1895 durkheim_rulessocmethod !

25
THE RULES OF SOCIOLOGICAL METHOD Br EMILE DURKHEIM EIGHTH EDITION, TRANSLATED BY SARAH A. SOLOVAY and JOHN H. MUELLER AND EDITED BY GEORGE E. G. CATLIN @ THE FREE PRESS A Division o~ Macmillan Publishing Co., Inc. Yov.x Collier MaemiUan Publishers LONDON m

Upload: lornap

Post on 14-Oct-2014

85 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: 1895 Durkheim_RulesSocMethod !

THE RULESOF SOCIOLOGICAL

METHODBr EMILE DURKHEIM

EIGHTH EDITION, TRANSLATED BY

SARAH A. SOLOVAY and JOHN H. MUELLER

AND EDITED BY

GEORGE E. G. CATLIN

@THE FREE PRESS

A Division o~ Macmillan Publishing Co., Inc.Yov.x

Collier MaemiUan PublishersLONDON

m

Page 2: 1895 Durkheim_RulesSocMethod !

l:K AUTHOR’S INTRODUCTION

which the principal problems should be formulated, thedirection research should take, the specific methods of workwhich may enable it to reach its conduslons--all these re-mained completely undetermined.

A happy combination of circumstances, among the mostimportant of which may rightly be placed the proposal toestablish a regular course in sociology in the Faculty ofLetters at Bordeaux, enabled us to devote ourselves early tothe study of social science and, indeed, to make it our voca-tion. Therefore, we have been able to abandon these verygeneral questions and to attack a certain number of definiteproblems. The very force of events has thus led us to con-struct a method that is, we believe, more precise and moreexactly adapted to the distinctive characteristics of soda]phenomena. We wish here to expound the results of ourwork in applied sociology .in their entirety and to submitthem for discussion. They axe, of course, contained by im-plication in the book wMch we published recently on theDivisio~ in Social Labor. But it seems to us that it is of someadvantage to make them explicit and to give them separateformulation, accompanying them with proofs and illustra-tions drawn either from that work or from works still un-published. The public will thus be better able to judge ofthe direction we are trying to give to sociological studies.

r.,

¯ .Q:

CHAPTER I

WHAT IS A SOCIAL FACT?

Before inquiring into the method suited to the study ofsocial facts, it is important to know which facts are common-ly called "social." This information is all the more necessarysince the designation "social" is used with little precision.It is currently employed for practically all phenomena gen-erally diffused within society, however small their socialinterest. But on that basis, there axe, as itwere, no humanevents that may not be called social. Each individualdrinks, sleeps, eats, reasons; and it is to society’s interestthat these functions be exercised in an orderly manner. If,then, all these facts axe counted as "sodal" facts, sociologywould have no subject matter exdusively its own, and itsdomain would be confused with that of biology and psy-chology.

But in reality there is in every society a certain group ofphenomena which may-be diiterentiated from those studiedby the other natural sciences. When I fulfil my obligationsas brother, husband, or citizen, when I execute my contracts,I perform duties which are defined, externally to myself andmy acts, in law and in custom. ]Even if they conform to myown sentiments and I feel their reality subjectively, suchreality is still objective, for I did not create them; I merelyinherited them through my education. How many times ithappens, moreove.r, that we are ignorant of the details ofthe obligations incumbent upon us, and that in order toacquaint ourselves with them we must consult the law andits authorized interpreters! Similarly, the church-member

T

¯ . .i r , . ¯

...; >

¯ ...

¯ . .,,i~¸

.:.,.:,.:

."i¸

¯ ...:,;:. ;

",.!i(

".;!

Page 3: 1895 Durkheim_RulesSocMethod !

RULES .OF SOCIOLOGICAL METHOD

finds the beliefs and practices oi his religious life ready-madeat birth; their existence prior to his own implies their ex-istence outside of himself. The system of signs I use to ex-press my thought, the system of currency I employ to paymy debts, the instruments of credit I utilize in my commer-cial relations, the practices followed in my profession, etc.,function independently of my own use of them. And thesestatements can be repeated for each member of society.Here, then, are ways of acting, thinking, and feeling thatpresent the noteworthy property of existing outside the indi-vidual consciousness.

These types of conduct Or thought are not only externalto the individual but are, moreover, endowed with coercivepower, by virtue of which they impose themselves upon him,independent of his individual will. Of course, when I fullyconsent and conform to them, this constraint is felt onlyslightly, i~ at all, and is therefore unnecessary. But it is,nonetheless, an intrinsic characteristic of these facts, theproof thereof being that it asserts itself as soon as I attemptto resist it. If I attempt to violate the law, it reacts againstme so as to prevent my act before its accomplishment, or tonullify my violation by restoring the damage, if it is ac-complished and reparable, or to make me expiate it if itcannot be compensated for otherwise.

In the case of purely moral maxims~ the public conscienceexercises a che~ on every act which offends it by means ofthe surveilIauce it exercises over the conduct of citizens, andthe appropriate penalties at its disposal. In many cases theconstraint is less violent, but nevertheless it always exists.If I do not submit to the conventions of society, if in mydress I do not conform to the customs observed in mycountry and in my class, the ridicule I provoke, the social

:!i!~

¯ WHAT IS A SOCIAL FACT? $

isolation in which I am kept, produce, although in an at-" tenuated form, the same effects as a punishment in the strictsense of the word. The constraint is nonetheless efficaciousfor being indirect. I am not obliged to speak French withmy fellow-countrymen nor to use the legal currency, but Icannot possibly do otherwise. If I tried to escape this neces-sity, my attempt would fail miserably. As an industrialist,I am free to apply the technical methods of former centuries;but by doing so~ I should invite certain ruin. Even when Ifree myself from these rules and violate them successfully,I am always compelled to struggle with them. When finallyovercome, they make their constraining power sufficientlyfelt by the resistance they offer. The enterprises of all in-novators, including successful ones, come up against re-sistance of this kind.

Here, then, is a category of facts with very distinctivecharacteristics: it consists of ways of acting, thinking, andfeeling, external to the individual, and endowed with apower of coercion, by reason of which they control him.These ways of thinking could not be confused with biologicalphenomena, since they consist of representations and ofactions; nor with psychological phenomena, which exist onlyin.the individual consciousness and through it. They con-stitute, thus, anew variety of phenomena; and it is to themexclusively that the term "social" ought to be applied. Andthis term fits them quite well, for it is clear that, since theirsource is not in the individual, their substratum can be noother than society, either the political society as a whole orsome one of the partial groups it includes, such as religiousdenominations, political, literary, and occupational associa-tions, etc. On the other hand, this term "social" applies tothem exclusively, for it has a distinct meanh~g only if it

Page 4: 1895 Durkheim_RulesSocMethod !

4 RULES. OF SOCIOLOGICAL METHOD

designates exclusively the phenomena which arc not in-cluded in any of the categories of facts that have alreadybeen established and classified. These ways of thinking andacting therefore constitute the proper domain of sodology.It is true that, when we define them with this word "con-straint," we risk shocking the zealous partisans of absoluteindividualism. For those who profess the complete autono-my of the individual, man’s dignity is diminished wheneverhe is made to feel that he is not completely self-determinant.It is generally accepted today, however, that most of ourideas and our tendencies are not developed by ourselves butcome to us from without. How can they become a part ofus except by imposing themselves upon us? This is thewhole meaning of our definition. And it is generally ac-cepted, moreover, that soda] constraint is not necessarilyincompatible with the individual personality.~

Since the examples that we have just cited (legal andmoral regulations, religious faiths, financial systems, etc.)all consist of established beliefs and practices, one might beled to believe that social facts exist only where there is somesocial organization. But there are other facts without suchcrystallized form which have the same objectivity and thesame ascendency over the individual. These are called ’%0-cial currents." Thus the great movements of enthusiasm,indignation, and pity in a crowd do not originate in any oneof the particular individual consdousnesses. They come toeach one of us from without and can carry us away in spiteof ourselves. Of course, it may happen that, in abandoningmyself to them unreservedly, I do not feel the pressure theyexert upon me. But it is revealed as soon as I try to resist

x We do not intend to imply, however, tl~t all constraint is normal. We

shall return to tht¢ point l~ter.

-~

WHAT IS A SOCIAL FACT? ¯5

them. Let an individual attempt to oppose one of these col-

lective manifestations, and the emotions that he denies willturn against him. Now, if this power of external coercionasserts itself so dearly in cases of resistance, it must existalso in the first-mentioned cases, although we are. uncon-scious of it. We are then victims of the illusion of havingourselves created that which actually forced itself from with-out. If the complacency with which we permit ourselves tobe carded along conceals the pressure undergone, neverthe-less it does not abolish it. Thus, air is no less heavy becausewe do not detect its weight. So, even if we ourselves havespontaneously contributed to the production of the commonemotion, the impression we have received differs markedlyfrom that which we would hgve experienced if we had beenalone. Also, once the crowd has dispersed, that is, once thesesocial influences have ceased to act upon US and we are aloneagain, the emotions which have passed through the mindappear strange to us, and we no longer recognize them asours. We realize that these feelings have been impressedupon us to a much greater extent than they were created byus. It may even happen that they horrify us, so much werethey contrary to our nature. Thus, a group of individuals,most of whom are perfectly inoffensive, may, when gatheredin a crowd, be drawn into acts of atrocity. And what we sayof these transitory outbursts applies similarly to those morepermanent currents of opinion on religious, political, litera-ry, or artistic matters which are constantly being formedaround us, whether in society as a whole or in more limited

circles.To confirm this definition of the social fact by a character-

istic illustration from common experience, one need onlyobserve the manner in which children are brought up. Con-

if"

m

Page 5: 1895 Durkheim_RulesSocMethod !

