pleasure, which is the symptom of ill-health. There is certain lack of vulnerability to pain which is pathological. In circumstances where a healthy man would be suffering, the neurasthenic would experience a sensation of enjoyment, the morbid nature of which is indisputable. Conversely, pain accompanies many conditions, such as hunger, tiredness and childbirth, which are purely physiological phenomena. '
May we assert that health, consisting in the joyous development of vital energy, is recognisable when there is perfect adaptation of the organism to its environment, and on the other hand may we term sickness as all that which upsets that adaptation? But first -and we shall have to return to this point later - it is by no means demonstrated that every state of the organism corresponds to some external state. Furthermore, even if the criterion of adapta-
i s
88   The Rules of Sociological Method
tion were truly distinctive of a state of health, some other criterion would be needed for it to be recognisable. In any case we should need to be informed of the principle to decide whether one particular mode of adaptation is more 'perfect' than another.
Is it according to the manner in which one mode rather than another affects our chances of survival? Health would be the state of the organism in which those chances were greatest, whereas sickness would be anything which reduced those chances. Unquestionably sickness has generally the effect of really weakening the organism. Yet sickness is not alone in being capable of producing this result. In certain lower species the reproductive functions inevitably entail death, and even in higher species carry risks with them. Yet this is normal. Old age and infancy are subject to the same effect, for both the old person and the infant are more vulnerable to the causes of destruction. But are they therefore sick persons, and must we admit that* the healthy type is represented only by the adult? This would be singularly to restrict the domain of health and physiology. Moreover, if old age is already a sickness in itself, how does one distinguish between a healthy old person and a sick one? By the same reasoning menstruation would have to be classified under pathological phenomena, for by the troubles that it brings on, it increases for a woman the liability to illness. Yet how can one term unhealthy a condition whose absence or premature disappearance constitutes without question a pathological phenomenon? We argue about this question as if in a healthy organism each element, so to speak, had a useful part to play, as if every internal state corresponded exactly to some external condition and consequently contributed to maintaining the vital equilibrium and reducing the chances of dying. On the contrary it may legitimately be presumed that certain anatomical or functional arrangements serve no direct purpose, but exist simply because they are, and cannot cease, given the general conditions of life. They cannot, however, be characterised as morbid, for sickness is eminently something avoidable which is not intrinsic to the normal constitution of a living creature. It may even be true that, instead of strengthening the organism, these arrangements lower its powers of resistance and consequently increase the risk of death.
On the other hand it is by no means sure that sickness always entails the consequence by which people have sought to define it. Do not a number of illnesses exist that are too slight for us to be
Rules for the Distinction of the Normal from the Pathological 89
able to attribute to them any perceptible effect upon the basic functions of the organism? Even among the gravest afflictions there are some whose effects are wholly innocuous, if we know how to combat them with the weapons at our command. The gastritis-prone individual who follows a good, hygienic way of living can live as long as the healthy man. Undoubtedly he is forced to take precautions, but are we not all subject to the same constraint, and can life be sustained otherwise? Each of us has his own hygiene to follow. That of the sick person differs considerably from that of his average contemporary, living in the same environment. But this may be seen to be the sole difference between them. Sickness does not always leave us at a loss, not knowing what to do, in an irremediable state of inadaptability; it merely obliges us to adapt ourselves differently from most of our fellows. Who is there to say that some sicknesses even exist which in the end are not useful to us? Smallpox, a vaccine of which we use to inoculate ourselves, is a true disease that we give ourselves voluntarily, yet it increases our chance of survival. There may be many other cases where the damage caused by the sickness is insignificant compared with the immunities that it confers upon us.
Finally and most importantly, this criterion is very often inapplicable. At the very most it can be established that the lowest mortality rate known is encountered in a particular group of individuals, but it cannot be demonstrated that an even lower rate might not be feasible. Who is to say that other conditions might not be envisaged which would have the effect of lowering it still further? The actual minimum is not therefore proof of perfect adaptation and is consequently not a reliable index of the state of health, to come back to the preceding definition. Moreover, a group with this characteristic is very difficult to constitute and to isolate from all other groups. Yet this would be necessary to be able to observe the bodily ^constitution of its members which is the alleged cause of their superiority. Conversely, in the case of a generally fatal illness it is evident that the probability of survival is lower, but the proof is signally more difficult to demonstrate in the case of an affliction which does not necessarily cause death. In fact there is only one objective way to prove that creatures placed in closely defined conditions have less chance of survival than others: this is to show that in fact the majority do not live as long. Now although in cases of purely individual sickness this can often be
90 The Rules of Sociological Method
demonstrated, it is utterly impracticable in sociology. For here we have not the criterion of reference available to the biologist, namely, the figures of the average mortality rate. We do not even know how to determine approximately the moment when a society is born and when it dies. All these problems, which even in biology are far from being clearly resolved, still remain wrapped in mystery for the sociologist. Moreover, the events occurring in social life and which are repeated almost identically in all societies of the same type, are much too diverse to be able to determine to what extent any particular one has contributed to hastening a society's final demise. In the case of individuals, as there are very many, one can select those to be compared so that they present only the same one irregularity. This factor is thus isolated from all concomitant phenomena, so that one can study the nature of its influence upon the organism. If, for example, about a thousand rheumatism sufferers taken at random exhibit a mortality rate above the average, there are good grounds for imputing this outcome to a rheumatoidal tendency. But in sociology, since each social species accounts for only a small number of individuals, the field of comparison is too limited for groupings of this kind to afford valid proof.
