Durkheim Emile, The Rules of Sociological Method, The Free Press, 1938, pages 47-75. I have moved the footnotes to the end of the document.





Observation conducted according to the preceding rules covers two types of facts which are very dissimilar in certain respects: those which conform to given standards and those which "ought" to be different-in other words, normal and pathological phenomena. We have seen that it is necessary to include them both in the definition with which all research must begin. But if their nature is in certain respects identical, they constitute, nevertheless, two different varieties of facts, which need to be distinguished. Can science make this distinction?

The question is of the greatest importance, for on its solution depends the role assigned to science, and especially to the science of man. According to a theory whose partisans belong to most diverse schools, science can teach us nothing about what we ought to desire. It is concerned, they say, only with facts which all have the same value and interest for us; it observes and explains, but does not judge them. Good and evil do not exist for science. It can, indeed, tell us how given causes produce their effects, but not what ends should be pursued. In order to determine not what is but what is desirable, we need to resort to the unconscious, by whatever name it may be designated: "feeling," "instinct," "vital urge," etc. Science, says a writer already quoted, can indeed illuminate the world, but it leaves darkness in our hearts; the heart must find its own light. Science thus loses all, or almost all, practical effectiveness and, consequently, its principal justification for existence. Why strive for knowledge of reality if this knowledge cannot serve us in life? To this we can make reply that, by revealing the causes of phenomena, science furnishes the means of producing them. Every means is from another point of view, an end. In order to put it into operation, it must be willed quite as much as the end whose realization it prepares. There are always several routes that lead to a given goal; a choice must therefore be made between them. If science cannot indicate the best goal to us, how can it inform us about the best means to reach it? Why should it recommend the most rapid in preference to the most economical, the surest rather than the simplest, or vice versa? If science cannot guide us in the determination of ultimate ends, it is equally powerless in the case of those secondary and subordinate ends called "means."

It is true that the ideological method offers an escape from this mysticism, and it is perhaps this circumstance which has made for the persistence of this method. Its adherents were, indeed, too much inclined to rationalism to deny that human conduct needed to be directed by reflection; but, so long as they considered phenomena independently of all subjective data, they could not find a criterion which would permit their classification according to practical value. It seemed, then, that the only means of judging them in this respect was to relate them to some master-concept; and from this argument the use of a concept governing the collation of facts, instead of being derived from them, became indispensable to all rational sociology. We must realize, however, that if, under these conditions, applied science becomes reflective, reflection thus employed is certainly not scientific.

It is yet possible for us to vindicate the legitimate rights of reason in the solution of the problem just stated, without reverting to ideology. Briefly, for societies as for individuals, health is good and desirable; disease, on the contrary, is bad and to be avoided. If, then, we can find an objective criterion, inherent in the facts themselves, which enables us to distinguish scientifically between health and morbidity in the various orders of social phenomena, science will be in a position to throw light on practical problems and still remain faithful to its own method.

Since it cannot at present arrive at formulations concerning the individual, it can give us general indications, which may in turn be appropriately modified by direct contact with the individual through sensation. The state of health, as defined by science, cannot fit exactly any individual subject, since it can be established only with relation to average circumstances, from which everyone deviates more or less; nevertheless, it may serve as a valuable point of reference for regulating our conduct. The necessity for adjusting this standard to each individual case does not mean that knowledge of it is useless. Quite the contrary, it constitutes the norm on which all our practical reasoning has to be based. Under these conditions we are no longer justified in saying that thought is of no use to action. There is no longer a gulf between science and art; but on the contrary, there is break of continuity between them. Science, it is true, can only attain facts through art, but art is merely the prolongation of science. Moreover, may it not be assumed that the practical insufficiency of the latter will diminish when the laws established by it express more and more completely individual reality?


Pain is commonly regarded by the layman as the index of morbidity; and in general it is true that there is a relation between these two conditions, but a relation which lacks uniformity and precision. There are serious but painless maladies; while less serious afflictions, such as those resulting from a speck of coal dust in the eye, may cause real torture. In certain cases the very absence of pain, or even actual pleasure, are symptoms of morbidity. There is an insensibility to pain which is pathological. Circumstances causing suffering to a healthy man may give to a neurasthenic a sensation of enjoyment of an incontestably morbid nature. Conversely, pain accompanies many states belonging to normal physiology, such as hunger, fatigue, and parturition.

Shall we say that health, consisting in successful development of the vital forces, is recognizable by the perfect adaptation of the organism to its environment; and shall we, on the contrary, term "morbidity" whatever disturbs this adaptation? But first-and we shall return to this point later-it has by no means been proved that every state of the organism corresponds to some external state of the environment. And, further, even if this criterion of adaptation were truly distinctive of the state of health, another criterion would be needed in order to recognize it. We must be able to distinguish varying degrees of completeness of adaptation.