6 RULES OF SOCIOLOGICAL METHOD

sMering the facts as they are and as theyhave always been,it becomes immediately evident that all education is acontinuous effort to impose on the child ways of seeing, feel-ing, and acting which he could not have arrived at spon-taneously. From the very first hours of his life, we compelhim to eat, drink, and sleep at regular hours; we constrain

him to cleanliness, calmness, and obedience; later we exertpressure upon him in order that be may learn proper con-sideration for others, respect for customs and conventions,the need for work, etc. If, in time, this constraint ceases tobe felt, it is because it gradually gives rise to habits and tointernal tendencies that render constraint unnecessary; butnevertheless it is not abolished, for it is still the source fromwhich these habits were derived. It is true that, according toSpencer, a rational education ought to reject such methods,allowing the child to act in complete liberty; but as thispedagogic theory has never been applied by any knownpeople, it must be accepted only as an expression of personalopinion, not as a fact which can contradict the aforemen-tioned observations. What makes these facts particularlyinstructive is that the aim of education is, precisely, thesocialization of the human being; the process of education,therefore, gives us in a nutshell the historical fashion inwhich the social being is constituted. This unremitting pres-sure to which the child is subjected is the very pressure ofthe social milieu which tends to fashion him in its own image,and of which parents and teachers are merely the represent-atives and intermediaries.

It follows that sociological phenomena cannot be definedby their universality. A thought which we find in every in-dividual consciousness, a movement repeated by alI indi-viduals, is not thereby a social fact. If sociologists have been

WHAT IS A SOCIAL FACT? 7

satisfied with defining them by tl~ characteristic, it isbecause they confused them with what one might call

¯ their reincarnation in the individual. It is, however, the col-lective aspects of the beliefs, tendencies, and practices of agroup that characterize truly social phenomena. As for theforms that the collective states assume when refracted in theindividual, these are things of another sort. This duality isclearly demonstrated by the fact that these two orders ofphenomena are frequently found dissociated from one an-other. Indeed, certain of these social manners of acting andthinking acquire, by rea.~on of their repetition, a certainrigidity which on its own account crystallizes them, so tospeak, and isolates theIn from the particular events whichreflect them. They thus acquire a body, a tangible form, andconstitute a reality in their own right, quite distinct from theindividual facts which produce it. Collective habits are in-herent not only in the successive acts which they determinebut, by a privilege of which we find no example in the biolog-ical realm, they are given permanent expression in a formulawhich is repeated from mouth to mouth, transmitted byeducation, and fixed even in writing. Such is the origin andnature of legal and moral rules, popular aphorisms andproverbs, articles of faith wherein religious or politicalgroups condense their beliefs, standards of taste establishedby literary schools, etc. None of these can be found entirelyreproduced in the applications made of them by.individuals,since they can exist even without being actually applied.

No doubt, this dissociation does not always manifest itselfwith equal distinctness, but its obvious existence in thehnportant and numerous cases just cited is su~cient toprove that the social fact is a thing distinct from its indi-

vidual manifestations. Moreover, even when this dissocia-

.~.; ¯

: i.:: ! "

¯ ..!

’ iii

¯ ,!

¯ i

.ii

. i

"’.;,

...:i.:i... :

.!

."i’

;" i.",.,."

’i .i,~?:.:.:":i

Page 6: 1895 Durkheim_RulesSocMethod !

t i ttt

8 RULES OF SOCIOLOGICAL METHOD

tion is not immediately apparent, it may often be disclosedby certain devices of method. Such dissociation isindispen-sable if one wishes to separate Social facts from their alloysin order to observe them in a state of purity. Currents ofopinion, with an intensity varying according to the time andplace, impel certain groups either to more marriages, forexample, or to more suicides, or to a higher or lower birth-rate, etc. These currents are plainly social facts. At firstsight they seem inseparable from the forms they take inindividual cases. But statistics furnish us with the means ofisolating them. They are, in fact, represented with consider-able exactness by the rates of births, marriages, and suicides,that is, by the number obtained by dividing the averageannual total of marriages, births, suicides, by the number ofpersons whose ages lie within the range in which marriages,births, and suicides occur? Since each of these figures con-tains all the individual cases indiscriminately, the individualcircumstances which may have had a share in the productionof the phenomenon are neutralized and, consequently, donot contribute to its determination. The average, then, ex-presses a certain state of the group mind (l’~me collective).

Such are social phenomena, when disentangled from allforeign matter. As for their individual manifestations, theseare indeed, to a certain extent, social, since they partly re-produce a social model. Each of them also depends, and toa large extent, on the organopsychological constituti6n ofthe individual and on the particular circumstances in whichhe is placed. Thus they are not sociological phenomena inthe strict sense of the word. They belong to two realms atonce; one could call them sodopsychological. They interest

¯ Suiddes do not occur at every age, and they take place with varyingintensity at the tli~erent ages in which they occur.

iii

..¢.

W-HAT IS A SOCIAL FACT? 9

the sociologist without constituting the immediate subjectmatter of sociology. There exist in the interior of organismssimilar phenomena, compound in their nature, which formin their turn the subject matter of the "hybrid sciences,"such as physiological chemistry, for example.

The objection may be raised that a phenomenon is collec-tive only if it is common to all members of society, or at leastto most of them--in other.words, if it is truly general. Thismay be true; but it is general because it is collective (that is,more or less obligatory), and certainly not collective becausegeneral. It is a group condition repeated in the individualbecause imposed on him. It is to be found in each part be-cause it exists in the whole, rather than in the whole becauseit exists in the parts. This becomes conspicuously evident

¯ in those beliefs and practices which are transmitted to usready-made by previous generations; we receive and adoptthem because, being both collective and andent, they areinvested with a particular authority that education hastaught us to recognize and respect. It is, of course, true thata vast portion of our social culture is transmitted to us inthis way; but even when the social fact is due in part to ourdirect collaboration, its nature is not different. A collectiveemotion which bursts forth suddenly and violently in acrowd does not express merely what all the individual senti-ments had in common; it is something entirely different, aswe have shown. It results from their being together, a prod-uct of the actions and reactions which take place betweenindividual consciousnesses; and if each individual conscious-ness echoes the collective sentiment, it is by virtue of thespecial energy resident in its collective origin. If all heartsbeat in unison, this is not the result of a spontaneous andpre-established harmony but rather because an identical

m

Page 7: 1895 Durkheim_RulesSocMethod !

ro RULES OF SOCIOLOGICAL METHOD

force propels them in the same direction. Each is carriedalong by all.

We thus arrive at the point where we can formulate anddelimit in a precise way the domain of sociology. It com-prises 0nly a limited group of phenomena. A social fact isto be recognized by the power of external coercion which itexercises or is capable of exercising over individuals, and thepresence of this power may be recognized in its turn eitherby the existence of some specific sanction or by the resistanceoffered against every individual effort that tends to violateit. One can, however, define it also by its diffusion withinthe group, provided that, in conformity with our previousremarks, one takes care to add as a second and essentialcharacteristic that its own existence is independent of theindividual forms it assumes in its diffusion. This last crite-rion is perhaps, in certain cases, easier to apply than the pre-ceding one. In fact, the constraint is easy to ascertain whenit expresses itself externally by some direct reaction ofsociety, as is the case in law, morals, beliefs, customs, andeven fashions. But when it is only indirect, like the con-straint which an economic organization exercises, it cannotalways be so easily detected. Generality combined with ex-ternality may, then, be easier to establish. Nforeover, thissecond definition is but another form of the first; for if amode of behavior whose existence is external to individualconsciousnesses becomes general, this can only be broughtabout by its being imposed upon them.3

It will be seen how this definition of the sociaI fact diverges from thatwhich forms the basis of the ingenious system of M. Tarde. First of all, wewish to state that our researches have nowhere led us to observe that pre-ponderant influcace in the genesis of collective facts which M. Tacde at-tributes to imitation. Mo~over, from the preceding definition, which is nota theory but simply a r6sum6 of the immediate data of observation, it

:.:~~

¯ .iI;,N

i.:i!i~

’.i

WHAT IS A SOCIAL FACT? IZ

But these several phenomena present the same char-acteristic by which we defined the others. These "ways ofexisting" are imposed on the individual precisely in thesame fashion as the "ways of acting" of which we liave-spoken. Indeed, when we wish to know how a society isdivided politically, of what these divisions themselves arecomposed, and how complete is the fusion existing betweenthem, we shall not achieve our purpose by physical inspec-tion and by geographical observations; for these phenomenaare social, even when they have some basis in physicalnature. It is only by a study of public Iaw that a compre-hension of this organization is possible, for it is:this law thatdetermines the organization, as it equally determines ourdomestic and civil relations. This political organization is,then, no Iess obligatory than the social facts mentionedabove. If the population crowds into our cities instead ofscattering into the country, this is due to a trend of publicopinion, a collective drive that imposes fbi’s concentrationupon the individuals. We can no more choose the style ofour houses than of our clothing--at least, both are equallyobligatory. The channels of communication prescribe thedirection of internal migrations and commerce, etc., and

seems indeed to follow, not only that imitation does not always expressthe essential and char~teristic features of the social fact, but even that itnever expresses them. No doubt, every social fact is imitated; it has, as wehave just shown, a tendenc3r to become general, but that isbecauseit is social,i.e., obligatory. Its power of expansion is not the cause but the comsequenCeof its sociological character. If, further, only social facts produced thisconsequence, imitation could perhaps serve, if not to exphin them, at leastto define them. But an individual condition which produces a whole seriesof effects remains individual nevertheless. Moreover, one may ask whetherthe word "imitation’ ~ is indeed fitted to desolate an effect due to ¯ coerciveinfluence. Thus, by this single ¢xpresslou, very different phenomena, whichought to b¢ distinguished, ace confused.