Lacking this factual proof, there is no alternative to deductive reasoning, whose conclusions can have no value except as subjective presumptions. We will be able to demonstrate, not that a particular occurrence does in fact weaken the social organism, but that it should have that effect. To do this it will be shown that the occurrence cannot fail to entail a special consequence esteemed to be harmful to society, and on these grounds it will be declared pathological. But, granted that it does bring about this consequence, it can happen that its deleterious effects are compensated, even over-compensated, by advantages that are not perceived. Moreover, only one reason will justify our deeming it to be socially injurious: it must disturb the normal operation of the social functions. Such a proof presumes that the problem has already been solved. The proof is only possible if the nature of the normal state has been determined beforehand and consequently the signs whereby normality may be recognised are already known. Could one try to construct a priori the normal state from scratch? There is no need to show what such a construction would be worth. This is why it happens in sociology, as in history, that the same events
"V I
Rules for the Distinction of the Normal from the Pathological 91
are judged to be salutary or .disastrous, according to the scholar's personal convictions. Thus it constantly happens that a theorist lacking religious belief identifies as a pathological phenomenon the vestiges of faith that survive among the general collapse of religious beliefs, while for the believer it is the very absence of belief which is the great social sickness. Likewise for the socialist, the present economic organisation is a fact of social abnormality, whereas for the orthodox economist it is above all the socialist tendencies which are pathological. To support his view each finds syllogisms that he esteems well founded.
The common weakness in these definitions is the attempt to reach prematurely the essence of phenomena. Thus they assume that propositions have already been demonstrated which, whether true or -false, can only be proved when the progress of science is sufficiently advanced. This is nevertheless a case where we should conform to the rule already established. Instead of claiming to determine at the outset the relationship of the normal state, and the contrary state, to the vital forces, we should simply look for some immediately perceptible outward sign, but an objective one, to enable us to distinguish these two orders of facts from each other.
Every sociological phenomenon, just as every biological phenomenon, although staying essentially unchanged, can assume a different form for each particular case. Among these forms exist two kinds. The first are common to the whole species. They are to be found, if not in all, at least in most individuals. If they are not replicated exactly in all the cases where they are observed, but vary from one person to another, their variations are confined within very narrow limits. On the other hand, other forms exist which are exceptional. These are encountered only in a minority of cases, but even when they occur, most frequently they do not last the whole lifetime of an individual. They are exceptions in time as they are in space.1 We are therefore faced with two distinct types of phenomena which must be designated by different terms. Those facts which appear in the most common forms we shall call normal, and the rest morbid or pathological. Let us agree to designate as the average type the hypothetical being which might be constituted by assembling in one entity, as a kind of individual abstraction, the most frequently occurring characteristics of the species in their most frequent forms. We may then say that the normal type
92  The Rules of Sociological Method
merges into the average type and that any deviation from that standard of healthiness is a morbid phenomenon. It is true that the average type cannot be delineated with the same distinctness as an individual type, since the attributes from which it is constituted are not absolutely fixed but are capable of variation. Yet it can unquestionably be constituted in this way since it is the immediate subject matter of science and blends with the generic type. The physiologist studies the functions of the average organism; the same is true of the sociologist. Once we know how to distinguish between the various social species - this question will be dealt with later - it is always possible to discover the most general form presented by a phenomenon in any given species.
It can be seen that a fact can be termed pathological only in relation to a given species. The conditions of health and sickness cannot be defined in abstracto or absolutely. This rule is not questioned in biology: it has never occurred to anybody to think that what is normal in a mollusc should be also for a vertebrate. Each species has its own state of health, because it has an average type peculiar to it, and the health of the lowest species is no less than that of the highest. The same principle is applicable to sociology, although it is often misunderstood. The habit, far too widespread, must be abandoned of judging an institution, a practice or a moral maxim as if they were good or bad in or by themselves for all social types without distinction.
