Archaeology Of Knowledge, Introduction

Original Publication: L’Archéologie du savoir (Gallimard, 1969)
— Foucault, Michel. Introduction.” In Archaeology Of Knowledge, edited by A. M. Sherida Smith, 3-20. Vintage, 1982.

For many years now his­to­ri­ans have pre­ferred to turn their at­ten­tion to long pe­ri­ods, as if, be­neath the shifts and changes of po­lit­i­cal events, they were try­ing to re­veal the sta­ble, al­most in­de­struc­tible sys­tem of checks and bal­ances, the ir­re­versible processes, the con­stant read­just­ments, the un­der­ly­ing ten­den­cies that gather force, and are then sud­denly re­versed af­ter cen­turies of con­ti­nu­ity, the move­ments of ac­cu­mu­la­tion and slow sat­u­ra­tion, the great silent, mo­tion­less bases that tra­di­tional his­tory has cov­ered with a thick layer of events. The tools that en­able his­to­ri­ans to carry out this work of analy­sis are partly in­her­ited and partly of their own mak­ing: mod­els of eco­nomic growth, quan­ti­ta­tive analy­sis of mar­ket move­ments, ac­counts of de­mo­graphic ex­pan­sion and con­trac­tion, the study of cli­mate and its long-term changes, the fix­ing of so­ci­o­log­i­cal con­stants, the de­scrip­tion of tech­no­log­i­cal ad­just­ments and of their spread and con­ti­nu­ity. These tools have en­abled work­ers in the his­tor­i­cal field to dis­tin­guish var­i­ous sed­i­men­tary strata; lin­ear suc­ces­sions, which for so long had been the ob­ject of re­search, have given way to dis­cov­er­ies in depth. From the po­lit­i­cal mo­bil­ity at the sur­face down to the slow move­ments of material civil­i­sa­tion”, ever more lev­els of analy­sis have been es­tab­lished: each has its own pe­cu­liar dis­con­ti­nu­ities and pat­terns; and as one de­scends to the deep­est lev­els, the rhythms be­come broader. Beneath the rapidly chang­ing his­tory of gov­ern­ments, wars, and famines, there emerge other, ap­par­ently un­mov­ing his­to­ries: the his­tory of sea routes, the his­tory of corn or of gold-min­ing, the his­tory of drought and of ir­ri­ga­tion, the his­tory of crop ro­ta­tion, the his­tory of the bal­ance achieved by the hu­man species be­tween hunger and abun­dance. The old ques­tions of the tra­di­tional analy­sis (What link should be made be­tween dis­parate events? How can a causal suc­ces­sion be es­tab­lished be­tween them? What con­ti­nu­ity or over­all sig­nif­i­cance do they pos­sess? Is it pos­si­ble to de­fine a to­tal­ity, or must one be con­tent with re­con­sti­tut­ing con­nec­tions?) are now be­ing re­placed by ques­tions of an­other type: which strata should be iso­lated from oth­ers? What types of se­ries should be es­tab­lished? What cri­te­ria of pe­ri­odi­s­a­tion should be adopted for each of them? What sys­tem of re­la­tions (hierarchy, dom­i­nance, strat­i­fi­ca­tion, uni­vo­cal de­ter­mi­na­tion, cir­cu­lar causal­ity) may be es­tab­lished be­tween them? What se­ries of se­ries may be es­tab­lished? And in what large-scale chrono­log­i­cal table may dis­tinct se­ries of events be de­ter­mined?

At about the same time, in the dis­ci­plines that we call the his­tory of ideas, the his­tory of sci­ence, the his­tory of phi­los­o­phy, the his­tory of thought, and the his­tory of lit­er­a­ture (we can ig­nore their speci­ficity for the mo­ment), in those dis­ci­plines which, de­spite their names, evade very largely the work and meth­ods of the his­to­rian, at­ten­tion has been turned, on the con­trary, away from vast uni­ties like periods” or centuries” to the phe­nom­ena of rup­ture, of dis­con­ti­nu­ity. Beneath the great con­ti­nu­ities of thought, be­neath the solid, ho­mo­ge­neous man­i­fes­ta­tions of a sin­gle mind or of a col­lec­tive men­tal­ity, be­neath the stub­born de­vel­op­ment of a sci­ence striv­ing to ex­ist and to reach com­ple­tion at the very out­set, be­neath the per­sis­tence of a par­tic­u­lar genre, form, dis­ci­pline, or the­o­ret­i­cal ac­tiv­ity, one is now try­ing to de­tect the in­ci­dence of in­ter­rup­tions. Interruptions whose sta­tus and na­ture vary con­sid­er­ably. There are the epis­te­mo­log­i­cal acts and thresh­olds de­scribed by Bachelard: they sus­pend the con­tin­u­ous ac­cu­mu­la­tion of knowl­edge, in­ter­rupt its slow de­vel­op­ment, and force it to en­ter a new time, cut it off from its em­pir­i­cal ori­gin and its orig­i­nal mo­ti­va­tions, cleanse it of its imag­i­nary com­plic­i­ties; they di­rect his­tor­i­cal analy­sis away from the search for silent be­gin­nings, and the never-end­ing trac­ing-back to the orig­i­nal pre­cur­sors, to­wards the search for a new type of ra­tio­nal­ity and its var­i­ous ef­fects. There are the dis­place­ments and trans­for­ma­tions of con­cepts: the analy­ses of G. Canguilhem may serve as mod­els; they show that the his­tory of a con­cept is not wholly and en­tirely that of its pro­gres­sive re­fine­ment, its con­tin­u­ously in­creas­ing ra­tio­nal­ity, its ab­strac­tion gra­di­ent, but that of its var­i­ous fields of con­sti­tu­tion and va­lid­ity, that of its suc­ces­sive rules of use, that of the many the­o­ret­i­cal con­texts in which it de­vel­oped and ma­tured. There is the dis­tinc­tion, which we also owe to Canguilhem, be­tween the mi­cro­scopic and macro­scopic scales of the his­tory of the sci­ences, in which events and their con­se­quences are not arranged in the same way: thus a dis­cov­ery, the de­vel­op­ment of a method, the achieve­ments, and the fail­ures, of a par­tic­u­lar sci­en­tist, do not have the same in­ci­dence, and can­not be de­scribed in the same way at both lev­els; on each of the two lev­els, a dif­fer­ent his­tory is be­ing writ­ten. Recurrent re­dis­tri­b­u­tions re­veal sev­eral pasts, sev­eral forms of con­nec­tion, sev­eral hi­er­ar­chies of im­por­tance, sev­eral net­works of de­ter­mi­na­tion, sev­eral tele­olo­gies, for one and the same sci­ence, as its pre­sent un­der­goes change: thus his­tor­i­cal de­scrip­tions are nec­es­sar­ily or­dered by the pre­sent state of knowl­edge, they in­crease with every trans­for­ma­tion and never cease, in turn, to break with them­selves (in the field of math­e­mat­ics, M. Serres has pro­vided the the­ory of this phe­nom­e­non). There are the ar­chi­tec­tonic uni­ties of sys­tems of the kind analysed by M. Guéroult, which are con­cerned not with the de­scrip­tion of cul­tural in­flu­ences, tra­di­tions, and con­ti­nu­ities, but with in­ter­nal co­her­ences, ax­ioms, de­duc­tive con­nec­tions, com­pat­i­bil­i­ties. Lastly, the most rad­i­cal dis­con­ti­nu­ities are the breaks ef­fected by a work of the­o­ret­i­cal trans­for­ma­tion which es­tab­lishes a sci­ence by de­tach­ing it from the ide­ol­ogy of its past and by re­veal­ing this past as ide­o­log­i­cal. To this should be added, of course, lit­er­ary analy­sis, which now takes as its unity, not the spirit or sen­si­bil­ity of a pe­riod, nor groups”, schools”, generations”, or movements”, nor even the per­son­al­ity of the au­thor, in the in­ter­play of his life and his creation”, but the par­tic­u­lar struc­ture of a given œu­vre, book, or text.