’,:il !i

’.’.i".

:,,...¯ ..:;.

¯ ...:..

¯ :...

’."i"

"...’~

’.!.

Page 8: 1895 Durkheim_RulesSocMethod !

ii;ll

RULES OF SOCIOLOGICAL METHOD

even their extent. Consequently, at the very most, it shouldbe necessary to add to the list of phenomena which we haveenumerated as presenting the distinctive criterion of a socialfact only one additional category, "ways of existing"; and,as this enumeration was not meant to be rigorously exhaus-five, the addition would not be absolutely necessary.

Such an addition is perhaps not necessary, for these "waysof existing" are only crystallized "ways of acting." Thepolitical structure of a society is merely the way in which itscomponent segments have become accustomed to live withone another. If their relations are traditionally intimate, thesegments tend to fuse with one another, or, in the contrarycase, to retain their identity. The type of habitation im-posed upon us is merely the way in which our contempora-ries and our ancestors have been accustomed to constructtheir houses. The methods of communication are merely thechannels which the regular currents of commerce and .migra-tions have dug, by flowing in the Same direction. To be sure,if the phenomena of a structural character alone presentedthis permanence, one might believe that they constituted adistinct species. A legal regulation is an arrangement no lesspermanent than a type of architecture, and yet the regula-tion is a "physiological" fact. A simple moral maxim is as-suredly somewhat more malleable, but it is much more rigidthan a simple professional custom or a fashion. There is thusa whole series of degrees without a break in continuity be-tween the facts of the most articulated structure and thosefree currents of social life which are not yet definitelymolded. The differences between them are, therefore, onlydifferences in the degree of consolidation they present. Bothare simply life, more or less crystallized, l~o doubt, it maybe of some advantage to reserve the term "morphological"

,,’.~.

¯::7

:!ii~

W~IAT IS A SOCIAL FACTP 13

for those social facts which concern the sodal substratum,but only on condition of not overlooking the fact that theyare of the same nature as the others. Our definition will theninclude the whole relevant range of facts if we say: A socia~fact is every way of acting, fixed or not, capable of exerclsing onthe individual an external constraint; or again, every way ofacting which is general througkout a given society, while a~ thesame time existing in its own right independent of its indi-vidual manifestations.4

4 This close connection between life and structure, organ and function,may be easily proved in socioloKF because between these two extreme terms

¯ there exits a whole series of immediately observable intermediate stageswhich show the bond between thegn. Biology is not in the same favorableposition. But we may well believe that the inductions on this subject madeby sociology are applicable to biology and that, in organisms as well as insocieties, only differences in degree exist betwecm these two orders of facts.

Page 9: 1895 Durkheim_RulesSocMethod !

CHAPTER II

RULES FOR THE OBSERVATIONOF SOCIAL FACTS

The first and most fundamental rule is: Consider socialfacts ~s ~h{ngs.

IAt the moment when a new order of phenomena becomes

the subject matter of a science, these phenomena are alreadyrepresented in the mind not only by rather definite percep-tions but also by some kind of crudely formed concepts.Before the first rudiments of physics and chemistry ap-peared, men already had some notions concerning physico-chemical phenomena which transcended mere perception,such as are found, for example, mingled in all religions. Thereason for this is that thought and reflection are prior toscience, whichmerely uses them more methodically. Mancannot live in an environment without forming some ideasabout it according to which he regulates his behavior. But,because these ideas are nearer to us ’and more within ourmental reach than the realities to which they correspond,we tend naturally to substitute them for the latter and tomake them the very subject of our speculations. Instead ofobserving, describing, and comparing things, we are contentto focus our consciousness upon, to analyze, and to combineour ideas. Instead of a science concerned with realities, weproduce no more than an ideological analysis. To be sure,this analysis does not necessarily exclude all observation.One may appeal to the facts in order to confirm one’s

z4

"]’l ....

. .#

,i1.,:ii

,..,,

.:; $

.:.’.~

:i!!

..;:2:.)t

:’ii:?:il

TIlE OBSERVATION OF SOCIAL FACTS x5

hypotheses or the final conclusions to which they lead.But in this case, facts intervene only seconda~ly as examplesor confirmatory proofs; they are not the central subject ofscience. Such a science therefore proceeds from ideas tothings, not from things to ideas.

It is clear that this method cannot give objective results.These ideas or concepts, whatever name one gives them, arenot legitimate substitutes for things. Products of everydayexperience, their primary function is to put our actions inharmony with our environment; they are created by expe-rience and for it. Now, a representation may successfullyfulfil this function while theoretically false. Several centu-ries have elapsed since Copernicus dissipated the illusionsof the senses con£erning the movements of heavenly bodies;and yet we still habitually regulate our time according tothese illusions. In order to evoke the reaction required bythe nature of a certain stimulus, an idea need not expressthat nature faithfully, but need only inform us about theuseful or disadvantageous qualities of the thL-xg. Further,the ideas thus formed are only approximately correct in thegeneral run of cases. In fact, many times they are as dan-gerously incorrect as they are inadequate. By elaboratingsuch ideas in some fashion, one will therefore never arriveat a discovery of the laws of reality. On the contrary, theyare like a veil drawn between the thing and ourselves, con=cealing them from us the more successfully as we think themmore transparent.

Not only must such a science necessarily remain in a stateof stagnation, but it even lacks the materials upon which itmight grow. It comes into existence but to disappear, asit were, and is replaced by art. Its concepts are supposed tocontain all that is essential in reality, since they are (wrong-

Page 10: 1895 Durkheim_RulesSocMethod !

i6 RULES OF SOCIOLOGICAL METHOD

ly) identified with reality itself. Therefore, they seem tohave all that is necessary to enable us not only to understandwhat is but to prescribe what ought to be, and to describethe means of bringing it to pass. For that which is good coin-cides with that which is in conformity with the nature ofthings; that which is contrary to this nature is bad; and themeans to attain the one and escape the other derive fromthis same nature. If, therefore, reality can be thus under-stood at a glance, the study of present phenomenal realityis no longer of any practical interest; and, as this interest isthe justification for its study, it is henceforth without apurpose. Thus, an incentive is given to turn from the verysubject of our science, namely, the present and the past, andto proceed at once to the future. Instead of seeking a com-prehension of facts already acquired, it undertakes immedi-ately to discover new ones, more in accord with the endspursued by men. If men think they know what the essenceof matter is, they immediately start to look for the philos-opher’s stone. This encroachment of art on science, whichprevents the development of the latter, is facilitated, more-over, by the very circumstances which determine theawakening of scientific reflection. For, since it comes intobeing only for the purpose of satisfying vital necessities, itfinds itself quite naturally oriented toward the attainmentof practical results. The needs which it is called to relieveare always urgent, and consequently hasten it on to a con-clusion; they demand remedies, not explanations.

This procedure conforms so closely to the natural bent ofthe human mind that it is to be found in the beginnings ofthe physical sciences. It is this point of view which char-acterizes alchemy rather than chemistry, astrology rather

. ...~..

":-2

..-~

.;IJ

".~i.~!~¯ ~

.!!.’..~...~

:!.a1

..5

.,!~

" f’~

?:ii

THE OBSERVATION OF SOCIAL FACTS x7

than astronomy. It is in terms of this procedure that Baconcharacterizes the method employed by the scholars of histime and Which he sought to reform. The ideas just men-tioned are those notiones ~ulgares or praenoiiones~ which hepoints out to be the basic ideas of all sciences," where theytake the place of facts. 3 These idola, which are illusions thatdistort the real aspect of things, are nevertheless mistakenfor the things themselves. Therefore the mind, encounteringno resistance in this imaginary world and conscious of norestraint, gives itself up to boundless ambitions and comesto believe in the possibility of constructing, or rather recon-structing, the world, by virtue of its o~rn resources exclusive-ly and at the whim of its desires.

If such was the case with the natural sciences, it would bemuch more so in the history of sociology. Man already hadideas on law, morality, the family, the state, and societyitself before the advent of social science, for these ideas werenecessary conditions of his life. In sociology especially, theseprejudices or "idols," to use Bacon’s expression again, axelikely to exercise undue ascendancy over the mind and to besubstituted for the study of facts. Indeed, social things areactualized onlythrough men; they are a product of humanactivity. They appear to be nothing but the overt manifes-tation of ideas perhaps innate, contained in the mind; theyare nothing but the application of these ideas to the diversecircumstances involving the relations of men. The organiza-tion of the family, of contracts, of punishment, of the state,and of society appears thus to be simply the embodiment ofthe ideas we hold concerning society, the state, justice, etc.Consequently, these and similar facts seem to have reality

z Norm organum, I, ~6. = lb/&~ p. xT. ~ Ibi¢t., p. 36.

,!i

¯ ii"

.i"

.i

m

Page 11: 1895 Durkheim_RulesSocMethod !

x8 RULES OF SOCIOLOGICAL METHOD

only in and through the ideas which are their germ, and theideas therefore become the proper subjectmatter of soci-ology.