Since the reference point for judging the state of health or sickness varies according to the species, it can vary also within the same species, if that happens to change. Thus from the purely biological viewpoint, what is normal for the savage is not always so for the civilised person and vice versa.2 There is one order of variations above all which it is important to take into account because these occur regularly in all species: they are those which relate to age. Health for the old person is not the same as it is for the adult, just as the adult's is different from the child's. The same is likewise true of societies.3 Thus a social fact can only be termed normal in a given species in relation to a particular phase, likewise determinate, of its development. Consequently, to know whether the term is merited for a social fact, it is not enough to observe the form in which it occurs in the majority of societies which belong to a species: we must also be careful to observe the societies at the corresponding phase of their evolution.
We may seem to have arrived merely at a definition of terms, for
Rules for the Distinction of the Normal from the Pathological 93
we have done no more than group phenomena according to their similarities and differences and label the groups formed in this way. Yet in reality the concepts so formed, while they possess the great merit of being identifiable because of characteristics which are objective and easily perceptible, are not far removed from the notion commonly held of sickness and health. In fact, does not everybody consider sickness to be an accident, doubtless bound up with the state of being alive, but one which is not produced normally? This is what the ancient philosophers meant when they declared that sickness does not derive from the nature of things but is the product of a kind of contingent state immanent in the organism. Such a conception is assuredly the negation of all science, for sickness is no more miraculous than health, which also inheres in the nature of creatures. Yet sickness is not grounded in their normal nature, bound up with their ordinary temperament or linked to the conditions of existence upon which they usually depend. Conversely the type of health is closely joined for everybody to the type of species. We cannot conceive incontro-vertibly of a species which in itself and through its own basic constitution would be incurably sick. Health is the paramount norm and consequently cannot be in any way abnormal.
It is true that health is commonly understood as a state generally preferable to sickness. But this definition is contained in the one just stated. It is not without good reason that those characteristics which have come together to form the normal type have been able to generalise themselves throughout the species. This generalisation is itself a fact requiring explanation and therefore necessitating a cause. It would be inexplicable if the most widespread forms of organisation were not also - at least in the aggregate - the most advantageous. How could they have sustained themselves in such a wide variety of circumstances if they did not enable the individual better to resist the causes of destruction? On the other hand, if the other forms are rarer it is plainly because - in the average number of cases - those individuals displaying such forms have greater difficulty in surviving. The greater frequency of the former class is thus the proof of their superiqrity.4
II
This last observation even provides a means of verifying the results of the preceding method.
94   The Rules of Sociological Method *
Since the generality which outwardly characterises normal phenomena, once directly established by observation, is itself an explicable phenomenon, it demands explanation. Doubtless we can have the prior conviction that it is not without a cause, but it is better to know exactly what that cause is. The normality of the phenomenon will be less open to question if it is demonstrated that the external sign whereby it was revealed to us is not merely apparent but grounded in the nature of things -* if in short, we can convert this factual normality into one which exists by right. Moreover, the demonstration of this will not always consist in showing that the phenomenon is useful to the organism, although for reasons just stated this is most frequently the case. But, as previously remarked, an arrangement may happen to be normal without serving any useful purpose, simply because it inheres in the nature of a creature. Thus it would perhaps be useful for childbirth not to occasion such violent disturbances in the female organism, but this is impossible. Consequently the normality of a phenomenon can be explained only through it being bound up with the conditions of existence in the species under consideration, either as the mechanically essential effect of these conditions or as a means allowing the organism to adapt to these conditions.5
This proof is not merely useful as a check. We must not forget that the advantage of distinguishing the normal from the abnormal is principally to throw light upon practice. Now, in order to act in full knowledge of the facts, it is not sufficient to know what we should want, but why we should want it., Scientific propositions relating to the normal state will be more immediately applicable to individual cases when they are accompanied by the reasons for them, for then it will be more feasible to pick out those cases where it is appropriate to modify their application, and in what way.
Circumstances even exist where this verification is indispensable, because the first method, if it were applied in isolation, might lead to error. This is what occurs in transition periods when the whole species is in the process of evolving, without yet being stabilised in a new and definitive form. In that situation the only normal type extant at the time and grounded in the facts is one that relates to the past but no longer corresponds to the new conditions of existence. A fact can therefore persist through a whole species but no longer correspond to the requirements of the situation. It
Rules for the Distinction of the Normal from the Pathological 95
therefore has only the appearance of normality, and the generality it displays is deceptive; persisting only through the force of blind habit, it is no longer the sign that the phenomenon observed is closely linked to the general conditions of collective existence. Moreover, this difficulty is peculiar to sociology. It does not exist, in a manner of speaking, for the biologist. Only very rarely do animal species require to assume unexpected forms. The only normal modifications through which tl\ey pass are those which occur regularly in each individual, principally under the influence of age. Thus they are already known or knowable, since they have already taken place in a large number of cases. Consequently at every stage in the development of the animal, and even in periods of crisis, the normal state may be ascertained. This is also still true in sociology for those societies belonging to inferior species. This is because, since a number of them have already run their complete course, the law of their normal evolution has been, or at least can be, established. But in the case of the highest and most recent societies, by definition this law is unknown, since they have not been through their whole history. The sociologist may therefore be at a loss to know whether a phenomenon is normal, since he lacks any reference point.