And the great prob­lem pre­sented by such his­tor­i­cal analy­ses is not how con­ti­nu­ities are es­tab­lished, how a sin­gle pat­tern is formed and pre­served, how for so many dif­fer­ent, suc­ces­sive minds there is a sin­gle hori­zon, what mode of ac­tion and what sub­struc­ture is im­plied by the in­ter­play of trans­mis­sions, re­sump­tions, dis­ap­pear­ances, and rep­e­ti­tions, how the ori­gin may ex­tend its sway well be­yond it­self to that con­clu­sion that is never given - the prob­lem is no longer one of tra­di­tion, of trac­ing a line, but one of di­vi­sion, of lim­its; it is no longer one of last­ing foun­da­tions, but one of trans­for­ma­tions that serve as new foun­da­tions, the re­build­ing of foun­da­tions. What one is see­ing, then, is the emer­gence of a whole field of ques­tions, some of which are al­ready fa­mil­iar, by which this new form of his­tory is try­ing to de­velop its own the­ory: how is one to spec­ify the dif­fer­ent con­cepts that en­able us to con­ceive of dis­con­ti­nu­ity (threshold, rup­ture, break, mu­ta­tion, trans­for­ma­tion)? By what cri­te­ria is one to iso­late the uni­ties with which one is deal­ing; what is a sci­ence? What is an œu­vre? What is a the­ory? What is a con­cept? What is a text? How is one to di­ver­sify the lev­els at which one may place one­self, each of which pos­sesses its own di­vi­sions and form of analy­sis? What is the le­git­i­mate level of for­mal­i­sa­tion? What is that of in­ter­pre­ta­tion? Of struc­tural analy­sis? Of at­tri­bu­tions of causal­ity?

In short, the his­tory of thought, of knowl­edge, of phi­los­o­phy, of lit­er­a­ture seems to be seek­ing, and dis­cov­er­ing, more and more dis­con­ti­nu­ities, whereas his­tory it­self ap­pears to be aban­don­ing the ir­rup­tion of events in favour of sta­ble struc­tures.

But we must not be taken in by this ap­par­ent in­ter­change. Despite ap­pear­ances, we must not imag­ine that cer­tain of the his­tor­i­cal dis­ci­plines have moved from the con­tin­u­ous to the dis­con­tin­u­ous, while oth­ers have moved from the tan­gled mass of dis­con­ti­nu­ities to the great, un­in­ter­rupted uni­ties; we must not imag­ine that in the analy­sis of pol­i­tics, in­sti­tu­tions, or eco­nom­ics, we have be­come more and more sen­si­tive to over­all de­ter­mi­na­tions, while in the analy­sis of ideas and of knowl­edge, we are pay­ing more and more at­ten­tion to the play of dif­fer­ence; we must not imag­ine that these two great forms of de­scrip­tion have crossed with­out recog­nis­ing one an­other.