This approach is justified by the fact that the mind, over-run, as it is, by the details of social life which invade it fromall sides, does not perceive these details dearly enough tofed their reality. Unable to perceive the relationships whichwould properly organize these details, they give rather easilythe impression of being isolated in a vacuum, of being a sub-stance that is half-unreal and indefinitely plastic. That iswhy so many thinkers have seen in our soda1 organizationonly artifidal and more or Iess arbitrary combinations. Butif the concrete and particular detailed forms escape us, atleast we have an approximate idea of the general aspects ofcollective existence; and these schematic and crystallizedrepresentations are the superficial concepts which we employ"in ordinary life. We cannot doubt their existence, since weperceive it simultaneously with our own. Not only are theywithin us, but, as they are a product of repeated experiences,they derive from repetition and from the habit resultingfrom it, a sort of dominance and authority. We fed theirresistance when we try to shake them off. We are bound toconfer the character of reality on phenomena which opposeus. ALL arguments thus converge to make us find the truesocial reality in these phenomena.

And in truth, up to the present, sodology has dealt moreor less exdusivdy with concepts andnot with things. Comte,it is true, declared that social phenomena are natural facts,subject to natural laws. He thereby implicitly recognizedtheir character as things, for in nature there are only things.But when he passes beyond these philosophical generalitiesand attempts to apply his prindple and develop from it the

.?

.q

-i!:’i

i

,’,I

~.L".J

. ..!J..q

.:~

.~.-4".r

!i,.�

:iI

,~1

:,u

;i

x ¯

!ii !!’ii

¯:il

..’,~ ¯

i:::.~ ".

¯ ¯ ~{. .

~i

THE OBSERVATION’ OF SOCIALFACTS ~9

science implied in it, he too, takes ideas for the subjectmatter of study. It is the course Of human progress thatforms the chief subject of his sodoiogy. He begins with theidea that there is a continuous evolution of the humanspedes, consisting in an ever more complete perfection ofhuman nature; and his problem is to discover the order ofthis evolution..~Tow, the existence of this assumed evolutioncan be established only by an already completed science; itcannot, then, constitute the immediate subject of research,excepting as a conception of the mind ai~d not as a thing.And indeed, this "representation" is so completely subjec-tive that, as a matter of fact, this progress of humanityactually cannot be said to exist at all. It is only the indi-vidual societies which are born, develop, and die that can beobserved, and therefore have objective existence. If themore recent societies were merely a continuation of theirpredecessors, each more advanced type could be consideredas a duplication of the type immediately preceding, withsomething added ;one could then place them all in sequentialorder, as it were, classifying together those displaying thesame degree of development; and the series thus constitutedmight be regarded as representative of humauity. But thefacts are not so simple. A group which succeeds another isnot simply a prolongation of the latter with some newly ac-quired characteristics; it is qualitatively different from it,having gained some properties and lost others. It consti-tutes a new individuality; and all these distinct individuali-ties, being heterogeneous, cannot be juxtaposed in the samecontinuous series, and surely not in a single series. For-thesuccession of societies cannot be represented as in a singleplane; it resembles, rather, a tree with branches extending indivergent directions. In short, Comte identified historical

¯ ... .~

:;f

.i

¯ "’,i

i’

~. .,..,

:... ;21

.. !: i:i

...:.:~:

Page 12: 1895 Durkheim_RulesSocMethod !

2O RULES OF SOCIOLOGICAL METHOD

development with the idea he had of it, which does not differmuch from that of the layman. Viewed from a distance,history does convey well enough this serial and simple as-pect, appearing as a mere succession of individuals proceed-ing in the same direction because they have the same humannature. Since, moreover, it seems inconceivable to thesewriters that social evolution could possibly be anything butthe development of some human idea, it appears quitenatural to define it by the idea men form about it. Now, inproceeding thus, not only does one remain in the sphere ofideology but one assigns to sociology a concept which is noteven truly sociological.

Spencer rejects this concept, but only to replace it byanother which has the same faulty origin. With him socie-ties, and not humanity, become the subject matter ofscience. However, in.the definition he gives of society atthe outset, the thing itself disappears, giving way to thepreconception he has of it. He postulates as a se]/-evidentproposition that "a society is formed only when, in additionto juxtaposition, there i~ co-operation"--that only by thlscombination does the union of individuals become a societyin the strict sense of the world.4 Then, starting from the ideathat co-operation is the essence of social llfe, he distinguishesbetween two classes of societies according to the nature ofthe co-operation prevailing in them. "There is," he says, "aspontaneous co-operation which grows up without thoughtduring the pursuit of private ends; and there is co-operationwhich, consciously devised, implies distinct recognition ofcommon ends.’’s The former he terms "industrial"; the lat-

4 H. Spencer, T~ P~nciples of Sociology (New York: D. Appleton & Co.),

II, 244;s Ibid., p. 24~;.

i;!

3~

i

.~

::i!

’;i}!

:I.,i

-!i

.?1

’?i

!I

THE OBSERVATION OF SOCIAL FACTS 2i

ter "military societies." This distinction is Certainly thegerminal idea of his sodoIogy!

But this initial definition defines as a thing that which ismerely an idea. It is presented as the expression of an im-mediately observable fact since the definition is formulatedat the outset as an axiom. However, mere inspection doesnot reveal that co-operation is the core of sodal life. Such

¯ an affirmation would be sdenti~cally legitimate only afterall the manifestations of collective existence had been re-viewed, and it had been shown that they are all variousforms of co-operation. So here again a certain conception ofsocial reality is substituted for reality itself. ~ What is thusdefined is dearly not society but Spencer’s idea of it. Andhe has no scruples in proceeding thus, because for him, also,society is and can be only the embodiment of an idea, name-ly, this veo~ idea of co-operation by which he defines it. 7 Itwould be easy to show that, in each of the particular prob-lems he treats, his method remains the same. Thus, al-though he claims to proceed empirically, the facts accumu-lated in his sociology seem to function prindpally as argu-ments, since they are employed to illustrate analyses of con-cepts rather than to describe and explain things. Actually,all the essential points of his doctrines are capable of directdeduction from his definition of society and the different

¯ forms of co-operation. For, if our only choice is between atyrannically imposed co-operation and a free and spontane-ous one, the latter is only too evidently the ideal towardwhich humanity does and ought to tend.

6 A conception, moreover, subject to controversy. (See D/v/.d~ du tree.el

so~/, H, ~, 4-)"Co-operation, then, is at once that w~ch cannot exist without a society,

and that (or wt~ch a society exists" (o~. ¢~., II, 244).

¯ i/

’i

¯ :ii,il ¯ ¯¯¯ ..: ..,

Page 13: 1895 Durkheim_RulesSocMethod !

52 RULES OF SOCIOLOGICAL 1V~ETHOD

Th~se lay notions are to be found not only in the basicprindples of the science but also constantly in the course ofthe arguments. In the present state of knowledge, we cannotbe certain of the exact nature of the state, of sovereignty,political liberty, democracy, socialism, communism, etc.Our method should, then, require our avoidance of all useof these concepts so long as they have not been scientificallyestablished. And yet the words which express them recurconstantly in the discussions of sociologists. They are freelyemployed with great assurance, as though they correspondedto things well known and precisely defined, whereas theyawaken in us nothing but confused ideas, a tangle of vagueimpressions, prejudices, and emotions. We ridicule todaythe strange polemics built up by the doctors of the MiddleAges upon the basis of their concept of cold, warm, humid,dry, etc.; and we do not realize that we continue to applythat same method to that very order of phenomena which,because of its extreme complexity, admits of it less than anyother.

In the special branches of sociology this ideological char-acter is even more pronounced, especially in the case ofethics. One may, indeed, say that there is not a single sys-tem of ethics which has not developed from an initial ideain which its entire development was contained implicitly.Some believe that man possesses that idea at birth. Others,on the contrary, believe that it evolves more or less slowlyin the course of history. But for all empiricists as well as forrationalists, this idea is the sole true datum in ethics. Asfor the details of legal and moral laws, it is affirmed that theyhave, as it were, no existence in their own right but are mere-ly applications of this fundamental notion to the particularcircumstances of life, varied somewhat to suit the different

.%,

I

".E

%,

a

i"!

’i

J

1

...,~

.L~4

,j

¯ it...#.

/1

.’~..

THE OBSERVATION OF SOCIAL FACTS 53

cases. Accordingly, the Subject matter of the science ofethics cannot be this system of precepts which has no reality,but must be the idea from which the precepts are derivedand of which they are only diverse applications. Further-more, all the problems ordinarily raised in ethics refer not tothings but to ideas. Moralists think it necessary to deter-mine with precision the essence of the ideas of law andetI~cs, and not the nature of ethics and law. They have notyet arrived at the very simple truth that, as our ideas(repr&entaHo~) of physical things are derived from thesethings themselves and express them more or less exactly, soour idea of ethics must be derived from the observablemanifestation of the rules that are functioning under oureyes, rules that reproduce them in systematic form. Con-sequently, these rules, and not.our superficial idea of them,are actually the subject matter of science, just as actualphysical bodies~ and not the layman’s idea of them,¯ consti-tute the subject matter of physics. Therefore, it is only thesuperstructure of ethics, viz., its prolongations and echoes inthe individual consciousness, that becomes the basis of theethical systems of these writers. And this methodis appliednot only to the most general problems of this science butlikewise to special questions. From the fundamental ethicalconcepts which are treated first, the moralist proceeds to thederived ideas of family, country, responsibility, charity, andjustice; and it is always with ideas that his reflection is con-cerned.