He can get out of this difficulty by proceeding along the lines we have just laid down. Having established by observation that the fact is general, he will trace back the conditions which determined this general character in the past and then investigate whether these conditions still pertain in the present or, on the contrary, have changed. In the first case he will be justified in treating the phenomenon as normal; in the other eventuality he will deny it that characteristic. For instance, to know whether the present economic state of the peoples of Europe, with the lack of organisation6 that characterises it, is normal or not, we must investigate what in the past gave rise to it. If the conditions are still those appertaining to our societies, it is because the situation is normal, despite the protest that it stirs up. If, on the other hand, it is linked to that old social structure which elsewhere we have termed segmentary7 and which, after providing the essential skeletal framework of societies, is now increasingly dying out, we shall be forced to conclude that this now constitutes a morbid state, however universal it may be. It is by the same method that all such controversial questions of this nature will have to be
96   The Rules of Sociological Method
resolved, such as those relating to ascertaining whether the weakening of religious belief and the development of state power are normal phenomena or not.8
Nevertheless this method should in no case be substituted for the previous one, nor even be the first one employed. Firstly it raises questions which require later discussion and which cannot be tackled save at an already fairly advanced stage of science. This is because, in short, it entails an almost comprehensive explanation of phenomena, since it presupposes that either their causes or their functions are determined. At the very beginning of our research it is important to be able to classify facts as normal or abnormal, except for a few exceptional cases, in order to assign physiology and pathology each to its proper domain. Next, it is in relation to the normal type that a fact must be found useful or necessary in order to be itself termed normal. Otherwise it could be demonstrated that sickness and health are indistinguishable, since the former necessarily derives from the organism suffering from it. It is only with the average organism that sickness does not sustain the same relatibnship. In the same way the application of a remedy, since it is useful to the sick organism, might pass for a normal phenomenon, although it is plainly abnormal, since only in abnormal circumstances does it possess this utility. This method can therefore only be used if the normal type has previously been constituted, which could only have occurred using a different procedure. Finally, and above all, if it is true that everything which is normal is useful without being necessary, it is untrue that everything which is useful is normal. We can indeed be certain that those states which have become generalised in the species are more useful than those which have continued to be exceptional. We cannot, however, be certain that they are the most useful that exist or can exist. We have no grounds for believing that all the possible combinations have been tried out in the course of the process; among those which have never been realised but are conceivable, there are perhaps some which are much more advantageous than those known to us. The notion of utility goes beyond that of the normal, and is to the normal what the genus is to the species. But it is impossible to deduce the greater from the lesser, the species from the genus, although we may discover the genus from the species, since it is contained within it. This is why, once the general nature of the phenomena has been ascertained, we
Rules for the Distinction of the Normal from the Pathological 97
may confirm the results of the first method by demonstrating how it is useful.9 We can then formulate the three following rules:
(1) A social fact is normal for a given social type, viewed at a given phase of its development, when it occurs in the average society of that species, considered at the corresponding phase of its evolution.
(2) The results of the preceding method can be verified by demonstrating that the general character of the phenomenon is related to the general conditions of collective life in the social type under consideration.
(3) This verification is necessary when this fact relates to a social species which has not yet gone through its complete evolution.
Ill
We are so accustomed to resolving glibly these difficult questions and to deciding rapidly, after cursory observation and by dint of syllogisms, whether a social fact is normal or not, that this procedure will perhaps be adjudged uselessly complicated. It seems unnecessary to have to go to such lengths to distinguish sickness from health. Do we not make these distinctions every day? This is true, but it remains to be seen whether we make them appositely. The difficulty of these problems is concealed because we see the biologist resolve them with comparative ease. Yet we forget that it is much easier for him than for the sociologist to see how each phenomenon affects the strength of the organism and thereby to determine its normal or abnormal charcter with an accuracy which is adequate for all practical purposes. In sociology the complexity and the much more changing nature of the facts constrain us to take many more precautions, as is proved by the conflicting judgements on the same phenomenon emitted by the different parties concerned. To show clearly how great this circumspection must be, we shall illustrate by a few examples to what errors we are exposed when we do not constrain ourselves in this way and in how different a light the most vital phenomena appear when they are dealt with methodically.
If there is a fact whose pathological nature appears indisputable, it is crime. All criminologists agree on this scdre. Although they explain this pathology differently, they none the less unanimously
98   The Rules of Sociological Method
acknowledge it. However, the problem needs to be treated less summarily.