Or shall we take as this criterion the effect health and morbidity may have on our probabilities of survival? Health would then be the state of an organism in which these probabilities are at a maximum; and morbidity, on the contrary, would include everything which reduces them. Unquestionably, morbidity weakens the organism. But it is not alone in producing this result. The functions of reproduction inevitably cause death in certain lower species, and they are accompanied by risks even in the higher orders. They are, however, normal. Senility and infancy have the same effects, for both the old and the very young are peculiarly susceptible to the causes of destruction. Are infancy and old age morbid types then? And can only the adult be healthy? How strangely would the domain of health and physiology then be restricted!

If, moreover, old age is already synonymous with morbidity, how distinguish the healthy from the diseased old person? From the same point of view, one is obliged to place menstruation among the morbid phenomena, for the disturbances it causes increase female susceptibility to disease. But would we then be justified in designating as "morbid" a state whose absence or premature disappearance constitutes an incontestably pathological phenomenon? People argue about this question as if, in a healthy organism, each element played a useful role, as if each internal state corresponded exactly to some external condition and, consequently, helped to maintain vital equilibrium and to diminish the chance of death. But it is, on the contrary, legitimate to suppose that some anatomical or functional arrangements are of no direct use , but are merely the products of the general conditions of life. We cannot, however, call them morbid, for morbidity is, above all, something escapable, something not essential to the constitution of the organism. It may even be true that, instead of strengthening the organism, these anatomical and functional arrangements diminish its resistance and, consequently, increase the risks of death.

Moreover, it is not certain that morbidity always has the consequences by which it is proposed to define it. Are there not many ailments too trivial to have an appreciable effect on the fundamental health of the organism? Even among the more serious, there are some whose consequences are not at all harmful, if we know how to combat them. The dyspeptic who follows a good hygienic routine can live to be as old as the healthy man. He is, no doubt, compelled to take certain precautions; but is this not true of us all? Can life be maintained otherwise? Each one of us has his appropriate regimen; that of the sick man does not resemble that practiced by the man who is the average of his time and environment; but this is the only difference between them. Disease does not always leave us helpless or in a state of irremediable maladaptation; it only constrains us to adapt ourselves differently from most of our fellows. There may even exist maladies which will eventually prove useful. The smallpox with which we inoculate ourselves by vaccine is a real disease that we voluntarily undergo, and yet it increases our chances of survival. There are perhaps many other cases in which the disturbance caused by the malady is insignificant compared with the immunities it confers.

Finally and above all, this criterion is most often practically inapplicable. The most we can do is to establish that the lowest mortality-rate known occurs in a specified group of individuals; the impossibility of still lower rates cannot be proved. How do we know that other conditions would not reduce the mortality-rate still further? This actual minimum is not, then, a proof of perfect adaptation, nor, therefore, a sure index of the state of health, if we adhere to the preceding definition. Further, a group exhibiting this characteristic would be very difficult to establish and to isolate yet this would be necessary in order to discover the organic constitution peculiar to it and the supposed cause of its superiority. And while, in the case of a disease which usually terminates fatally, it is evident that the probability of survival of the individual in question is diminished, it is singularly difficult to use that kind of proof in the case of affliction not of the kind to bring on death directly. There is, indeed, only one objective way of proving that individuals have fewer chances of survival in certain conditions than in others, and that is simply to show that, actually, most of them do not live as long. Now, if in the case of purely individual pathology this demonstration is often possible, it is entirely impracticable in social pathology. For here the data at the disposal of the biologist, namely, the average rate of mortality, is wanting. We cannot determine approximately the moments of the birth and death of a society. All these problems, which are far from being solved in biology, are still more mysterious to the sociologist. Moreover, the events occurring in the course of social life and repeating themselves almost identically in all societies of the same type are much too varied for anyone to be able to determine in what measure one of them has contributed to hasten the end. Since individuals are very numerous, we can exercise a selection and compare only those who have a single irregularity in common; this factor is then isolated from all concomitant factors. It is thereby possible to study its influence on the organism. If, for example, a thousand rheumatic patients, chosen at random, present a mortality appreciably higher than the average, there is good reason for attributing this result to that disease. But in sociology, as each social species contains but a small number of individual groups, the field of possible comparisons is too restricted to make demonstrations of this kind valid.

Now, in the absence of such empirical proof, only deductive reasonings are possible, whose conclusions can have no other than presumptive value. We then prove not that a given event actually does weaken the social organism, but that it "must" logically have that effect. In order to prove this, we affirm that it entails a certain consequence judged as harmful to society; and for this reason it is called morbid. But even granted that it actually produces this consequence, it is possible that this disturbance is compensated, and even overcompensated, for by advantages that are overlooked. [Further, there is only one possible reason for characterizing the aforementioned consequence as harmful, namely, that it disturbs the normal functioning of society. Such a proof, however, presupposes that the problem has already been solved; for such could be the case only if one has determined in advance what is normal, together with the distinctive criteria by which normality can be identified. It is unnecessary to discuss here the value of a wholly aprioristic construction of the concept of normality. Thus in sociology, as in history, the same events may be characterized, according to the personal sentiments of the scholar, either as beneficial or disastrous. Irreligious writers describe as abnormal the remnants of faith surviving in the midst of the general upheaval of religious beliefs; while to the religious individual, unbelief (agnosticism) is the great social malady of today. Similarly, to the socialist the present economic organization is a monstrosity, while to the orthodox economist it is the socialist tendencies that are pathological par excellence. And to support its opinions each party finds reasonings which it considers well argued.