In political economy the same situation exists. Its sub-ject matter, says John Stuart Mill, consists of those socialfacts the goal of which, principally or exclusively, is the ac-quisition of wealth,s But in order to be able to relate the

s Sy~r,~ of Lof/¢, HI, 496-

!

.~:..". i

~ii!

"i

i(i

Page 14: 1895 Durkheim_RulesSocMethod !

24 RULES OF SOCIOLOGICAL METHOD

facts thus defined as things, to the observation of thescholar, it would, at the very least, be necessary to indicateby what criteria the facts satisfying this condition are to berecognized. Now, when a science is in its infancy, we do nothave the right to aflTrm the existence of such facts, to saynothing of asserting the possibility of their identi~cation.Indeed, in every branch of research, it is possible to establishthat facts have a meaning, and what the meaning is, onlywhen the explanation of the facts is sui~ciently advanced.There is no problem more complex or less likely to be solvedon the first attempt. Nothing, then, assures us in advanceof the existence of a sphere of social activity wherein thedesire for wealth really plays such a preponderant role. Con-sequently, the subject matter of economics, so defined, com-prises not the realities given to immediate observation butmerely conjectures that are the product of pure intellect.They are "facts" imagined by the economist as being related

to the above-mentioned end, and they are facts to the extentthat he recognizes them as facts. For example, when he un-dertakes the study of what he calls "production," he thinkshe can straightway enumerate and review the prindpalagents of that process. He does not, then, determine themby observing the conditions upon which the thing he wasstudying depends, for then he .would have begun by adescription of his observations from which he drew his con-dusion. If, from the beginning of his research and in a few

words, he proceeds to this classification, it is because he hasobtained it by a simple, logical analysis. He starts from the

idea of production; in analyzing it, he finds that it implieslogically the ideas of natural forces, of work, and of tools or

~3

..!

’7

3

.’..~

".’i

.i.,.,!

-I

,,~,

..t

,] "

.

THE OBSERVATION OF SOCIAL FACTS 25

capital, and he likewise treats in their turn these derivativeideas3

The most fundamental of all economic theories, that ofvalue, is manifestly constructed according to this samemethod. If value had been studied as any fact of realityought tO be studied, the economist would indicate, first ofall, by what ctmracteristics one might recognize the thing sodesignated, then classify its varieties, investigate by method-ical inductions what the causes of its variations are, andfinally compare these various results in order to abstract ageneral formula. Theory would be introduced only when sci-ence had reached a sufficient stage of advancement. On thecontrary, it is introduced at the very outset. In order toconstruct economic theory, the economist is content to med-itate and to focus his attention on his own idea of value, thatis, as an object capable oi being exchanged; he finds thereinthe idea of utility, scarcity, etc., and with these products ofhis analysis he constructs his definition. To be sure, he con-

firms it by several examples. But, considering the innumer-able facts such a theory must account for, how can one granteven the slightest validity to the necessarily limited numberof facts thus cited at random?

Thus, the actual contribution of scientific investigation toeconomics and ethics is very limited, while that of art ispreponderant. Ethical theory is limited merely to a few dis-cussions on the idea of duty, the good and right. And eventhese abstract speculations do not constitute a science,

The ideological nature of economics is implied even in the expressionsused by economists. The question is always one of the concept of utility,savings, investment, expenditure. (S~ Gide, Pr~’~pe~ d’Scono~� ~oH~-~e,Book Ill, chap. i, § i; chap. [i, § i; chap. iii, § i.)

: !

[:

.:i .

.:.:.

¯ i

,...

.3!

Page 15: 1895 Durkheim_RulesSocMethod !

RULES OF SOCIOLOGICAL METHOD

strictly speaking, since their object is the determination notof that which is, in fact, the supreme rule of morality but ofwhat it ought to be. Similarly, economists are today prin-dpally occupied with the problem of whether society o~g/~to be organized on an individualistic or socialistic basis,whether it is bdL~r that the state should intervene in indus-trial and commercial relations, or whether it is bstt~r toabandon them to private initiative; whether one ought touse a single monetary standard, or a bimetallic system, etc.It contains few laws in the proper sense of the word; evenwhat are commonly called "laws" are generally unworthyof this designation since they are mereIy maxims for action,or practical precepts in disguise. The famous law of supplyand demand, for example, has never been inductively estab-lished, as should be the case with a law referring to economicreality. No experiment or systematic comparison has everbeen undertaken for the purpose of establishing that, i~fact, economic relations do conform to this law. All thatthese economists could do, and actually did do, was todemonstrate by dialectics that, in order properly to promotetheir interests, individuals ought to proceed according tothis law, and that every other line of action would be harm-ful to those who engage in it and would imply a serious errorof judgment. It is fair and logical that the most productiveindustries should be the most attractive and that the holdersof the products most in demand and most scarce should sellthem at the highest prices. But this quite logical necessityresembles in no way the necessity that the true laws of na-ture present. The latter express the regulations accordingto which facts are really interconnected, not the way inwhich it is good that they should be interconnected.

What we say of this law may be repeated for all those that

¯ ¯iI

THE OBSERVATION OF SOCIAL FACTS 27

orthodox economics designates as "natural" and which,moreover, are scarcely more than particular cases of it.They are natural, if one likes, in the sense that they enun-date the means which it is really or seemingly natural toemploy in order to attain a certain hypothetical end, butthey do not deserve this designation if natural law means aninductively determined way of behavior in nature. In brief,they are merely maxims of practical wisdom; and they havebeen more or less plausibly presented as the very expressionof reality only because it was supposed, rightly or wrongly,that these counsels were indeed followed by the average manin the average case.

In spite of all these doctrines, socialphenomena are thingsand ought to be treated as things. To demonstrate thisproposition, it isunnecessary tophilosophize on their natureand to discuss the analogies they present with the phenome-na of lower realms of existence. It is sufficient to note thatthey are the unique data of the sociologist. All that is given,all that is subject to observation, has thereby the characterof a thing. To treat phenomena as things is to treat them asdata, and these constitute the point of departure of science.Now, social phenomena present this character incontestably.What is given is not the idea that men form of value; forthat is inaccessible, but only the values established in thecourse of economic relations; not conceptions of the moralideal, but the totality of rules which actually determineconduct; not the idea¯ of utility or wealth, but all the detailsof economic organization. Even assuming the possibilitythat sodal life is merely the development of certain ¯ideas,these ideas are nevertheless not immediately given. Theycannot be perceived or known directly, but only throughthe phenomenal reality expressing them. We do not know

’...

:iil¯ ¢::,

...:~

Page 16: 1895 Durkheim_RulesSocMethod !

i .i

28 RULES OF SOCIOLOGICAL METHOD

a priori whether ideas form the basis of the diverse currentsof social life, nor what they are. Only after having tracedthese currents back to their sources shall we know whencethey issue. . .

We must, therefore, consider social phenomena in them-selves as distinct from the consciously formed representa-tions of them in the mind; we must study them objectivelyas external things, for it is this character that they presentto us. If this exteriority should prove to be only apparent,the advance of science will bring the disillusionment and weshall see our conception of social phenomena change, as itwere, from the objective to the subjective. But in any case,the solution cannot be anticipated; and even if we finallyarrive at the result that social phenomena do not possess allthe intrinsic characteristics of the thing, we ought at first totreat them as if they had. This rule is applicable, then, toall social reality without exception. Even phenomena whichgive the strongest impression of being arbitrary arrange-ments ought to be thus considered. The voluntary characterof a practice or an institution should nev,r be assumed b,fore-hand. Moreover, if we may introduce our personal observa-tion, it has always been our experience that, when this pro-cedure is followed, facts most arbitrary in appearance willcome to present, after more attentive observation, qualitiesof consistency and regularity that are symptomatic of theirobjectivity.

The foregoing statements concerning the distinctive char-acteristics of the social fact give us sufficient assuranceabout the nature of this objectivity to prove that it is notillusory. Indeed, the most important characteristic of a"thing" is the impossibility of its modification by a simpleeffort of the will. Not that the thing is refractory to all

ii

¯ : ii!¯: !,~:i

THE OBSERVATION OF SOCIAL FACTS 29

modification, but a mere act of the wili is insufficient to pro-duce a change in R; it requires a more or lessstrenuous effortdue to the resistance which it offers, and, moreover, the effortis not always successful. We have already seen that socialfacts have this characteristic. Far from being a product ofthe will, they determine it from without; they are like moldsin which our actions are inevitably shaped. This necessityis often inescapable. But even when we triumph over it, theopposition encountered signifies clearly to us the presence ofsomething not depending upon ourselves. Thus, in consider-ing social phenomena as things, we merely adjust our con-ceptions in conformity to their nature.

Clearly, the reform needed in sociology is at all pointsidentical with that which has transformed psychology in thelast thirty years. Just as Comte and Spencer declare thatsocial facts are facts of nature, without, however, treatingthem as things, so the different empirical schools had longrecognized the natural character of psychological phenome-na, but continued to apply to them a purely ideologicalmethod. In fact, the empiricists, not less than their adver-saries, proceeded exclusively by introspection. Now, thefacts obtained thereby are too few in number, too fleetingand plastic, to be able to control and to correct the corre-sponding ideas fixed in us by habit. If they are not sub-jected to some other check, nothing counterbalances them;consequently, they take the place of facts and become thesubject matter of science. Thus, neither Locke nor Condillacstudied psychological phenomena objectively. They did notstudy sensation in itself but their particular idea of it.Therefore, although in certain respects they prepared theway for scientific psychology, its actual origin is to be datedmuch later, when it had finally been established thatstates

i.’."