Let us in fact apply the rules previously laid down. Crime is not only observed in most societies of a particular species, but in all societies of all types. There is not one in which criminality does not exist, although it changes in form and the actions which are termed criminal are not everywhere the same. Yet everywhere and always there have been men who have conducted themselves in such a way as to bring down punishment upon their heads. If at least, as societies pass from lower to higher types, the crime rate (the relationship between the annual crime figures and population figures) tended to fall, we might believe that, although still remaining a normal phenomenon, crime tended to lose that character of normality. Yet there is no single ground for believing such a regression to be real. Many facts would rather seem to point to the existence of a movement in the opposite direction. From the beginning of the century statistics provide us with a means of following the progression' of criminality. It has everywhere increased, and in France the increase is of the order of 300 per cent. Thus there is no phenomenon which represents more incontrovert-ibly all the symptoms of normality, since it appears to be closely bound up with the conditions of all collective life. To make crime a social illness would be to concede that sickness is not something accidental, but on the contrary derives in certain cases from the fundamental constitution of the living creature. This would be to erase any distinction between the physiological and the pathological. It can certainly happen that crime itself has normal forms; this is what happens, for instance, when it reaches an excessively high level. There is no doubt that this excessiveness is pathological in nature. What is normal is simply that criminality exists, provided that for each social type it does not reach or go beyond a certain level which it is perhaps not impossible to fix in conformity with the previous rules.10
We are faced with a conclusion which is apparently somewhat paradoxical. Let us make no mistake: to classify crime among the phenomena of normal sociology is not merely to declare that it is an inevitable though regrettable phenomenon arising from the incorrigible wickedness of men; it is to assert that it is a factor in public health, an integrative element in any healthy society. At first sight this result is so surprising that it disconcerted even
Rules for the Distinction of the Normal from the Pathological 99
ourselves for a long time. However, once that first impression of surprise has been overcome it is not difficult to discover reasons to explain this normality and at the same time to confirm it.
In the first place, crime is normal because it is completely impossible for any society entirely free of it to exist.
Crime, as we have shown elsewhere, consists of an action which offends certain collective feelings which are especially strong and clear-cut. In any society, for actions regarded as criminal to cease, the feelings that they offend would need to be found in each individual consciousness without exception and in the degree of strength requisite to counteract the opposing feelings. Even supposing that this condition could effectively be fulfilled, crime would not thereby disappear; it would merely change in form, for the very cause which made the well-springs of criminality to dry up would immediately open up new ones.
Indeed, for the collective feelings, which the penal law of a people at a particular moment in its history protects, to penetrate individual consciousnesses that had hitherto remained closed to them, or to assume, greater authority - whereas previously they had not possessed enough - they would have to acquire an intensity greater than they had had up to then. The community as a whole must feel them more" keenly, for they cannot draw from any other source the additional force which enables them to bear down upon individuals who formerly were the most refractory. For murderers to disappear, the horror of bloodshed must increase in those strata of society from which murderers are recruited; but for this to happen the abhorrence must increase throughout society. Moreover, the very absence of crime would contribute directly to bringing about that result, for a sentiment appears much more respectable when it is always and uniformly respected. But we overlook the fact that these strong states of the common consciousness cannot be reinforced in this way without the weaker states, the violation of which previously gave rise to mere breaches of convention, being reinforced at the same time, for the weaker states are no more than the extension and attenuated form of the stronger ones. Thus, for example, theft and mere misappropriation of property offend the same altruistic sentiment, the respect for other people's possessions. However, this sentiment is offended less strongly by the latter- action than the former. Moreover, since the average consciousness does not have suffi-
100   The Rules of Sociological Method
cient intensity of feeling to feel strongly about the lesser of these two offences, the latter is the object of greater tolerance. This is why the misappropriator is merely censured, while the thief is punished. But if this sentiment grows stronger, to such a degree that it extinguishes in the consciousness the tendency to theft that men possess, they will become more sensitive to these minor offences, which up to then had had only a marginal effect upon them. They will react with greater intensity against these lesser faults, which will become the object of severer condemnation, so that, from the mere moral errors that they were, some will pass into the category of crimes. For example, dishonest contracts or those fulfilled dishonestly, which only incur public censure or civil redress, will become crimes. Imagine a community of saints in an exemplary and perfect monastery. In it crime as such will be unknown, but faults that appear venial to the ordinary person will arouse the same scandal as does, normal crime in ordinary consciences. If therefore that community has the power to judge and punish, it will term such acts criminal and deal with them as such. It is for the same reason that the completely honourable man judges his slightest moral failings with a severity that the mass of people reserves for acts that are truly criminal. In former times acts of violence against the person were more frequent than they are today because respect for individual dignity was weaker. As it has increased, such crimes have become less frequent, but many acts which offended against that sentiment have been incorporated into the penal code, which did not previously include them.11