The common flaw in these definitions is their premature attempt to grasp the essence of phenomena. They presuppose propositions which, true or not, can be proved only at a more advanced stage of science. This is just the case where we should apply the rule previously established. Instead of aspiring to determine at the outset the relations of the normal and the morbid to vital forces, let us simply seek some external. and perceptible characteristic which will enable us merely to distinguish these two orders of facts.

All sociological phenomena (as well as all biological phenomena) can assume different forms in different cases while still conserving their essential characteristics. We can distinguish two kinds of such forms. Some are distributed in the entire range of the species; they are to be found, if not in all individuals, at least in the majority of them. If they are not found to be identical in all the cases in question, but vary in different persons, these variations do occur within narrow limits. There are, of course, exceptional variations besides these; but these are, first, to be found only in the minority of cases; and, secondly, where they do occur, they most often do not persist throughout the life of the individual. They are an exception both in time and in space., Here are, then, two distinct varieties of phenomena to which we ought to assign different terms. We shall call "normal" these social conditions that are the most generally distributed, and the others "morbid" or "pathological." If we designate as "average type" that hypothetical being that is constructed by assembling in the same individual, the most frequent forms, on ' e may say that the normal type merges with the average type, and that every deviation from this standard of health is a morbid phenomenon. It is true that the average type cannot be determined with the same distinctness as an individual type, since its constituent attributes are not absolutely fixed but are likely to vary. But the possibility of its constitution is beyond doubt, since, blending as it does with the generic type, it is the immediate subject matter of science. It is the functions of the average organism that the physiologist studies; and the sociologist does the same. Once we know how to distinguish the various social species one from the other-a problem which will be treated below-it is always possible to find the most general form of a phenomenon in a given species.

It is clear that a condition can be defined as pathological only in relation to a given species. The conditions of health and morbidity cannot be defined in the abstract and absolutely. This rule is not denied in biology; it has never occurred to anyone to assume that what is normal for a mollusk is normal also for a vertebrate. Each species has a health of its own, because it has an average type of its own. Hence, there exists a state of health for the lowest species as well as for the highest. The same principle applies to sociology, although it is often misunderstood here. One should completely abandon the still too widespread habit of judging an institution, a practice or a moral standard as if it were good or bad in and by itself, for all social types indiscriminately.

Since the point of reference for judging health or morbidity varies with the species, it varies also for a single species, if this species itself changes. Thus, from the purely biological point of view, what is normal for the savage is not normal for the civilized man, and vice versa. There are, especially the variations depending on age, which are important because they occur regularly in all species. The health of the aged person is not that of the adult, and, similarly, the health of the latter is not that of the child; the same is true of societies.3 A social fact can, then, be called normal for a given social species only in relation to a given phase of its development; consequently, to know if it has a right to this appellation, it is not enough to observe the form it takes in the generality of societies belonging to this species; we must also take special care to consider them at the corresponding phase of their evolution.

It may seem that we have simply given a definition of terms, for we have only grouped phenomena according to their resemblances and differences, and given names to the groups thus formed. But, in reality, the concepts thus constituted, while having the great advantage of being recognizable by objective and readily perceptible characteristics, do not differ from the lay conception of health and morbidity. Is not morbidity commonly regarded as an accident which the living organism endures but does not itself ordinarily produce? That is what the ancient philosophers meant when they said that it does not originate from the nature of things-that it is the result of some circumstance in the organism. Such a conception is, surely, a negation of all science; for morbidity is no more miraculous than is health: it is equally grounded in the nature of things. But it is not grounded in their normal nature; it is not inherent in their ordinary constitution or bound up with the conditions of existence upon which they generally depend. Conversely, nobody distinguishes the type of the healthy specimen from the type of the species. One cannot, without contradiction, even conceive of a species which would be incurably diseased in itself and by virtue of its fundamental constitution. The healthy constitutes the norm par excellence and can consequently be in no way abnormal.

Health is also commonly thought of as a state generally preferable to morbidity. But this definition is already implied in the preceding one. There must, indeed, have been some reason by virtue of which the characteristics, which in the aggregate form the normal type, have been able to spread to the entire species. This dissemination is itself a fact in need of explanation, and a cause must therefore be assigned to it. It would be incomprehensible if the most widespread forms of organization would not at the same time be, at least in their aggregate, the most advantageous. How could they have maintained themselves under so great a variety of circumstances if they had not enabled the individual better to resist the elements of destruction? On the other hand, the reason for the rarity of the other characteristics is evidently that the average organism possessing them has greater difficulty in surviving. The greater frequency of the former is, thus, a proof of their superiority.4


This last statement furnishes a means of checking the results of the preceding method.