In fact, the same prob­lems are be­ing posed in ei­ther case, but they have pro­voked op­po­site ef­fects on the sur­face. These prob­lems may be summed up in a word: the ques­tion­ing of the doc­u­ment. Of course, it is ob­vi­ous enough that ever since a dis­ci­pline such as his­tory has ex­isted, doc­u­ments have been used, ques­tioned, and have given rise to ques­tions; schol­ars have asked not only what these doc­u­ments meant, but also whether they were telling the truth, and by what right they could claim to be do­ing so, whether they were sin­cere or de­lib­er­ately mis­lead­ing, well in­formed or ig­no­rant, au­then­tic or tam­pered with. But each of these ques­tions, and all this crit­i­cal con­cern, pointed to one and the same end: the re­con­sti­tu­tion, on the ba­sis of what the doc­u­ments say, and some­times merely hint at, of the past from which they em­anate and which has now dis­ap­peared far be­hind them; the doc­u­ment was al­ways treated as the lan­guage of a voice since re­duced to si­lence, its frag­ile, but pos­si­bly de­ci­pher­able trace. Now, through a mu­ta­tion that is not of very re­cent ori­gin, but which has still not come to an end, his­tory has al­tered its po­si­tion in re­la­tion to the doc­u­ment: it has taken as its pri­mary task, not the in­ter­pre­ta­tion of the doc­u­ment, nor the at­tempt to de­cide whether it is telling the truth or what is its ex­pres­sive value, but to work on it from within and to de­velop it: his­tory now or­gan­ises the doc­u­ment, di­vides it up, dis­trib­utes it, or­ders it, arranges it in lev­els, es­tab­lishes se­ries, dis­tin­guishes be­tween what is rel­e­vant and what is not, dis­cov­ers el­e­ments, de­fines uni­ties, de­scribes re­la­tions. The doc­u­ment, then, is no longer for his­tory an in­ert ma­te­r­ial through which it tries to re­con­sti­tute what men have done or said, the events of which only the trace re­mains; his­tory is now try­ing to de­fine within the doc­u­men­tary ma­te­r­ial it­self uni­ties, to­tal­i­ties, se­ries, re­la­tions. History must be de­tached from the im­age that sat­is­fied it for so long, and through which it found its an­thro­po­log­i­cal jus­ti­fi­ca­tion: that of an age-old col­lec­tive con­scious­ness that made use of ma­te­r­ial doc­u­ments to re­fresh its mem­ory; his­tory is the work ex­pended on ma­te­r­ial doc­u­men­ta­tion (books, texts, ac­counts, reg­is­ters, acts, build­ings, in­sti­tu­tions, laws, tech­niques, ob­jects, cus­toms, etc.) that ex­ists, in every time and place, in every so­ci­ety, ei­ther in a spon­ta­neous or in a con­sciously or­gan­ised form. The doc­u­ment is not the for­tu­nate tool of a his­tory that is pri­mar­ily and fun­da­men­tally mem­ory; his­tory is one way in which a so­ci­ety recog­nises and de­vel­ops a mass of doc­u­men­ta­tion with which it is in­ex­tri­ca­bly linked.

To be brief, then, let us say that his­tory, in its tra­di­tional form, un­der­took to memorise” the mon­u­ments of the past, trans­form them into doc­u­ments, and lend speech to those traces which, in them­selves, are of­ten not ver­bal, or which say in si­lence some­thing other than what they ac­tu­ally say; in our time, his­tory is that which trans­forms doc­u­ments into mon­u­ments. In that area where, in the past, his­tory de­ci­phered the traces left by men, it now de­ploys a mass of el­e­ments that have to be grouped, made rel­e­vant, placed in re­la­tion to one an­other to form to­tal­i­ties. There was a time when ar­chae­ol­ogy, as a dis­ci­pline de­voted to silent mon­u­ments, in­ert traces, ob­jects with­out con­text, and things left by the past, as­pired to the con­di­tion of his­tory, and at­tained mean­ing only through the resti­tu­tion of a his­tor­i­cal dis­course; it might be said, to play on words a lit­tle, that in our time his­tory as­pires to the con­di­tion of ar­chae­ol­ogy, to the in­trin­sic de­scrip­tion of the mon­u­ment.

This has sev­eral con­se­quences. First of all, there is the sur­face ef­fect al­ready men­tioned: the pro­lif­er­a­tion of dis­con­ti­nu­ities in the his­tory of ideas, and the emer­gence of long pe­ri­ods in his­tory proper. In fact, in its tra­di­tional form, his­tory proper was con­cerned to de­fine re­la­tions (of sim­ple causal­ity, of cir­cu­lar de­ter­mi­na­tion, of an­tag­o­nism, of ex­pres­sion) be­tween facts or dated events: the se­ries be­ing known, it was sim­ply a ques­tion of defin­ing the po­si­tion of each el­e­ment in re­la­tion to the other el­e­ments in the se­ries. The prob­lem now is to con­sti­tute se­ries: to de­fine the el­e­ments proper to each se­ries, to fix its bound­aries, to re­veal its own spe­cific type of re­la­tions, to for­mu­late its laws, and, be­yond this, to de­scribe the re­la­tions be­tween dif­fer­ent se­ries, thus con­sti­tut­ing se­ries of se­ries, or tables”: hence the ever-in­creas­ing num­ber of strata, and the need to dis­tin­guish them, the speci­ficity of their time and chronolo­gies; hence the need to dis­tin­guish not only im­por­tant events (with a long chain of con­se­quences) and less im­por­tant ones, but types of events at quite dif­fer­ent lev­els (some very brief, oth­ers of av­er­age du­ra­tion, like the de­vel­op­ment of a par­tic­u­lar tech­nique, or a scarcity of money, and oth­ers of a long-term na­ture, like a de­mo­graphic equi­lib­rium or the grad­ual ad­just­ment of an econ­omy to cli­matic change); hence the pos­si­bil­ity of re­veal­ing se­ries with widely spaced in­ter­vals formed by rare or repet­i­tive events. The ap­pear­ance of long pe­ri­ods in the his­tory of to­day is not a re­turn to the philoso­phers of his­tory, to the great ages of the world, or to the pe­ri­odi­s­a­tion dic­tated by the rise and fall of civil­i­sa­tions; it is the ef­fect of the method­olog­i­cally con­certed de­vel­op­ment of se­ries. In the his­tory of ideas, of thought and of the sci­ences, the same mu­ta­tion has brought about the op­po­site ef­fect; it has bro­ken up the long se­ries formed by the progress of con­scious­ness, or the tele­ol­ogy of rea­son, or the evo­lu­tion of hu­man thought; it has ques­tioned the themes of con­ver­gence and cul­mi­na­tion; it has doubted the pos­si­bil­ity of cre­at­ing to­tal­i­ties. It has led to the in­di­vid­u­al­i­sa­tion of dif­fer­ent se­ries, which are jux­ta­posed to one an­other, fol­low one an­other, over­lap and in­ter­sect, with­out one be­ing able to re­duce them to a lin­ear schema. Thus, in place of the con­tin­u­ous chronol­ogy of rea­son, which was in­vari­ably traced back to some in­ac­ces­si­ble ori­gin, there have ap­peared scales that are some­times very brief, dis­tinct from one an­other, ir­re­ducible to a sin­gle law, scales that bear a type of his­tory pe­cu­liar to each one, and which can­not be re­duced to the gen­eral model of a con­scious­ness that ac­quires, pro­gresses, and re­mem­bers.