In order to exhaust all the logically possible hypotheses, it will perhaps be asked why this unanimity should not cover all collective sentiments without exception, and why even the weakest sentiments should not evoke sufficient power to forestall any dissentient voice. The moral conscience of society would be found in its entirety in every individual, endowed with sufficient force to prevent the commission of any act offending against it, whether purely conventional failings or crimes. But such universal and absolute uniformity is utterly impossible, for the immediate physical environment in which each one of us is placed, our hereditary antecedents, the social influences upon which we depend, vary from one individual to another and consequently cause a diversity of consciences. It is impossible for everyone to be alike in this matter, by virtue of the fact that we each have our own
Rules for the Distinction of the Normal from the Pathological 101
organic constitution and occupy different areas in space. This is why, even among lower peoples where individual originality is very little developed, such originality does however exist. Thus, since there cannot be a society in which individuals do not diverge to some extent from the collective type, it is also inevitable that among these deviations some assume a criminal character. What confers upon them this character is not the intrinsic importance of the acts but the importance which the common consciousness ascribes to them. Thus if the latter is stronger and possesses sufficient authority to make these divergences very weak in absolute terms, it will also be more sensitive and exacting. By reacting against the slightest deviations with an energy which it elsewhere employs against those what are more weighty, it endues them with the same gravity and will brand them as criminal.
Thus crime is necessary. It is linked to the basic conditions of social life, but on this very account is useful, for the conditions to which it is bound are themselves indispensable to the normal evolution of morality and law.
Indeed today we can no longer dispute the fact that not only do law and morality vary from one social type to another, but they even change within the same type if the conditions of collective existence are modified. Yet for these transformations to be made possible, the collective sentiments at the basis of morality should not prove unyielding to change, and consequently should be only moderately intense. If they were too strong, they would no longer be malleable. Any arrangement is indeed an obstacle to a new arrangement; this is even more the case the more deep-seated the original arrangement. The more strongly a structure is articulated, the more it resists modification; this is as true for functional as for anatomical patterns. If there were no crimes, this condition would not be fulfilled, for such a hypothesis presumes that collective sentiments would have attained a degree of intensity unparalleled in history. Nothing is good indefinitely and without limits. The authority which the moral consciousness enjoys must not be excessive, for otherwise no one would dare to attack it and it would petrify too easily into an immutable form. For it to evolve, individual originality must be allowed to manifest itself. But so that the originality of the idealist who dreams of transcending his era may display itself, that of the criminal, which falls short of the age, must also be possible. One does not go without the other.,
102   The Rules of Sociological Method
Nor is this all. Beyond this indirect utility, crime itself may play a useful part in this evolution. Not only does it imply that the way to necessary changes remains open, but in certain cases it also directly prepares for these changes. Where crime exists, collective sentiments are not only in the state of plasticity necessary to assume a new form, but sometimes it even contributes to determining beforehand the shape they will take on. Indeed, how often is it only an anticipation of the morality to come, a progression towards what will be! According to Athenian law, Socrates was a criminal and his condemnation was entirely just. However, his crime - his independence of thought - was useful not only for humanity but for his country. It served to prepare a way for a new morality and a new faith, which the Athenians then needed because the traditions by which they had hitherto lived no longer corresponded to the conditions of their existence. Socrates's case is not an isloated one, for it recurs periodically in history. The freedom of thought that we at present enjoy could never have been asserted if the rules that forbade it had hot been violated before they were solemnly abrogated. However, at the time the violation was a crime, since it was an offence against sentiments still keenly felt in the average consciousness. Yet this crime was useful since it was the prelude to changes> which were daily becoming more necessary, Liberal philosophy has had as its precursors heretics of all kinds whom the secular arm rightly punished throught the Middle Ages and has continued to do so almost up to the present day.
From this viewpoint the fundamental facts of criminology appear to us in an entirely new light. Contrary to current ideas, the criminal no longer appears as an utterly unsociable creature, a sort of parasitic element, a foreign, unassimilable body introduced into the bosom of society.12 He plays a normal role in social life. For its part, crime must no longer be conceived of as an evil which cannot be circumscribed closely enough. Far from there being cause for congratulation when it drops too noticeably below the normal level, this apparent progress assuredly coincides with and is linked to some social disturbance. Thus the number of crimes of assault never falls so low as it does in times of scarcity.13 Consequently, at the same time, and as a reaction, the theory of punishment is revised, or rather should be revised. If in fact crime is a sickness, punishment is the cure for it and cannot be conceived of otherwise;
Rules for the Distinction of the Normal from the Patholbgical 103
thus all the discussion aFoused revolves round knowing what punishment should be to fulfil its role as a remedy. But if crime is in no way pathological, the object of punishment cannot be to cure it and its true function must be sought elsewhere.