::iilli’.:

i:!!:

Page 17: 1895 Durkheim_RulesSocMethod !

3° RULES OF SOCIOLOGICAL METHOD

of consciousness can and ought to be considered from with-out, and not from the point of view of the consciousnessexperiencing them. Such is the great revolution accom-pushed in this branch of studies. All the specific proceduresand all the new methods by which this science has beenenriched are only diverse means of realizing more completelythis fundamental idea. It remah~s for sociology to make thissame advance, to pass from the subjective stage, which it hasstill scarcely outgrown, to the objective.

Fortunately, this transformation is less difficult to effecthere than in psychology. Indeed, psychological facts arenaturally given as conscious states of the individual, fromwhom they do not seem to be even separable. Internal bydefinition, it seems that they can be treated as extemaI onlyby doing violence to their nature. Not only is an effort ofabstraction necessary, but in addition a whole series of pro-cedures and artifices in order to hold them continuouslywithin this point of view. Social facts, on the contrary, qual-ify far more naturally and immediately as things. Law isembodied in codes; the currents of daily life are recorded instatistical figures and historical monuments; fashions arepreserved in costumes; and taste in works of art. By theirvery nature they tend toward an independent existence out-side the individual consciousnesses, which they dominate.In order to disclose their character as things, it is unneces-sary to manipuIate them ingeniously. From this point ofview, sociology has a significant advantage over psychology,an advantage not hitherto perceived, and one which shouldhasten its development. Its facts are perhaps more difficultto interpret because more complex, but they are more easilyarrived at. Psychology, on the contrary, has difficulties notonly in the manipulation of its facts but also in rendering

i ¯

...j

.L

..t,

...,J...3

THE OBSERVATION OF SOCIAL FACTS 3x

them explicit. Consequently,¯ we believe that, once this prin-ciple of sociological method is generally recognized andpracticed, sociology will progress with a rapidity difficult toforecast from its present tardiness of development and willeven overtake psychology, whose present relative advantageis due solely to historical priority.~*

IIBut the experience of our predecessors has shown that,

in order to assure the practical realization of the truth justenunciated, it is not enough to be thoroughly convincedone’s self, or even to set forth a theoretical demonstrationof it. The mind is so naturalIy inclined to underrate anddisregard this particular truth that a relapse into the olderrors will inevitably follow unless sociologists are willing tosubmit themselves to a rigorous discipline. We shall there-fore formulate the principal rules for such a discipline, all ofthem corollaries of the foregoing theorem.

x. The first corollary is: All precor~eptions mu~t be eradi-cated. A special demonstration of this rule is unnecessary; itfollows easily from all our previous statements. It is, more-over, the basis of all scientific method. The logical doubt ofDescartes is, in its essence, only an application of it. If, atthe moment of the foundation of science, Descartes resolvesto question all ideas he had previously received,it is becausehe wishes to employ only scientifically developed concepts,that is, concepts constructed according to the method in-stituted by himself; all those having some other origin, then,

~* It is true that the greater complexity of social facts makes the sciencemore difficult. But, in compensation, precisely bvcause sociology is thelatest comer, it is in a position to profit by the progress made in the scieacesconcerned with lower stages of existence and to learn from them. Tkisutilization of previous experimtmts win certainly accelerate its development.

.’i;

¯~!ii

!i:ii

:i.

..!.:

Page 18: 1895 Durkheim_RulesSocMethod !

i1"1-[~1

RULES OF SOCIOLOGICAL METHOD

must be rejected, at least provisionally. We have alreadyseen that Bacon’s theory of the "idols" has the same mean-ing. The two great doctrines that have been so often op-posed to one another thus agree on this essential point. Thesociologist ought, therefore, whether at the moment of thedetermination of his research objectives or in the course ofhis demonstrations, to repudiate resolutely the use of con-cepts orginating outside of science for totally unscientificneeds. He must emancipate himself from the fallacious ideasthat dominate the mind of the layman; he must throw off,once and for all, the yoke of these empirical categories, whichfrom long continued habit have become tyrannical. At thevery least, if at times he is obliged to resort to them, heought to do so fully conscious of their trifling value, so thathe will not assign to them a role out of proportion to theirreal importance.

The frequent interference of sentiment makes this eman-cipation from lay ideas particularly dif~cult in sociology.Indeed, our political and religious beliefs and our moralstandards carry with them an emotional tone that is notcharacteristic of our attitude toward physical objects; con-sequentty, this emotional character infects our manner ofconceiving and explaining them. The ideas we form ofthings have a vital interest for us, just as the objects, them-selves, and thus assume an authority which brooks no con-tradiction. Every opinion that disturbs them is treated withhostility. Yff a proposition is not in agreement, for example,with one’s idea of patriotism or of individual dignity, it isdenied, whatever its proofs may be. We cannot admit itstruth; it is given no consideration at all; and our emotion, tojustify our attitude, has no difficulty in suggesting reasonsthat are readily found convincing. These ideas may, indeed,

.ii,l"1

i;.i". "..’: J..

:i!~1:;.’

THEOBSERVATION OF SOCIAL FACTS 33

have such prestige that they do not even tolerate scientificexamination. The very fact of submitting them, as well asthe phenomena they represent, to cold, dry analysis, isrevolting to certain minds. Whoever undertakes the studyof morality objectively, and as an external reality, seems tothese sensitive creatures to be devoid of alI moral sense, justas the vivisectionist seems to the layman devoid of commonsensibility. Far from admitting that these sentiments shouldthemselves be drawn under scientific scrutiny; it is to themthat these writers feel they must appeal in order to treatscientifically the parallel social facts.

~’Woe to the schoIar," writes an eloquent historian ofreligions, "who approaches divine matters without havingin the depths of his consciousness, in the innermost inde-structible regions of iris being, where the souls of his ances-tors sleep, an unknown sanctuary from which rises now andthen the~ aroma of incense, a line of a psalm, a sorrowful ortriumphal cry that as a child he sent to heaven along withhis brothers, and that creates immediate communion withthe prophets of yore!’’zx

One cannot protes.t too strongly against this mysticaldoctrine, which, I~e all mysticism, is essentially a disguisedempiricism, the negation of all science. Sentiments pertain-ing to social things enjoy no privilege not possessed by othersentiments, for their origin is the same. They, too, havebeen formed in the course of history; they are a product ofhuman experience, which is, however, confused and unorgan-ized. They are not due to some transcendental insight intoreality but result from all sorts of impressions and emotionsaccumulated according to circumstances, without order andwithout methodical interpretation. Far from conveying in-

"~ J. Darmesteter, Lss Pro~tcs d/I~ra~ p. 9.

¯. ’!

.:.~,"1,

:i~ii

...’

.;!i

:i’:’

:i~!i¯ " .,:l

Page 19: 1895 Durkheim_RulesSocMethod !

ii. r lit

34’ RULES OF SOCIOLOGICAL METHOD

sights superior to rational ones, these sentiments are simplystrong but confused states of mind. To accord them a dom-inant role means giving supremacy to the inferior facultiesof intelligence over the superior, condemning one’s self topure logomachy. Such a science can satisfy only those whoprefer to think with their feelings end emotions rather thanwith their understanding, and who prefer the immediateand confused syntheses of first impression to the patient andluminous analyses of reason. Sentiment is a subject forscientific study, not the criterion of scientific truth. More-over, every science encounters analogous resistances at theoutset. There was a time when sentiments relating to thethings of the physical world opposed with equal energy theestablishment of the physical sciences, because they, too,had a religious or moral character. We believe, therefore,that this prejudice, pursued from One science to the next, willfinally disappear also from its last retreat, sociology, leavinga free field for the true scientific endeavor.

2. As it happens, this first rule for sociology is entirelynegative. It teaches the sociologist to escape the realm oflay ideas and to turn his attention toward facts, but it doesnot tell him how to take hold of the facts in order to studythem objectively.

Every scientific investigation is directed toward a limitedclass of phenomena, included in tl~ same definition. Thefirst step of the sociologist, then, ought to be to define thethings he treats, in order that his subject matter may beknown. This is the first and most indispensable condition ofall proofs and verifications. A theory, indeed, can bechecked only if we know how to recognize the facts of whichit is intended to give an account. Moreover, since this initialdefinition determines the very subject matter of science, this

’i

IIII III 11111

THE OBSERVATION OF SOCIAL FACTS 35

subject matter will or will not be a thing, depending on thenature of the definition.

In order to be objective, the definition must obviouslydeal with phenomena not as ideas but in terms of their in-herent properties. It must characterize them by elementsessential to their nature, not by their conformity to an intel-lectual ideal. Now, at the very beginning of research, whenthe facts have not yet been analyzed, the only ascertainablecharacteristics are those external enough to be immediatelyperceived. Those that are less obvious may be perhaps moresignificant, and their explanatory value is more important;but they are unknown tO science at this stage, and they canbe anticipated only by substituting some hypothetical con-ception.in the place of reality. It is imperative, then, thatthe material included under this fundamental definition besought among the more external characteristics of sociologi-cal phenomena. On the other hand, this definition shouldinclude, without exception or distinction, all phenomenapresenting to an equal extent these characteristics, for wehave neither the reason nor the means for choosing amongthem. These characteristics are our only clue to reality; con-sequentIy, they must be given complete authority in ourselection of facts. No other criterion could even partiallyjustify any suspension of, or exception to, this rule. Whence¯ our second corotlary: The subject matter of every sociologicalstudy should comprise a group of phenomena defined in ad-vance by certain common external characteristics, and all pke-nogg~ena so defined should be included w~2hin this group.