Second con­se­quence: the no­tion of dis­con­ti­nu­ity as­sumes a ma­jor role in the his­tor­i­cal dis­ci­plines. For his­tory in its clas­si­cal form, the dis­con­tin­u­ous was both the given and the un­think­able: the raw ma­te­r­ial of his­tory, which pre­sented it­self in the form of dis­persed events - de­ci­sions, ac­ci­dents, ini­tia­tives, dis­cov­er­ies; the ma­te­r­ial, which, through analy­sis, had to be re­arranged, re­duced, ef­faced in or­der to re­veal the con­ti­nu­ity of events. Discontinuity was the stigma of tem­po­ral dis­lo­ca­tion that it was the his­to­ri­an’s task to re­move from his­tory. It has now be­come one of the ba­sic el­e­ments of his­tor­i­cal analy­sis. its role is three­fold. First, it con­sti­tutes a de­lib­er­ate op­er­a­tion on the part of the his­to­rian (and not a qual­ity of the ma­te­r­ial with which he has to deal): for he must, at least as a sys­tem­atic hy­poth­e­sis, dis­tin­guish the pos­si­ble lev­els of analy­sis, the meth­ods proper to each, and the pe­ri­odi­s­a­tion that best suits them. Secondly, it is the re­sult of his de­scrip­tion (and not some­thing that must be elim­i­nated by means of his analy­sis): for he is try­ing to dis­cover the lim­its of a process, the point of in­flec­tion of a curve, the in­ver­sion of a reg­u­la­tory move­ment, the bound­aries of an os­cil­la­tion, the thresh­old of a func­tion, the in­stant at which a cir­cu­lar causal­ity breaks down. Thirdly, it is the con­cept that the his­to­ri­an’s work never ceases to spec­ify (instead of ne­glect­ing it as a uni­form, in­dif­fer­ent blank be­tween two pos­i­tive fig­ures); it as­sumes a spe­cific form and func­tion ac­cord­ing to the field and the level to which it is as­signed: one does not speak of the same dis­con­ti­nu­ity when de­scrib­ing an epis­te­mo­log­i­cal thresh­old, the point of re­flex­ion in a pop­u­la­tion curve, or the re­place­ment of one tech­nique by an­other. The no­tion of dis­con­ti­nu­ity is a para­dox­i­cal one: be­cause it is both an in­stru­ment and an ob­ject of re­search; be­cause it di­vides up the field of which it is the ef­fect; be­cause it en­ables the his­to­rian to in­di­vid­u­alise dif­fer­ent do­mains but can be es­tab­lished only by com­par­ing those do­mains. And be­cause, in the fi­nal analy­sis, per­haps, it is not sim­ply a con­cept pre­sent in the dis­course of the his­to­rian, but some­thing that the his­to­rian se­cretly sup­poses to be pre­sent: on what ba­sis, in fact, could he speak with­out this dis­con­ti­nu­ity that of­fers him his­tory — and his own his­tory — as an ob­ject? One of the most es­sen­tial fea­tures of the new his­tory is prob­a­bly this dis­place­ment of the dis­con­tin­u­ous: its trans­fer­ence from the ob­sta­cle to the work it­self; its in­te­gra­tion into the dis­course of the his­to­rian, where it no longer plays the role of an ex­ter­nal con­di­tion that must be re­duced, but that of a work­ing con­cept; and there­fore the in­ver­sion of signs by which it is no longer the neg­a­tive of the his­tor­i­cal read­ing (its un­der­side, its fail­ure, the limit of its power), but the pos­i­tive el­e­ment that de­ter­mines its ob­ject and val­i­dates its analy­sis.