Thus the rules previously enunciated are far from having as their sole reason to satisfy a logical formalism which lacks any great utility. This is because, on the contrary, according to whether they are applied or not, the most essential social facts totally change their character. If the example quoted is particularly cogent - and this is why we thought we should dwell upon it - there are nevertheless many others which could usefully be cited. There is no society where it is not the rule that the punishment should fit the crime - and yet for the Italian school of thought this principle is a mere invention of legal theoreticians devoid of any solid basis.14 For these criminologists the whole institution of punishment, as it has functioned up to the present among all known peoples, is a phenomenon which goes against nature. We have already seen that for Garofalo the criminality peculiar to the lower forms of society has nothing natural about it. For the socialists it is capitalist organisation, despite its widespread nature, which constitutes a deviation from the normal state and is an organisation brought about by violence and trickery. On the other hand for Spencer it is our administrative centralisation and the extension of governmental power which are the radical vices of our societies, in spite of the fact that both have developed entirely regularly and universally over the course of history. The belief is that one is never obliged systematically to decide on the normal or abnormal character of social facts according to their degree of generality. It is always by a great display of dialectic that these questions are resolved.
However, by laying this criterion on one side, not only is one exposed to confusion and partial errors like those just discussed, but science itself becomes impossible. Indeed its immediate object is the study of the normal type, but if the most general facts can be pathological, it may well be that the normal type has never really existed. Hence what use is it to study facts? They can only confirm bur prejudices and root us more deeply in our errors, since they spring from them. If punishment and responsibility, as they exist in history, are merely a product of ignorance and barbarism, what use is it to strive to know them in order to determine their normal forms? Thus the mind is led to turn away from a reality which from
104  The Rules of Sociological Method
then on lacks interest for us, turning in upon itself to seek the materials necessary to reconstruct that reality. For sociology to deal with facts as things, the sociologist must feel a need to learn from them. The principal purpose of any science of life, whether individual or social, is in the end to define and explain the normal state and distinguish it from the abnormal; If normality does not inhere in the things themselves, if on the contrary it is a characteristic which we impose upon them externally or, for whatever reason, refuse to do so, this salutary state of dependence on things is lost. The mind complacently faces a reality that has not much to teach it. It is no longer contained by the subject matter to which it applies itself, since in some respects it determines that subject matter. The different rules that we have established up to now are therefore closely linked. For sociology really to be a science of things, the generality of phenomena must be taken as the criterion of their normality.
Moreover, our method has the advantage of regulating action at the same time as thought. If what is deemed desirable is not the object of observation, but can and must be determined by some sort of mental calculus, no limit, in a manner of speaking, can be laid down to the free inventions of the imagination in their search for the best. For how can one assign to perfection bounds that it cannot exceed? By definition it escapes all limitations. The goal of humanity thus recedes to infinity, discouraging not a few by its very remoteness, arousing and exciting others, on the other hand, who, so as to draw a little nearer to it, hasten their steps and throw themselves into revolutionary activity. This practical dilemma is avoided if what is desirable is declared to be what is healthy, and if the state of health is something definite, inherent in things, for at the same time the extent of our effort is given and defined. There is no longer need to pursue desperately an end which recedes as we move forward; we need only to work steadily and persistently to maintain the normal state, to re-establish it if it is disturbed, and to rediscover the conditions of normality if they happen to change. The duty of the statesman is no longer to propel societies violently towards an ideal which appears attractive to him. His role is rather that of the doctor: he forestalls the outbreak of sickness by maintaining good hygiene, or when it does break out, seeks to cure it.15
Rules for the Distinction of the Normal from the Pathological   105
Notes
1. Through this we can distinguish the case of sickness from monstrosity. The second is an exception only in space; it is not met with in the average member of the species, but it lasts the whole lifetime of the individuals in which it is to be found. Yet it is clear that these two orders of facts differ only in degree and basically are of the same nature. The boundaries drawn between them are very imprecise, for sickness can also have a lasting character and abnormality can evolve. Thus in defining them we can hardly separate them rigidly. The distinction between them cannot be more categorical than that between the morphological and the physiological, since after all morbidity is abnormal in the physiological order just as monstrosity is in the anatomical order.
2. For example, the savage who had the reduced digestive tube and developed nervous system of the civilised healthy being would be considered sick in relationship to his environment.
3. This section of our argument is abridged, for we can only reiterate here regarding social facts in general what we have said elsewhere concerning the division of moral facts into the normal and abnormal. (Gf. Division du travail social, pp.33-9.)