Since the general distribution of normal phenomena is itself an explainable phenomenon, after it has been definitely established by observation it should be explained. Although it is assumed, of course, that it is not without its cause, science demands that we know exactly what this cause is. The normality of the phenomenon will, indeed, be more certain if it is demonstrated that the external sign, which had at first revealed it, is not purely adventitious but grounded in the nature of things-if, in a word, one can erect this normality of fact into a normality by logical necessity. Furthermore, this demonstration will not always consist in showing that the trait is useful to the organism, although this is most frequently the case; but it can also happen, as we have remarked above, that a situation is normal without being at all useful, simply because it is necessarily implied in the nature of the being. Thus, it would perhaps be desirable if parturition did not occasion such violent disturbances in the female organism, but this is impossible; consequently, the normality of the phenomenon is to be explained by the mere fact that it is bound up with the conditions of existence of the species under consideration, either as a mechanically necessary effect of these conditions or as a means permitting the organisms to adapt thernselves.5

This proof is not simply useful as a check. It must not be forgotten that, if there is any advantage in distinguishing the normal from the abnormal, it is especially helpful in making our practice more intelligent. To act with full knowledge of the facts, we need to know not only the proper procedure but also the reasons for it. Scientific propositions concerning the normal state will be more immediately applicable to individual cases when accompanied by their reasons, for then we shall be better able to recognize in which cases, and in which direction, they should be modified in their application.

There are circumstances in which this verification is absolutely necessary, since the first method, if used alone, might lead to error. This applies to periods of transition, when the entire species is in process of evolution, without having yet become stabilized in its new form. The only normal type that is valid under such circumstances is the type from the previous condition, and yet it no longer corresponds to the new conditions of existence.

A phenomenon can thus persist throughout the entire range of a species although no longer adapted to the requirements of the situation. It is then normal only in appearance. Its universality is now an illusion, since its persistence, due only to the blind force of habit, can no longer be accepted as an index of a close connection with the general conditions of its collective existence. This difficulty is especially peculiar to sociology. It scarcely exists in biology. Very rarely, indeed, are animal species compelled to take on unforeseen forms. The only normal modifications they undergo are those occurring regularly in each individual, principally under the influence of age. Therefore, the norm is easily established, since it can be observed in a great many cases. The normal state can be known at each moment of the development of the animal and even in periods of crisis, and the same is also true in sociology for societies belonging to the lower cultures. Since many of them have already completed their cycle of development, the law of their normal evolution is (or, at least, can be) established. For the highest and most recent societies this law is by definition unknown, since they have not yet accomplished their entire course. Having no point of reference, the sociologist may then be embarrassed in deciding whether a phenomenon is normal or not.

The procedure just indicated will free him from this difficulty. After having established by observation that a particular fact is general, he will go back to the conditions which determined this generality in the past and will then investigate whether these conditions are still given in the present or if, on the contrary, they have changed. In the ,first case he may properly designate the phenomenon as normal; and, in the second, refuse it this designation. For example, in order to determine whether the present economic state of Europe, with the absence of organization' characterizing it, is normal or not, we shall investigate the causes which brought it about. If these conditions still exist in our present-day society, this situation is normal in spite of the dissent it arouses. But if, on the contrary, it is found to be related to the old social structure which we have elsewhere qualified as segmental' and which, after having been the essential framework of societies, progressively disappears, we shall have to conclude that the present situation, however universal, is pathological. By the same method should be settled all controversial questions of this kind, such as those concerning the normality of the decline in religious beliefs or of the development of state powers.

Nevertheless, this method can by no means be substituted for the preceding one, nor even be employed as a first resort. in the first place, it raises questions which we shall treat below and which can be approached only at a rather advanced stage of science, for it implies, on the whole, an almost complete explanation of the phenomena concerned, and assumes that either their causes or their functions have already been determined. Now, it is important, from the very beginning of research, to be able to classify facts as normal and abnormal, save for a few exceptional cases, so that the proper domains can be assigned to physiology and pathology, respectively. Further, it is in relation to the normal type that a fact must be useful or necessary, if it is to be designated as normal itself. Otherwise it could be demonstrated that morbidity is indistinguishable from health, since it is merely an outgrowth of the afflicted organism. In this case, however, it is morbid because it does not maintain its proper relation to the average organism. Similarly, since a remedy is useful to the sick, its application could then be considered as a normal phenomenon. The remedy itself possesses this utility. One can, then, use this further method only after the normal type has been determined by some previous method. Finally and above all, if it is true that all that is normal is useful, without being necessary, it is not true that all that is useful is normal. Certainly the states that have become general in the species are more useful than those that have remained exceptional, but they do not possess the maximum of utility that exists or might possibly be brought about. We have no reason to believe that all possible combinations have been tried out in the course of experience; and, among those conceivable but never realized, there are some that are perhaps much more advantageous than those now known to us. The idea of utility is broader than that of normality; it has to the latter the same relation as the genus to the species. It is impossible to deduce the greater from the less, the genus from the species. But the species, since it is contained in the genus, can be discovered in the latter. That is why, once the generality of the phenomenon has been established, one can, by showing its utility, confirm the results of the first method.9 We may, therefore, formulate the three following rules:

1. A social fact is normal, in relation to a given social type at a given Phase of its development, when it is present in the average society of that species at the corresponding phase of its evolution.