Third con­se­quence: the theme and the pos­si­bil­ity of a to­tal his­tory be­gin to dis­ap­pear, and we see the emer­gence of some­thing very dif­fer­ent that might be called a gen­eral his­tory. The pro­ject of a to­tal his­tory is one that seeks to re­con­sti­tute the over­all form of a civil­i­sa­tion, the prin­ci­ple ma­te­r­ial or spir­i­tual of a so­ci­ety, the sig­nif­i­cance com­mon to all the phe­nom­ena of a pe­riod, the law that ac­counts for their co­he­sion — what is called metaphor­i­cally the face” of a pe­riod. Such a pro­ject is linked to two or three hy­pothe­ses; — it is sup­posed that be­tween all the events of a well-de­fined spa­tio-tem­po­ral area, be­tween all the phe­nom­ena of which traces have been found, it must be pos­si­ble to es­tab­lish a sys­tem of ho­mo­ge­neous re­la­tions: a net­work of causal­ity that makes it pos­si­ble to de­rive each o them, re­la­tions of anal­ogy that show how they sym­bol­ise one an­other, or how they all ex­press one and the same cen­tral core; it is also sup­posed that one and the same form of his­toric­ity op­er­ates upon eco­nomic struc­tures, so­cial in­sti­tu­tions and cus­toms, the in­er­tia of men­tal at­ti­tudes, tech­no­log­i­cal prac­tice, po­lit­i­cal be­hav­iour, and sub­jects them all to the same type of trans­for­ma­tion; lastly, it is sup­posed that his­tory it­self may be ar­tic­u­lated into great units — stages or phases — which con­tain within them­selves their own prin­ci­ple of co­he­sion. These are the pos­tu­lates that are chal­lenged by the new his­tory when it speaks of se­ries, di­vi­sions, lim­its, dif­fer­ences of level, shifts, chrono­log­i­cal speci­fici­ties, par­tic­u­lar forms of re­han­dling, pos­si­ble types of re­la­tion. This is not be­cause it is try­ing to ob­tain a plu­ral­ity of his­to­ries jux­ta­posed and in­de­pen­dent of one an­other: that of the econ­omy be­side that of in­sti­tu­tions, and be­side these two those of sci­ence, re­li­gion, or lit­er­a­ture; nor is it be­cause it is merely try­ing to dis­cover be­tween these dif­fer­ent his­to­ries co­in­ci­dences of dates, or analo­gies of form and mean­ing. The prob­lem that now pre­sents it­self — and which de­fines the task of a gen­eral his­tory — is to de­ter­mine what form of re­la­tion may be le­git­i­mately de­scribed be­tween these dif­fer­ent se­ries; what ver­ti­cal sys­tem they are ca­pa­ble of form­ing; what in­ter­play of cor­re­la­tion and dom­i­nance ex­ists be­tween them; what may be the ef­fect of shifts, dif­fer­ent tem­po­ral­i­ties, and var­i­ous re­han­dlings; in what dis­tinct to­tal­i­ties cer­tain el­e­ments may fig­ure si­mul­ta­ne­ously; in short, not only what se­ries, but also what series of se­ries” — or, in other words, what tables” it is pos­si­ble to draw up. A to­tal de­scrip­tion draws all phe­nom­ena around a sin­gle cen­tre — a prin­ci­ple, a mean­ing, a spirit, a world-view, an over­all shape; a gen­eral his­tory, on the con­trary, would de­ploy the space of a dis­per­sion.

Fourth and last con­se­quence: the new his­tory is con­fronted by a num­ber of method­olog­i­cal prob­lems, sev­eral of which, no doubt, ex­isted long be­fore the emer­gence of the new his­tory, but which, taken to­gether, char­ac­terise it. These in­clude: the build­ing-up of co­her­ent and ho­mo­ge­neous cor­pora of doc­u­ments (open or closed, ex­hausted or in­ex­haustible cor­pora), the es­tab­lish­ment of a prin­ci­ple of choice (according to whether one wishes to treat the doc­u­men­ta­tion ex­haus­tively, or adopt a sam­pling method as in sta­tis­tics, or try to de­ter­mine in ad­vance which are the most rep­re­sen­ta­tive el­e­ments); the de­f­i­n­i­tion of the level of analy­sis and of the rel­e­vant el­e­ments (in the ma­te­r­ial stud­ied, one may ex­tract nu­mer­i­cal in­di­ca­tions; ref­er­ences — ex­plicit or not — to events, in­sti­tu­tions, prac­tices; the words used, with their gram­mat­i­cal rules and the se­man­tic fields that they in­di­cate, or again the for­mal struc­ture of the propo­si­tions and the types of con­nec­tion that unite them); the spec­i­fi­ca­tion of a method of analy­sis (the quan­ti­ta­tive treat­ment of data, the break­ing-down of the ma­te­r­ial ac­cord­ing to a num­ber of as­sign­a­ble fea­tures whose cor­re­la­tions are then stud­ied, in­ter­pre­ta­tive de­ci­pher­ment, analy­sis of fre­quency and dis­tri­b­u­tion); the de­lim­i­ta­tion of groups and sub-groups that ar­tic­u­late the ma­te­r­ial (regions, pe­ri­ods, uni­tary processes); the de­ter­mi­na­tion of re­la­tions that make it pos­si­ble to char­ac­terise a group (these may be nu­mer­i­cal or log­i­cal re­la­tions; func­tional, causal, or ana­log­i­cal re­la­tions; or it may be the re­la­tion of the signifier” (signs) to the signified” (signifé).

All these prob­lems are now part of the method­olog­i­cal field of his­tory. This field de­serves at­ten­tion, and for two rea­sons. First, be­cause one can see to what ex­tent it has freed it­self from what con­sti­tuted, not so long ago, the phi­los­o­phy of his­tory, and from the ques­tions that it posed (on the ra­tio­nal­ity or tele­ol­ogy of his­tor­i­cal de­vel­op­ment (devenir), on the rel­a­tiv­ity of his­tor­i­cal knowl­edge, and on the pos­si­bil­ity of dis­cov­er­ing or con­sti­tut­ing a mean­ing in the in­er­tia of the past and in the un­fin­ished to­tal­ity of the pre­sent). Secondly, be­cause it in­ter­sects at cer­tain points prob­lems that are met with in other fields — in lin­guis­tics, eth­nol­ogy, eco­nom­ics, lit­er­ary analy­sis, and mythol­ogy, for ex­am­ple. These prob­lems may, if one so wishes, be la­belled struc­tural­ism. But only un­der cer­tain con­di­tions: they do not, of them­selves, cover the en­tire method­olog­i­cal field of his­tory, they oc­cupy only one part of that field — a part that varies in im­por­tance with the area and level of analy­sis; apart from a num­ber of rel­a­tively lim­ited cases, they have not been im­ported from lin­guis­tics or eth­nol­ogy (as is of­ten the case to­day), but they orig­i­nated in the field of his­tory it­self — more par­tic­u­larly, in that of eco­nomic his­tory and as a re­sult of the ques­tions posed by that dis­ci­pline; lastly, in no way do they au­tho­rise us to speak of a struc­tural­ism of his­tory, or at least of an at­tempt to over­come a conflict” or opposition” be­tween struc­ture and his­tor­i­cal de­vel­op­ment: it is a long time now since his­to­ri­ans un­cov­ered, de­scribed, and analysed struc­tures, with­out ever hav­ing oc­ca­sion to won­der whether they were not al­low­ing the liv­ing, frag­ile, pul­sat­ing history” to slip through their fin­gers. The struc­ture/​de­vel­op­ment op­po­si­tion is rel­e­vant nei­ther to the de­f­i­n­i­tion of the his­tor­i­cal field, nor, in all prob­a­bil­ity, to the de­f­i­n­i­tion of a struc­tural method.