For example, we note the existence of certain acts, allpresenting the external characteristic that they evoke fromsociety the particular reaction called punishment.. We con-stitute them as a separate group, to which we give a common

III

!./ :j

"’;i

.ii,

..d

Page 20: 1895 Durkheim_RulesSocMethod !

35 RULES OF SOCIOLOGICAL METHOD

label; we call every punished act a crime, and crime thusdefined becomes the object of a special science, criminology.Similarly, we observe within all known societies small groupswhose special characteristic is that they are composed pre-ponderantly of individuals who are blood-kin, united bylegal bonds. We classify together the facts relating thereto,and give a particular name to the group of facts so created,"domestic relations." We call every aggregate of this kinda family, and this becomes the subject of a special investiga-tion which has not yet received a specific name in sociologi-cal terminology. In passing from the family in general to thedifferent family types, the same rule should be applied. Forexample, the study of the dan and the matriarchal or thepatriarchal family should begin with a definition constructedaccording to the same method. The field of each problem,whether general or particular, must be similarly circum-scribed.

By proceeding thus, the sociologist, from the very .first, isfirmly grounded in reality. Indeed, the pattern of such aclassification does not depend .on him or on the cast of hisindividual mind but on the nature of things. The criteriaaccording to which they are placed in a particular cate-gory can be recognized by everyone; and the concepts thusformed do not always, or even generally, tally with that ofthe layman. For example, manfestations of free thought orviolations of etiquette, so regularly and severely penalizedin many societies, are evidently considered crimes in thecommon-sense view even in these societies. Similarly, in theusual acceptance of the words a clan is not a family. Butsuch discrepancies are not important, for it is not our aimsimply to discover a method for identifying with sufficientaccuracy the facts to which the words of ordinary language

.i

"3

~i:~:~

THE OBSERVATION OF SOCIALFACTS 37

refer and the ideas they conveF. We need, rather, to formu-late entirely new concepts, appropriate to the requirementsof science and expressed in" an appropriate terminology. Ofcourse, lay concepts are not entirely useless to the scholar;they serve as suggestions and guides. They inform us of theexistence, somewhere, of an aggregation of phenomenawhich, bearing the same name, must, in consequence,probably have certain characteristics in common. Sincethese concepts havre always had some reference to phenome-na, they even indicate to us at times, though roughly, wherethese phenomena are to be found. But, as they have beencrudely formed, they quite naturally do not coincide exactIywith the scientific concepts, which have been establishedfor a set purpose.=

This rule, as obvious and important as it is, is seldomobserved in sociology. Precisely because it treats everydaythings, such as the family, property, crime, etc., the sociol-ogist most often thinks it unnecessary to define them rigor-ously at the outset. We are so accustomed to use theseterms, and they recur so constantly in our conversation, thatit seems unnecessary to render their meaning precise. Wesimply refer to the common notion, but this common notionis very often ambiguous. As a result of this ambiguity,things that are very different in reality are given the same

¯ ~ In actual practice one always starts with the lay concept and the layterm. One inquires whether, among the things which this word confusedlyconnotes, there are some which present common external characteristics.If this is the case, and if the concept formed by the grouping of the factsthus brought together coincides, if not totally (which is rare), at least to large extent, with the lay concept, it will be possil~le to’continue to designatethe former by the same term as the latter, that is, to retain in science theexpression used in everyday language. But if the gap is too considerable, ifthe common notion confuses a plurality of distinct ideas, the creation of newand distinctive terms becomes necessary.

...:,

i.i:

."i

".’.i

!(

.’[

Page 21: 1895 Durkheim_RulesSocMethod !

Im imll

38 RULES OF SOCIOLOGICAL METHOD

name and the same explanation, and this leads to boundlessconfusion.

For example, two sorts of monogamous unions exist:those monogamous in fact, and those monogamous by law.In the former, the husband has only one wife, although he isallowed by law to possess several; in the latter, polygamy islegally forbidden. In several animal species and in certainprimitive societies monogamy "in fact" is to be found, notsporadically, but with the same prevalence as if imposed bylaw. When a tribe is dispersed over a vast area, there is Iitflesocial contact, and consequently the individuals live isolatedfrom one another. In such a case each man naturally seeksonly one wife, because in this state of isolation it is di~cultfor him to secure several. Compulsory monogamy, on thecontrary, is observed only in the highest societies. Thesetwo types of conjugal unions have, then, a very differentsignificance; and yet the same word serves to designate themboth. We commonly call certain animals ’tmonogamous," al-though they have nothing resembling legal control. NowSpencer, in his study of marriage, uses the word "monog-amy" in its ordinary equivocal meaning, without definingit. As a result the evolution of marriage seems to him topresent an unaccountable anomaly, since he thinks he ob-serves a higher form of the sexual union as early as the firstphases of historical development, while it seems to disappearin the intermediate period, only to reappear later. He thenconcludes that there is no positive correlation between so-cial progress in general and progress toward a perfect typeof family life. A timely definition would have preventedthis error,xs

zs The same absence of definition caused the occasional statem~ts thatdemocracy is realized both at the beginning and at the end of history. Thetruth is that primitive and modem democracy are very different from oneanother.

THE OBSERVATION OF SOCIAL FACTS 39

In other cases great care may be exercised in defining theobjects of investigation; but instead of grouping under thesameheading all phenomena having the same external prop-erties, only a selected number of them are included. Thus,only certain ones are designated as a kind of "~lite," andthese alone are regarded as coming within the category. Asfor the others, they are considered as having usurped thesedistinctive signs and are disregarded. It is ~asy to foreseethat in this way only a subjective and incomplete picturecan be attained. Such an omission can be made only byapplying a preconcei.ved idea, since, at the beginning ofscience, no research could possibly have already establishedthe legitimacy of this usurpation, even if it were possible tohave done so. The only possible reason for retaining thephenomena chosen was, then, that they conformed, morethan the others, to a certain ideal conception concerning thissort of reality.

For example, M. Garofalo, at the beginning of hisCrir~inalogie, demonstrates very well that "the sociologicalconcept of crime’’~4 has to form the point of departure of thisscience. Only, in setting up his concept, he does not compareindlscriminately all acts which have been repressed byregular punishments in the different social types. He com-pares only certain ones among them, namely, those offendingthe most general and universal of the-moral feelings. Themoral sentiments which have disappeared in the course ofevolution are not, to him, grounded in the nature of things,since they have not survived; conseqt~ently, the acts whichhave been deemed criminal because of their violation ofthese particular sentiments seem to him to have owed this.designation only to accidental and more or less pathologicalcircumstances. But it is by virtue of an entirely personal

.4 0p. cir., p. ~.

.’i

:.!

. i,Ei.i

Page 22: 1895 Durkheim_RulesSocMethod !

4o RULES OF SOCIOLOGICAL METHOD

conception of morality that he makes this elimination. Hestarts from the idea that ¯moral evolution, taken at its veryfount or near its source, carries with it all sorts Of dross andimpurities, which it then progressively eliminates, and thatit is only today that it has succeeded in freeing itself fromatl the adventitious elements which, in primitive times,troubled its course. But this principle is neither an evidentaxiom nor a demonstrated truth; it is only a hypothesis, andindeed one without justification. The variable aspects of themoral sense are not less grounded in the nature of thingsthan are the immutable; the variations in standards ofmorality merely testify to the corresponding variations inlife. In zoSlogy, the forms peculiar to the lower species arenot regarded as less natural than those occurring at the otherpoints on the evolutionary scale. Similarly, these acts whichwere condemned as crimes by primitive societies and havesince lost this designation are really criminal in relation tothese societies, quite like those which we continue to represstoday. The former correspond to the changing, the latter tothe constanh conditions of social life; but the former are notany more artificial than those acts which are consideredcrimes today.

But, even if these acts had unduly assumed the criminalcharacter, they ought not to be sharply separated from theothers; for the pathological forms of a phenomenon are notdifferent in nature from the normal forms, and it is thereforenecessary to observe the former as welI as the latter in orderto determine this nature. Morbidity is not absolutely anti-thetical to health; these are two varieties of the same phe-nomenon, and each tends to explain the other. This is a ruleIong recognized and practiced in biology and in psychology,and the sociologist is equally under an obligation to respect

oi

THEOBSERVATION OF SOCIAL FACTS 4I

it. Unless one asserts that the same phenomenon can be duesometimes to one cause and sometimes to another, that is,unless one denies the principle of causality, the causes whichimpress on an act the mark of crime, in an abnormal manner,c.ann6t differ qualitatively from those producing the sameeffect in a normal manner; they differ only in degree or theydiffer because they do not act in the same environment. Theabnormal crime, then, is still a crime and ought, consequent-ly, to be included in the definition of crime. What M. Garo-falo actually does is to take as the genus that which is ontya species or merely a simple variety. The facts to which hisdefinition of criminality applies represent only an infinites-imal minority among those it should include, for it appliesneither to religious crimes, nor to violations of etiquette,ceremonial, tradition, etc. If-these have disappeared fromour modern codes, they make up, on the contrary, almostthe entire penal law of former societies.