2. One can verify the results of the preceding method by showing that the generality of the phenomenon is bound up with the general conditions of collective life of the social type considered.

3. This verification is necessary when the fact in question occurs in a social species which has not yet reached the full course of its evolution.


The custom of resolving these difficult questions with a pat phrase and of deciding hastily, from superficial observations supported by syllogisms, whether a social fact is normal or not prevails to such an extent that our procedure will perhaps be judged needlessly complicated. It seems unnecessary to go to such lengths in order to distinguish between morbidity and health. It is true that we make these distinctions every day, but it remains to be seen whether we make them correctly. The fact that the biologist solves these problems with relative case obscures in our minds the difficulties they involve. We forget that it is much easier for him than for the sociologist to observe how the resistance of the organism is affected by each phenomenon and to determine thereby its normal or abnormal character with sufficient exactness for practical purposes. In sociology the greater complexity and inconstancy of the facts oblige us to take many more precautions, and this is all too evident in the contradictory judgments on the same phenomenon given by different scholars. In order to show clearly the great necessity for circumspection, we shall illustrate by a few examples the errors resulting from the opposite attitude and show in how different a light the most essential phenomena appear when treated methodically.

If there is any fact whose pathological character appears incontestable, that fact is crime. All criminologists are agreed on this point. Although they explain this pathology differently, they are unanimous in recognizing it. But let us see if this problem does not demand a more extended consideration.

We shall apply the foregoing rules. Crime is present not only in the majority of societies of one particular species but in all societies of all types. There is no society that is not confronted with the problem of criminality. Its form changes; the acts thus characterized are not the same everywhere; but, everywhere and always, there have been men who have behaved in such a way as to draw upon themselves penal repression. If, in proportion as societies pass from the lower to the higher types, the rate of criminality, i.e., the relation between the yearly number of crimes and the population, tended to decline, it might be believed that crime, while still normal, is tending to lose this character of normality. But we have no reason to believe that such a regression is substantiated. Many facts would seem rather to indicate a movement in the opposite direction. From the beginning of the [nineteenth] century, statistics enable us to follow the course of criminality. It has everywhere increased. In France the increase is nearly 300 per cent. There is, then, no phenomenon that presents more indisputably all the symptoms of normality, since it appears closely connected with the conditions of all collective life. To make of crime a form of social morbidity would be to admit that morbidity is not something accidental, but, on the contrary, that in certain cases it grows out of the fundamental constitution of the living organism; it would result in wiping out all distinction between the physiological and the pathological. No doubt it is possible that crime itself will have abnormal forms, as, for example, when its rate is unusually high. This excess is, indeed, undoubtedly morbid in nature. What is normal, Simply, is the existence of criminality, provided that it attains and does not exceed, for each social type, a certain level, which it is perhaps not impossible to fix in conformity with the preceding rules.10

Here we are, then, in the presence of a conclusion in appearance quite paradoxical. Let us make no mistake. To classify crime among the phenomena of normal sociology is not to say merely that it is an inevitable, although regrettable phenomenon, due to the incorrigible wickedness of men; it is to affirm that it is a factor in public health, an integral part of all healthy societies. This result is, at first glance, surprising enough to have puzzled even ourselves for a long time. Once this first surprise has been overcome, however, it is not difficult to find reasons explaining this normality and at the same time confirming it.

In the first place crime is normal because a society exempt from it is utterly impossible. Crime, we have shown elsewhere, consists of an act that offends certain very strong collective sentiments. In a society in which criminal acts are no longer committed, the sentiments they offend would have to be found without exception in all individual consciousness, and they must be found to exist with the same degree as sentiments contrary to them. Assuming that this condition could actually be realized, crime would not thereby disappear; it would only change its form, for the very cause which would thus dry up the sources of criminality would immediately open up new ones.

Indeed, for the collective sentiments which are protected by the penal law of a people at a specified moment of its history to take possession of the public conscience or for them to acquire a stronger hold where they have an insufficient grip, they must acquire an intensity greater than that which they had hitherto had. The community as a whole must experience them more vividly, for it can acquire from no other source the greater force necessary to control these individuals who formerly were the most refractory. For murderers to disappear, the horror of bloodshed must become greater in those social strata from which murderers are recruited; but, first it must become greater throughout the entire society. Moreover, the very absence of crime would directly contribute to produce this horror; because any sentiment seems much more respectable when it is always and uniformly respected.