This epis­te­mo­log­i­cal mu­ta­tion of his­tory is not yet com­plete. But it is not of re­cent ori­gin ei­ther, since its first phase can no doubt be traced back to Marx. But it took a long time to have much ef­fect. Even now — and this is es­pe­cially true in the case of the his­tory of thought — it has been nei­ther reg­is­tered nor re­flected upon, while other, more re­cent trans­for­ma­tions — those of lin­guis­tics, for ex­am­ple — have been. It is as if it was par­tic­u­larly dif­fi­cult, in the his­tory in which men re­trace their own ideas and their own knowl­edge, to for­mu­late a gen­eral the­ory of dis­con­ti­nu­ity, of se­ries, of lim­its, uni­ties, spe­cific or­ders, and dif­fer­en­ti­ated au­tonomies and de­pen­dences. As if, in that field where we had be­come used to seek­ing ori­gins, to push­ing back fur­ther and fur­ther the line of an­tecedents, to re­con­sti­tut­ing tra­di­tions, to fol­low­ing evo­lu­tive curves, to pro­ject­ing tele­olo­gies, and to hav­ing con­stant re­course to metaphors of life, we felt a par­tic­u­lar re­pug­nance to con­ceiv­ing of dif­fer­ence, to de­scrib­ing sep­a­ra­tions and dis­per­sions, to dis­so­ci­at­ing the re­as­sur­ing form of the iden­ti­cal. Or, to be more pre­cise, as if we found it dif­fi­cult to con­struct a the­ory, to draw gen­eral con­clu­sions, and even to de­rive all the pos­si­ble im­pli­ca­tions of these con­cepts of thresh­olds, mu­ta­tions, in­de­pen­dent sys­tems, and lim­ited se­ries — in the way in which they had been used in fact by his­to­ri­ans. As if we were afraid to con­ceive of the Other in the time of our own thought.

There is a rea­son for this. If the his­tory of thought could re­main the lo­cus of un­in­ter­rupted con­ti­nu­ities, if it could end­lessly forge con­nec­tions that no analy­sis could undo with­out ab­strac­tion, if it could weave, around every­thing that men say and do, ob­scure syn­the­sis that an­tic­i­pate for him, pre­pare him, and lead him end­lessly to­wards his fu­ture, it would pro­vide a priv­i­leged shel­ter for the sov­er­eignty of con­scious­ness. Continuous his­tory is the in­dis­pens­able cor­rel­a­tive of the found­ing func­tion of the sub­ject: the guar­an­tee that every­thing that has eluded him may be re­stored to him; the cer­tainty that time will dis­perse noth­ing with­out restor­ing it in a re­con­sti­tuted unity; the promise that one day the sub­ject — in the form of his­tor­i­cal con­scious­ness — will once again be able to ap­pro­pri­ate, to bring back un­der his sway, all those things that are kept at a dis­tance by dif­fer­ence, and find in them what might be called his abode. Making his­tor­i­cal analy­sis the dis­course of the con­tin­u­ous and mak­ing hu­man con­scious­ness the orig­i­nal sub­ject of all his­tor­i­cal de­vel­op­ment and all ac­tion are the two sides of the same sys­tem of thought. In this sys­tem, time is con­ceived in terms of to­tal­i­sa­tion and rev­o­lu­tions are never more than mo­ments of con­scious­ness.

In var­i­ous forms, this theme has played a con­stant role since the nine­teenth cen­tury: to pre­serve, against all de­cen­trings, the sov­er­eignty of the sub­ject, and the twin fig­ures of an­thro­pol­ogy and hu­man­ism. Against the de­cen­tring op­er­ated by Marx — by the his­tor­i­cal analy­sis of the re­la­tions of re­duc­tion, eco­nomic de­ter­mi­na­tions, and the class strug­gle — it gave place to­wards the end of the nine­teenth cen­tury, to the search for a to­tal his­tory, in which all the dif­fer­ences of a so­ci­ety might be re­duced to a sin­gle form, to the or­gan­i­sa­tion of a world—view, to the es­tab­lish­ment of a sys­tem of val­ues, to a co­her­ent type of civil­i­sa­tion. To the de­cen­tring op­er­ated by the Nietzschean ge­neal­ogy, it op­posed the search for an orig­i­nal foun­da­tion that would make ra­tio­nal­ity the te­los of mankind, and link the whole his­tory of thought to the preser­va­tion of this ra­tio­nal­ity, to the main­te­nance of this tele­ol­ogy, and to the ever nec­es­sary re­turn to this foun­da­tion. Lastly, more re­cently, when the re­searches of psy­cho­analy­sis, lin­guis­tics, and eth­nol­ogy have de­cen­tred the sub­ject in re­la­tion to the laws of his de­sire, the forms of his lan­guage, the rules of his ac­tion, or the games of his myth­i­cal or fab­u­lous dis­course, when it be­came clear that man him­self, ques­tioned as to what he was, could not ac­count for his sex­u­al­ity and his un­con­scious, the sys­tem­atic forms of his lan­guage, or the reg­u­lar­i­ties of his fic­tions, the theme of a con­ti­nu­ity of his­tory has been re­ac­ti­vated once again; a his­tory that would be not di­vi­sion, but de­vel­op­ment (devenir); not an in­ter­play of re­la­tions, but an in­ter­nal dy­namic; not, a sys­tem, but the hard work of free­dom; not form, but the un­ceas­ing ef­fort of a con­scious­ness turned upon it­self, try­ing to grasp it­self in its deep­est con­di­tions: a his­tory that would be both an act of long, un­in­ter­rupted pa­tience and the vi­vac­ity of a move­ment, which, in the end, breaks all bounds. If one is to as­sert this theme, which, to the immobility” of struc­tures, to their closed” sys­tem, to their nec­es­sary synchrony”, op­poses the liv­ing open­ness of his­tory, one must ob­vi­ously deny in the his­tor­i­cal analy­ses them­selves the use of dis­con­ti­nu­ity, the de­f­i­n­i­tion of lev­els and lim­its, the de­scrip­tion of spe­cific se­ries, the un­cov­er­ing of the whole in­ter­play of dif­fer­ences. One is led there­fore to an­thro­pol­o­gise Marx, to make of him a his­to­rian of to­tal­i­ties, and to re­dis­cover in him the mes­sage of hu­man­ism; one is led there­fore to in­ter­pret Nietzsche in the terms of tran­scen­den­tal phi­los­o­phy, and to re­duce his ge­neal­ogy to the level of a search for ori­gins; lastly, one is led to leave to one side, as if it had never arisen, that whole field of method­olog­i­cal prob­lems that the new his­tory is now pre­sent­ing.- For, if it is as­serted that the ques­tion of dis­con­ti­nu­ities, sys­tems and trans­for­ma­tions, se­ries and thresh­olds, arises in all the his­tor­i­cal dis­ci­plines (and in those con­cerned with ideas or the sci­ences no less than those con­cerned with eco­nom­ics and so­ci­ety), how could one op­pose with any sem­blance of le­git­i­macy development” and system”, move­ment and cir­cu­lar reg­u­la­tions, or, as it is some­times put crudely and un­think­ingly, history” and structure”?