The same flaw in method causes certain observers to denythe existence of any species of morality among savages.’s

They start with the idea that our morality is ~ke morality.It is evident, however, that our morality is either unknownor in a rudimentary state among prknitive peoples and thatthis discrimination is clearly arbitrary. If we apply our sec-ond corollary in this case, everything changes. To decidewhether a precept belongs to the moral order, we must de-termine whether or not it presents the external mark ofmorality; this mark is a widespread repressive sanction, thatis, a condemnation by public opinion that punishes allviolations of the precept. Whenever we are presented with

¯ s See Lubbock, Or/g/n of C~dl~iza~i~, chap. viii: It is a still morn wide-spread, sad not less false, opinion that the ancient religions are amoral orimmoral. The truth is that they have a morality of their own.

¯i¯!

.!i:

Page 23: 1895 Durkheim_RulesSocMethod !

42 RULES OF SOCIOLOGICAL METHOD

a fact having this characteristic, we have no right to denyits moral character, for this characteristic proves that it hasthe same nature as other moral facts. Not only are socialregulations of this kind met with in primitive societies, butthey are even more numerous there than in civilized socie-ties. A large number of acts which today are left to the freechoice of individuals are obligatory among them. Thus wemay realize the errors we commit by omitting definitions orby defining inadequately.

But, it will be said that, in defining phenomena by theirapparent characteristics, we are allowing to certain super-ficial properties a significance greater than that of morefundamental attributes. Are we not, by a veritable inver-sion of logical order, beginning at the summit instead of thebase? Thus, when we define crime in terms of punishment,one is almost inevitably exposed to the accusation of de-riving crime from punishment, or, as a wen-known quotationputs it, of considering the scaffold, and not the crime, as thesource of ignominy. This reproach rests upon a confusion.Since the definition in question is placed at the beginningsof the science, it cannot possibly aim at a statement con-ceming the essence of reality; tlmt must be attained subse-quently. The sole function of the definition is to establishcontact with things; and since the latter can be grasped bythe mind only from its exteriors, the definition expressesthem in terms of their external qualities. It does not explainthese things thereby; it furnishes merely a just basis forfurther explanations. Certainly, punishment is not the es-sence of crime; but it does constitute a symptom thereof,and consequently, in order to understand crime, we mustbegin with punishment.

The aforementioned objection would be well founded only

TI-IE OBSERVATION OF SOCIAL FACTS 43

if these external characteristics were at the same time acci-dental, that is, if they were not bound up with the funda-mental properties of things. Under these conditions indeed,after science had pointed them out, it could not possibly gofarther; it could not penetrate the deeper layers of reality,since there would be no necessary connection between sur-face and essence. But, if the principle of causality is valid,when certain characteristics are found identically and with-out exceptions in all the phenomena of a certain order, onemay be assured that they are closely connected with thenature of the latter and bound up with it. And if to a givengroup of acts there is attached also the peculiarity of apenal sanction, an intimate bond must exist between punish-ment and the intrinsic attributes of these acts. Consequent-ly, however superficial they may be, these properties, pro-vided that they have been sygtematicatly observed, clearlypoint out to the scientist the course which he must follow inorder to penetrate more to the core of the things in question.They are the first and indispensable llnk in the sequence tobe unfolded by science in the course of its explanations.

Since objects are perceived only through sense percep-tion, we can conclude: Science, to be objective, ought tostart, not with concepts formed independent to them, butwith these same perceptions. It ought to borrow the ma-terials for its initial definitions directly from perceptualdata. And, as a matter of fact, one need only reflect on thereal nature of scientific work to understand that it cannotproceed otherwise. It needs concepts that adequately ex-press things as they actually are, and not as everyday llfefinds it useful to conceive them. Now those concepts formu-lated without the discipline of science do not fulfil thiscondition. Science, then, has to create new concepts; it must

Page 24: 1895 Durkheim_RulesSocMethod !

44 RULES OF SOCIOLOGICAL METHOD

dismiss all lay notions and the terms expressing them, andreturn to sense perception, the primary and necessary sub-stance underlying all concepts. From sensation all generalideas flow, whether they be true or false, scientific or impres-sionistic. The point of departure of science, or speculativeknowledge, cannot be different from that of lay, or practical,knowledge. It is only beyond this point, namely, in the man-ner of elaboration of these common data, that divergencesbegin.

3- But sensation may easily be subjective. It is a rule inthe natural sciences to discard those data of sensation thatare too subjective, in order to retain exclusively those pre-senting a sufficient degree of objectivity. Thus the physicistsubstitutes, for the vague impressions of temperature andelectricity, the visual registrations of the thermometer orthe electrometer. The sociologist must take the same pre-cautions. The external characteristics in terms of which hedefines the objects of his researches should be as objective aspossible.

We may lay down as a principle that social facts lendthemselves more readily to objective representation in pro-portion as their separation from the individual facts ex-pressing them is more complete. Indeed, the degree of objec-tivity of a sense perception is proportionate to the degree ofstability of its object; for objectivity depends upon theexistence of a constant and identical point of reference towhich the representation can be referred and which permitsthe elimination of what is variable, and hence subjective, init. But if the points of reference themselves are variable, ifthey are perpetually shifting in relation to each other, thereis no common standard, and the scientist has no means ofdistinguishing between those impressions which are external

.t.

i:¯ :i.’.

¯ , .,...,-,..

THE OBSERVATION OF SOCIAL FACTS 4S

and those that are subjective. So long as social life is notseparated from the individual or particular events whichcomprise it, and has no separate existence, it will presentthis dilemma. As these events differ among themselves andchange in time, and as we assume the life of society to beinseparable from them, they communicate their mutabilityto it. Social life consists, then, of free currents perpetuallyin the process of transformation and incapable of beingmentally fixed by the observer, and the scholar cannot ap-proach the study of social reality from this angle. But weknow that it possesses the power of crystallization withoutceasing to be itself. Thus, apart from the individual acts towhich they give rise, coIIect~ve habits find expression indefinite forms: legal rules, moral regulations, popular prov-erbs, social conventions, etc. As these forms have a perma-nent existence and do not change with the diverse applica-tions made of them, they constitute a fixed object, aconstant standard within the observer’s reach, exclusive ofsubjective impressions and purely personal observations.A legal regulation is what it is, and there are no two ways ofIooking at it. Since, on the other hand, these practices aremerely social life consolidated, it is legitimate, except whereotherwise stated,~6 to study the latter through the former.

When, then, the sociologist undertakes t~ investigation ofsome order of social facts, ke must endeavor to consider themfrom an aspect that is indepengent of their individual manifes-tat~ns. It is this principle that we have applied in studyingthe diverse forms of social solidarity and their evolution,through the medium of the legal structure which reflects

x~ It.would be necessary, for example~ in order to invalidate this substitu-

tion, to have reason to. believe that, at a given moment, law no longer

expresses the actual state o~ social relations.

Page 25: 1895 Durkheim_RulesSocMethod !

46 RULES OF SOCIOLOGICAL METHOD

them?~ On the other hand, an attempt to distingmish andclassify the different family types on the basis of the literarydescription given us by travelers and historians is exposedto the danger of confusing the most diverse species and ofbringing together the most dissimilar types. If the legalstructure of the family and, more specifical~y, the right ofsuccession are taken as the basis of classification, objectivecriteria are at hand which, while not infallible, will preventmany errors, z8 In order to classify the different kinds ofcrhnes, one has to try to reconstruct the ways of living andthe occupational customs that are practiced in the differentworlds of crime. One will then recognize as many crimino-logical types as there are different forms o1 this organization.To achieve an understanding of customs and popular be-liefs, one must investigate the proverbs and epigrams thatexpress them. No doubt, in proceeding thus, we leave theconcrete data of collective life temporarily outside the realmof science; and yet, however changeable and unstable it maybe, its unintelligibility need not be assumed. In order to

follow a methodical course, we must establish the founda-tions of science on solid ground and not on shifting sand.We must approach the social realm where it offers the easiestaccess to scientific investigation. Only subsequently will itbe possible to push research further and, by successive ap-proximations, to encompass, little by little, this fleetingreality, which the human mind will never, perhaps, be ableto grasp completely.

,7 See Dh~ion d~ trava~ social, Book I.

,s Cf. the author’s "Introduction ~ Is sociologie de la famille," in Anna~sd~ la Faz~ ¢Ie~ Ie~tres de Bord~a~z, z889.

I,

i CHAPTER III

RULES FOR DISTINGUISHING BETWEENTHE NORMAL AND THE

PATHOLOGICAL

Observation conducted according to the preceding rulescovers two types of facts which are very dissimilar in certainrespects: those which conform to given standards and thosewhich "ought" to be differentnin other words, normal andpathological phenomena. We have seen that it is necessaryto include them both in the definition with which all researchmust begin. But if their nature is in certain respects iden-tical, they constitute, nevertheless, two different varietiesof facts, which need to be distinguished. Can science makethis distinction?

The question is of the greatest importance, for on itssolution depends the role assigned to science, and especiallyto the science of man. According to a theory whose partisansbelong to most diverse schools, science can teach us nothingabout what we ought to desire. It is concerned, they say,only with facts which all have the same vaJue and interestfor 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 notwhat ends should be pursued. In order to determine notwhat is but what is desirable, we need to resort to the un-conscious, by whatever name it may be designated: "feel-ing," "instinct," "vital urge," etc. Science, says a writeralready quoted, can indeed illuminate the world, but it

47