One easily overlooks the consideration that these strong states of the common consciousness cannot be thus reinforced without reinforcing at the same time the more feeble states, whose violation previously gave birth to mere infraction of convention-since the weaker ones are only the prolongation, the attenuated form, of the stronger. Thus robbery and simple bad taste injure the same single altruistic sentiment, the respect for that which is another's. However, this same sentiment is less grievously offended by bad taste than by robbery; and since, in addition, the average consciousness has not sufficient intensity to react keenly to the bad taste, it is treated with greater tolerance. That is why the person guilty of bad taste is merely blamed, whereas the thief is punished. But, if this sentiment grows stronger, to the point of silencing in all consciousness the inclination which disposes man to steal, he will become more sensitive to the offenses which, until then, touched him but lightly. He will react against them, then, with more energy; they will be the object of greater opprobrium, which will transform certain of them from the simple moral faults that they were and give them the quality of crimes. For example, improper contracts, or contracts improperly executed, which only incur public blame or civil damages, will become offenses in law.

Imagine a society of saints, a perfect cloister of exemplary individuals. Crimes, properly so called, will there be unknown; but faults which appear venial to the layman will create there the same scandal that the ordinary offense does in ordinary consciousness. If, then, this society has the power to judge and punish, it will define these acts as criminal and will treat them as such. For the same reason, the perfect and upright man judges his smallest failings with a severity that the majority reserve for acts more truly in the nature of an offense. Formerly, acts of violence against persons were more frequent than they are today, because respect for individual dignity was less strong. As this has increased, these crimes have become more rare; and also, many acts violating this sentiment have been introduced into the penal law which were not included there in primitive times.

In order to exhaust all the hypotheses logically possible, it will perhaps be asked why this unanimity does not extend to all collective sentiments without exception. Why should not even the most feeble sentiment gather enough energy to prevent all dissent? The moral consciousness of the society would be present in its entirety in all the individuals, with a vitality sufficient to prevent all acts offending it-the purely conventional faults as well as the crimes. But a uniformity so universal and absolute is utterly impossible; for the immediate physical milieu in which each one of us is placed, the hereditary antecedents, and the social influences vary from one individual to the next, and consequently diversify consciousness. It is impossible for all to be alike, if only because each one has his own organism and that these organisms occupy different areas in space. That is why, even among the lower peoples, where individual originality is very little developed, it nevertheless does exist.

Thus, since there cannot be a society in which the individuals do not differ more or less from the collective type, it is also inevitable that, among these divergences, there are some with a criminal character. What confers this character upon them is not the intrinsic quality of a given act but that definition which the collective conscience lends them. If the collective conscience is stronger, if it has enough authority practically to suppress these divergences, it will also be more sensitive, more exacting; and, reacting against the slightest deviations with the energy it otherwise displays only against more considerable infractions, it will attribute to them the same gravity as formerly to crimes. In other words, it will designate them as criminal.

Crime is, then, necessary; it is bound up with the fundamental conditions of all social life, and by that very fact it is useful, because these conditions of which it is a part are themselves indispensable to the normal evolution of morality and law.

Indeed, it is no longer possible today to dispute the fact that law and morality vary from one social type to the next, nor that they change within the same type if the conditions of life are modified. But, in order that these transformations may be possible, the collective sentiments at the basis of morality must not be hostile to change, and consequently must have but moderate energy. If they were too strong, they would no longer be plastic. Every pattern is an obstacle to new patterns, to the extent that the first pattern is inflexible. The better a structure is articulated, the more it offers a healthy resistance to all modification; and this is equally true of functional, as of anatomical, organization. If there were no crimes, this condition could not have been fulfilled; for such a hypothesis presupposes that collective sentiments have arrived at a degree of intensity unexampled in history. Nothing is good indefinitely and to an unlimited extent. The authority which the moral conscience enjoys must not be excessive; otherwise no one would dare criticize it, and it would too easily congeal into an immutable form. To make progress, individual originality must be able to express itself - In order that the originality of the idealist whose dreams transcend his century may find expression, it is necessary that the originality of the criminal, who is below the level of his time, shall also be possible. One does not occur without the other.

Nor is this all. Aside from this indirect utility, it happens that crime itself plays a useful role in this evolution. Crime implies not only that the way remains open to necessary changes but that in certain cases it directly prepares these changes. Where crime exists, collective sentiments are sufficiently flexible to take on a new form, and crime sometimes helps to determine the form they will take. How many times, indeed, it is only an anticipation of future morality -a step toward what will be! According to Athenian law, Socrates was a criminal, and his condemnation was no more than just. However, his crime, namely, the independence of his thought, rendered a service not only to humanity but to his country. It served to prepare a new morality and faith which the Athenians needed, since the traditions by which they had lived until then were no longer in harmony with the current conditions of life. Nor is the case of Socrates unique; it is reproduced periodically in history. It would never have been possible to establish the freedom of thought we now enjoy if the regulations prohibiting it had not been violated before being solemnly abrogated. At that time, however, the violation was a crime, since it was an offense against sentiments still very keen in the average conscience. And yet this crime was useful as a prelude to reforms which daily became more necessary. Liberal philosophy had as its precursors the heretics of all kinds who were justly punished by secular authorities during the entire course of the Middle Ages and until the eve of modern times.