The same con­ser­v­a­tive func­tion is at work in the theme of cul­tural to­tal­i­ties (for which Marx has been crit­i­cised, then trav­es­tied), in the theme of a search for ori­gins (which was op­posed to Nietzsche, be­fore an at­tempt was made to trans­pose him into it), and in the theme of a liv­ing, con­tin­u­ous, open his­tory. The cry goes up that one is mur­der­ing his­tory when­ever, in a his­tor­i­cal analy­sis - and es­pe­cially if it is con­cerned with thought, ideas, or knowl­edge - one is seen to be us­ing in too ob­vi­ous a way the cat­e­gories of dis­con­ti­nu­ity and dif­fer­ence, the no­tions of thresh­old, rup­ture and trans­for­ma­tion, the de­scrip­tion of se­ries and lim­its. One will be de­nounced for at­tack­ing the in­alien­able rights of his­tory and the very foun­da­tions of any pos­si­ble his­toric­ity. But one must not be de­ceived: what is be­ing be­wailed with such ve­he­mence is not the dis­ap­pear­ance of his­tory, but the eclipse of that form of his­tory that was se­cretly, but en­tirely re­lated to the syn­thetic ac­tiv­ity of the sub­ject, what is be­ing be­wailed is the development” (devenir) that was to pro­vide the sov­er­eignty of the con­scious­ness with a safer, less ex­posed shel­ter than myths kin­ship sys­tems, lan­guages, sex­u­al­ity, or de­sire; what is be­ing be­wailed is the pos­si­bil­ity of re­an­i­mat­ing through the pro­ject, the work of mean­ing, or the move­ment of to­tal­i­sa­tion, the in­ter­play of ma­te­r­ial de­ter­mi­na­tions, rules of prac­tice, un­con­scious sys­tems, rig­or­ous but un­re­flected re­la­tions, cor­re­la­tions that elude all lived ex­pe­ri­ence; what is be­ing be­wailed, is that ide­o­log­i­cal use of his­tory by which one tries to re­store to man every­thing that has un­ceas­ingly eluded him for over a hun­dred years. All the trea­sure of by­gone days was crammed into the old citadel of this his­tory; it was thought to be se­cure; it was sec­u­larised; it was made the last rest­ing-place of an­thro­po­log­i­cal thought; it was even thought that its most in­vet­er­ate en­e­mies could be cap­tured and turned into vig­i­lant guardians. But the his­to­ri­ans had long ago de­serted the old fortress and gone to work else­where; it was re­alised that nei­ther Marx nor Nietzsche were car­ry­ing out the guard du­ties that had been en­trusted to them. They could not be de­pended on to pre­serve priv­i­lege; nor to af­firm once and for all - and God knows it is needed in the dis­tress of to­day - that his­tory, at least, is liv­ing and con­tin­u­ous, that it is, for the sub­ject in ques­tion, a place of rest, cer­tainty, rec­on­cil­i­a­tion, a place of tran­quil­lised sleep.

At this point there emerges an en­ter­prise of which my ear­lier books Histoire de la folie (Madness and Civilisation), Naissance de la clin­ique, and Les Mots et les choses (The Order of Things) were a very im­per­fect sketch. An en­ter­prise by which one tries to mea­sure the mu­ta­tions that op­er­ate in gen­eral in the field of his­tory; an en­ter­prise in which the meth­ods, lim­its, and themes proper to the his­tory of ideas are ques­tioned; an en­ter­prise by which one tries to throw off the last an­thro­po­log­i­cal con­straints; an en­ter­prise that wishes, in re­turn, to re­veal how these con­straints could come about. These tasks were out­lined in a rather dis­or­dered way, and their gen­eral ar­tic­u­la­tion was never clearly de­fined. It was time that they were given greater co­her­ence or, at least, that an at­tempt was made to do so. This book is the re­sult.

In or­der to avoid mis­un­der­stand­ing, I should like to be­gin with a few ob­ser­va­tions.