4. It is true that Garofalo has attempted to distinguish the sick from the abnormal (Criminologie, pp.109, 110). But the sole two arguments on which he relies to make this distinction are:
(1) The word 'sickness' always signifies something which tends to the total or partial destruction of the organism. If there is not destruction, there is a cure, but never stability, such as exists in several abnormalities. But we have just seen that the abnormal is also, in the average case, a threat to the living creature. It is true that this is not always so, but the dangers that sickness entails likewise exist only in average circumstances. As for the absence of stability allegedly distinctive of the morbid, this leaves out of account chronic illnesses and is to divide the study of monstrosities from that of the pathological. The monstrosities are permanent.
(2) It is stated that the normal and abnormal vary according to different races, while the distinction between the physiological and the pathological is valid for all the human race. On the contrary, we have shown that what is morbid for the savage is not so for the civilised person. The conditions of physical health vary according to different environments.
5. It is true that one may speculate whether, when a phenomenon derives necessarily from the general conditions of life, this very fact does not make it useful. We cannot deal with this philosophical question, although we touch upon it a little later.
6. Cf. on this point a note we published in the Revue philosophique (November 1893) on 'La definition du socialisme'.
106   The Rules of Sociological Method
7. Segmentary societies, particularly those which have a territorial basis, are ones whose essential components correspond to territorial divisions (cf. Division du travail social, pp. 189-210).
8. In certain cases one may proceed somewhat differently and demonstrate whether a fact whose normal character is suspect justifies this suspicion by showing whether it is closely linked to the previous development of the social type under consideration, and even to the totality of social evolution in general; or on the other hand whether it contradicts both. By this means we have been able to show that the present weakening of religious beliefs and, more generally, of collective sentiments towards collective objects, is utterly normal; we have proved that such weakening becomes increasingly marked as societies evolve towards our present type, and that this type, in turn, is more developed (cf. Division du travail social pp.73 - 182). But basically this method is only a special case of the preceding one. For if the normality of the phenomenon has been established in this way, it is because at the same time it has been linked to the most general conditions of our collective existence. Indeed, on the one hand, if this regression of religious consciousness is more apparent as the structure of our societies becomes more precisely determinate, it is because it does not depend on any accidental cause but on the very constitution of our social environment. Moreover, on the other hand, since the special characteristics of that constitution are certainly more developed today than formerly, it is entirely normal that the phenomena that depend upon it should themselves be more developed. This method differs only from the preceding one in that the conditions which explain and justify the general character of the phenomenon have been induced' and not observed directly. We know that the phenomenon relates to the nature of the social environment without knowing by what, or how, it is connected.
9. But then it will be said that the realisation of the normal type is not the highest objective that can be proposed and, in order to go beyond it, one must also go beyond the bounds of science. We need not deal with this question here ex professo; let us merely reply: (1) that the question is purely theoretical because in fact the normal type, a state of health, is already somewhat difficult to determine and rarely enough attained for us to exercise our imagination to discover something better; (2) that these improvements, objectively more advantageous, are not for that reason objectively desirable. For if they do not correspond to any latent or actual tendency they would add nothing to happiness and, if they do correspond to some tendency, it is because the normal type has not been realised; (3) finally, that, in order to improve the normal type, it must first be known. One cannot therefore in any case go beyond science except by first re lying upon it.
10. From the fact that crime is a phenomenon of normal sociology it does not follow that the criminal is a person normally constituted from the biological and psychological viewpoints. The two questions
Rules for the Distinction of the Normal from the Pathological 107
are independent of each other. This independence will be better understood when we have shown later the difference which exists between psychical and sociological facts.
11. Calumny, insults, slander, deception, etc.
12. We have ourselves committed the error of speaking of the criminal in this way through not having applied our rule (cf. Division du travail social, pp.395, 396).
13. But, although crime is a fact of normal sociology, it does not follow that we should not abhor it. Pain has likewise nothing desirable about it: the individual detests it just as society detests crime, and yet it is a normal physiological function. Not only does it necessarily derive from the very constitution of every living creature, but it plays a useful and irreplaceable role in life. Thus it would be a peculiar distortion to represent our thinking as an apologia for crime. We would not even have envisaged protesting against such an interpretation were we not aware of the strange accusations and misunderstandings to which one is exposed in undertaking to study moral facts objectively and to speak of them in language that is not commonly used.
14. Cf. Garofalo, Criminologie, pp.299.
15. From the theory developed in this chapter it has sometimes been concluded that, in our view, the upward trend in criminality during the nineteenth century was a normal phenomenon. Nothing is farther from our thoughts. Several facts which we have pointed out in connexion with suicide (cf. le Suicide, p.420ff.) tend, on the contrary, to cause us to believe that this development has been, in general, pathological. However, it may be that a certain increase in certain forms of criminality would be normal, for every state of civilisation has its own criminality. But on this matter one can only hypothesise.