From this point of view the fundamental facts of criminality present themselves to us in an entirely new light. Contrary to current ideas, the criminal no longer seems a totally unsociable being, a sort of parasitic element, a strange and unassimilable body, introduced into the midst of society." On the contrary, he plays a definite role in social life. Crime, for its part, must no longer be conceived as an evil that cannot be too much suppressed. There is no occasion for self-congratulation when the crime rate drops noticeably below the average level, for we may be certain that this apparent progress is associated with some social disorder. Thus, the number of assault cases never falls so low as in times of want. With the drop in the crime rate, and as a reaction to it, comes a revision, or the need of a revision in the theory of punishment. If, indeed, crime is a disease, its punishment is its remedy and cannot be otherwise conceived; thus' all the discussions it arouses bear on he point of determining what the punishment must be in order to fulfil this role of remedy. If crime is not pathological at all, the object of punishment cannot be to cure it, and its true function must be sought elsewhere.

It is far from the truth, then, that the rules previously stated have no other justification than to satisfy an urge for logical formalism of little practical value, since, on the contrary, according as they are or are not applied, the most essential social facts are entirely changed in character. If the foregoing example is particularly convincing-and this was our hope in dwelling upon it-there are likewise many others which might have been cited with equal profit. There is no society where the rule does not exist that the punishment must be proportional to the offense; yet, for the Italian school, this principle is but an invention of jurists, without adequate basis.14

For these criminologists the entire penal system, as it has functioned until the present day among all known peoples, is a phenomenon contrary to nature. We have already seen that, for M. Garofalo, the criminality peculiar to lower societies is not at all natural. For socialists it is the capitalist system, in spite of its wide diffusion, which constitutes a deviation from the normal state, produced, as it was, by violence and fraud. Spencer, on the contrary, maintains that our administrative centralization and the extension of governmental powers are the radical vices of our societies, although both proceed most regularly and generally as we advance in history. We do not believe that scholars have ever systematically endeavored to distinguish the normal or abnormal character of social phenomena from their degree of generality. It is always with a great array of dialectics that these questions are partly resolved.

Once we have eliminated this criterion, however, we are not only exposed to confusions and partial errors, such as those just pointed out, but science is rendered all but impossible. Its immediate object is the study of the normal type. If, however, the most widely diffused facts can be pathological, it is possible that the normal types never existed in actuality; and if that is the case, why study the facts? Such study can only confirm our prejudices and fix us in our errors. If punishment and the responsibility for crime are only the products of ignorance and barbarism, why strive to know them in order to derive the normal forms from them? By such arguments the mind is diverted from a reality in which we have lost interest, and falls back on itself in order to seek within itself the materials necessary to reconstruct its world. In order that sociology may treat facts as things, the sociologist must feel the necessity of studying them exclusively.

The principal object of all sciences of life, whether individual or social, is to define and explain the normal state and to distinguish it from its opposite. If, however, normality is not given in the things themselves-if it is, on the contrary, a character we may or may not impute to them-this solid footing is lost. The mind is then complacent in the face of a reality which has little to teach it; it is no longer restrained by the matter which it is analyzing, since it is the mind, in some manner or other, that determines the matter.

The various principles we have established up to the present are, then, closely interconnected. In order that sociology may be a true science of things, the generality of phenomena must be taken as the criterion of their normality.

Our method has, moreover, the advantage of regulating action at the same time as thought. If the social values are not subjects of observation but can and must be determined by a sort of mental calculus, no limit, so to speak, can be set for the free inventions of the imagination in search of the best. For how may we assign to perfection a limit? It escapes all limitation, by definition. The goal of humanity recedes into infinity, discouraging some by its very remoteness and arousing others who, in order to draw a little nearer to it, quicken the pace and plunge into revolutions. This practical dilemma may be escaped if the desirable is defined in the same way as is health and normality and if health is something that is defined as inherent in things. For then the object of our efforts is both given and defined at the same time. It is no longer a matter of pursuing desperately an objective that retreats as one advances, but of working with steady perseverance to maintain the normal state, of re-establishing it if it is threatened, and of rediscovering its conditions if they have changed. The duty of the statesman is no longer to push society toward an ideal that seems attractive to him, but his role is that of the physician: he prevents the outbreak of illnesses by good hygiene, and he seeks to cure them when they have appeared.15

1 Hereby one can distinguish the morbid case from the monstrosity. The second is an exception in space only; it is not met with in the average of the species, but it persists throughout the life of the individual in which it is found. These two orders of facts differ, however, only in degree, and are fundamentally of the same nature; their boundaries, are very uncertain, for morbidity is capable of permanence and monstrosity of modification. They can, then, scarcely be rigidly separated in their definitions. The distinction between them cannot be more categorical than that between the morphological and the physiological, since, in short, the morbid is the abnormal in the Physiological order as the monstrous is the abnormal in the anatomical order.