- My aim is not to trans­fer to the field of his­tory, and more par­tic­u­larly to the his­tory of knowl­edge (connaissances), a struc­tural­ist method that has proved valu­able in other fields of analy­sis. My aim is to un­cover the prin­ci­ples and con­se­quences of an au­tochtho­nous trans­for­ma­tion that is tak­ing place in the field of his­tor­i­cal knowl­edge. It may well be that this, trans­for­ma­tion, the prob­lems that it raises, the tools that it uses, the con­cepts that emerge from it, and the re­sults that it ob­tains are not en­tirely for­eign to what is called struc­tural analy­sis. But this kind of analy­sis is not specif­i­cally used;

- my aim is most de­cid­edly not to use the cat­e­gories of cul­tural to­tal­i­ties (whether world-views, ideal types, the par­tic­u­lar spirit of an age) in or­der to im­pose on his­tory, de­spite it­self, the forms of struc­tural analy­sis. The se­ries de­scribed, the lim­its fixed, the com­par­isons and cor­re­la­tions made are based not on the old philoso­phies of his­tory, but are in­tended to ques­tion tele­olo­gies and to­tal­i­sa­tions;

- in so far as my aim is to de­fine a method of his­tor­i­cal analy­sis freed from the an­thro­po­log­i­cal theme, it is clear that the the­ory that I am about to out­line has a dual re­la­tion with the pre­vi­ous stud­ies. It is an at­tempt to for­mu­late, in gen­eral terms (and not with­out a great deal of rec­ti­fi­ca­tion and elab­o­ra­tion), the tools that these stud­ies have used or forged for them­selves in the course of their work. But, on the other hand, it uses the re­sults al­ready ob­tained to de­fine a method of analy­sis purged of all an­thro­pol­o­gism. The ground on which it rests is the one that it has it­self dis­cov­ered. The stud­ies of mad­ness and the be­gin­nings of psy­chol­ogy, of ill­ness and the be­gin­nings of a clin­i­cal med­i­cine, of the sci­ences of life, lan­guage, and eco­nom­ics were at­tempts that were car­ried out, to some ex­tent, in the dark: but they grad­u­ally be­came clear, not only be­cause lit­tle by lit­tle their method be­came more pre­cise, but also be­cause they dis­cov­ered — in this de­bate on hu­man­ism and an­thro­pol­ogy — the point of its his­tor­i­cal pos­si­bil­ity.

In short, this book, like those that pre­ceded it, does not be­long — at least di­rectly, or in the first in­stance — to the de­bate on struc­ture (as op­posed to gen­e­sis, his­tory, de­vel­op­ment); it be­longs to that field in which the ques­tions of the hu­man be­ing, con­scious­ness, ori­gin, and the sub­ject emerge, in­ter­sect, min­gle, and sep­a­rate off. But it would prob­a­bly not be in­cor­rect to say that the prob­lem of struc­ture arose there too.

This work is not an ex­act de­scrip­tion of what can be read in Madness and Civilisation, Naissance de la clin­ique, or The Order of Things. It is dif­fer­ent on a great many points. It also in­cludes a num­ber of cor­rec­tions and in­ter­nal crit­i­cisms. Generally speak­ing, Madness and Civilisation ac­corded far too great a place, and a very enig­matic one too, to what I called an experiment”, thus show­ing to what ex­tent one was still close to ad­mit­ting an anony­mous and gen­eral sub­ject of his­tory; in Naissance de la clin­ique, the fre­quent re­course to struc­tural analy­sis threat­ened to by­pass the speci­ficity of the prob­lem pre­sented, and the level proper to ar­chae­ol­ogy; lastly, in The Order of Things, the ab­sence of method­olog­i­cal sign­post­ing may have given the im­pres­sion that my analy­ses were be­ing con­ducted in terms of cul­tural to­tal­ity. It is mor­ti­fy­ing that I was un­able to avoid these dan­gers: I con­sole my­self with the thought that they were in­trin­sic to the en­ter­prise it­self, since, in or­der to carry out its task, it had first to free it­self from these var­i­ous meth­ods and forms of his­tory; more­over, with­out the ques­tions that I was asked, with­out the dif­fi­cul­ties that arose, with­out the ob­jec­tions that were made, I may never have gained so clear a view of the en­ter­prise to which I am now in­ex­tri­ca­bly linked. Hence the cau­tious, stum­bling man­ner of this text: at every turn, it stands back, mea­sures up what is be­fore it, gropes to­wards its lim­its, stum­bles against what it does not mean, and digs pits to mark out its own path. At every turn, it de­nounces any pos­si­ble con­fu­sion. It re­jects its iden­tity, with­out pre­vi­ously stat­ing: I am nei­ther this nor that. It is not crit­i­cal, most of the time; it is not a way of say­ing that every­one else is wrong. It is an at­tempt to de­fine a par­tic­u­lar site by the ex­te­ri­or­ity of its vicin­ity; rather than try­ing to re­duce oth­ers to si­lence, by claim­ing that what they say is worth­less, I have tried to de­fine this blank space from which I speak, and which is slowly tak­ing shape in a dis­course that I still feel to be so pre­car­i­ous and so un­sure.

Aren’t you sure of what you’re say­ing? Are you go­ing to change yet again, shift your po­si­tion ac­cord­ing to the ques­tions that are put to you, and say that the ob­jec­tions are not re­ally di­rected at the place from which you, are speak­ing? Are you go­ing to de­clare yet again that you have never been what you have been re­proached with be­ing? Are you al­ready prepar­ing the way out that will en­able you in your next book to spring up some­where else and de­clare as you’re now do­ing: no, no, I’m not where you are ly­ing in wait for me, but over here, laugh­ing at you?”

What, do you imag­ine that I would take so much trou­ble and so much plea­sure in writ­ing, do you think that I would keep so per­sis­tently to my task, if I were not prepar­ing — with a rather shaky hand — a labyrinth into which I can ven­ture, in which I can move my dis­course, open­ing up un­der­ground pas­sages, forc­ing it to go far from it­self, find­ing over­hangs that re­duce and de­form its itin­er­ary, in which I can lose my­self and ap­pear at last to eyes that I will never have to meet again. I am no doubt not the only one who writes in or­der to have no face. Do not ask who I am and do not ask me to re­main the same: leave it to our bu­reau­crats and our po­lice to see that our pa­pers are in or­der. At least spare us their moral­ity when we write.”