2 For example, a savage with the undersized digestive system and the overdeveloped nervous system of the healthy civilized man would be ill in relation to his environment.

3 We cut short this part of our discussion since we can only repeat here, with regard to social facts in general, what we have said elsewhere on the subject of the division of moral facts into normal and abnormal ones. (See Division du travail social, PP. 33-39.)

4 M. Garofalo has tried to distinguish between the morbid and the abnormal (Criminologie, pp. iog, i io). But he bases this distinction exclusively on the following two arguments: (ii) The word "morbidity" always signifies something tending toward the total or partial destruction of the organism; if there is not destruction, there is cure, but never stability, as in some anomalies. But we have just seen that the abnormal is a menace to the living being, on the average. It is true that this is not always the case, but the dangers of morbidity likewise exist only in average cases. And if the absence of stability is taken as the criterion of the morbid, chronic morbidities are overlooked, and the monstrosity is completely separated from the pathological. Monstrosities are fixed. (2) We are told that the normal and the abnormal vary with races, whereas the distinction between the physiological and the pathological holds good f or the entire human race. But we have just shown, On the contrary, that often phenomena which are morbid for the savage are not morbid for the civilized man. The conditions of physical health vary with the group.

5 One may well ask whether the necessary derivation of a phenomenon from the general conditions of life does not imply its utility. We cannot treat at length this philosophical question, which will, however, be touched on below.

6 Cf., on this point, the note we published in the Revue philosophique, November, 1893, on "La D6finition du socialisme."

7 Segmental societies and, more particularly, segmental societies with a territorial basis are those whose essential boundaries correspond to territorial divisions. (See Division du travail social, pp. 189-210.)

8 In certain cases, one can proceed in a slightly different way and prove, in the case of a fact whose normality is doubted, the validity or invalidity of this suspicion by showing its close connection with the previous development of the social type under consideration, and even to social evolution in general. Or one can show the opposite, that it contradicts both. In this way we have been able to prove that the present weakening of religious beliefs, and more generally of collective sentiments, is only normal; we have proved that this weakening becomes more and more accentuated as societies approach our present type and as the latter, in its turn, is more highly developed (ibid., PP. 73-182). But, fundamentally, this method is only a particular case of the preceding one. For this demonstration of the normality of the phenomenon in question implies that we relate it to the most general conditions of our collective existence. Indeed, on the one hand, the positive correlation between this regression of religious consciousness and the degree of articulation of the structure of our societies is based not on some accidental cause but on the very constitution of our social milieu. Since, on the other hand, the traits characteristic of this constitution are certainly more highly developed today than formerly, it is only normal that the phenomena depending on it be themselves more highly developed. This method differs from the preceding one only in one feature, namely, that the conditions explaining and justifying the generality of the phenomenon are inferred and not directly observed. We know that it is connected with the nature of the social milieu without knowing the nature and mode of this connection.

9 But it will be said that the determination of the normal type is not the highest objective possible and that, in order to transcend it, science has to be transcended also. We are avowedly not concerned with this question here; let us answer only: (i) that the question is entirely theoretical, for the normal type, the state of health, is difficult enough to determine and rarely enough attained for us to search our imagination in the attempt to find something better; (2) that these improvements of the normal type, although objectively advantageous, are not objectively desirable for that reason, for, if they do not correspond to a latent or actual tendency, they add nothing to happiness, and, if they do, the normal type is not realized; (3) and finally, that in order to improve the normal type, it must be known. One can, then, in any case, only transcend science by first making the best possible use of it.

10 From the fact that crime is a phenomenon of normal sociology, it does not follow that the criminal is an individual normally constituted from the biological and psychological points of view. The two questions are independent of each other. This independence will be better understood when we have shown, later on, the difference between psychological and sociological facts.

11 Calurnny, insults, slander, fraud, etc.

12 We have ourselves committed the error of speaking thus of the criminal, because of a failure to apply our rule (Division du travail social, PP. 395-96).

13 Although crime is a fact of normal sociology, it does not follow that we must not abhor it. Pain itself has nothing desirable about it; the individual dislikes it as society does crime, and yet it is a function of normal physiology. Not only is it necessarily derived from the very constitution of every living organism, but it plays a useful role in life, for which reason it cannot be replaced. It would, then, be a singular distortion of our thought to present it as an apology for crime. We would not even think of protesting against such an interpretation, did we not know to what strange accusations and misunderstandings one exposes oneself when one undertakes to study moral facts objectively and to speak of them in a different language from that of the layman.

14 See Garofalo, CriMinalogie, P. 299.

15 From the theory developed in this chapter, the conclusion has at times been reached that, according to us, the increase of criminality in the course of the nineteenth century was a normal phenomenon. Nothing is farther from our thought. Several facts indicated by us apropos of suicide (see Suicide, PP- 420 ff.) tend, on the contrary, to make us believe that this development is in general morbid. Nevertheless, it might happen that a certain increase Of certain forms of criminality would be normal, for each state of civilization has its own criminality. But on this, one can only formulate hypotheses.