The polemic over the number of deaths inflicted on Algerians by the Paris police in October 1961, we have argued, tends to dominate and to obscure ‘the wider significance of the event which is of considerable historical interest and importance’.Footnote ¹ It is therefore with some reluctance that we find ourselves being drawn back on to this terrain, and we would much rather engage in debate over some of the broader issues that are now being investigated. For example, Paris 1961 explores in detail the longue durée colonial roots of the techniques of police repression that Maurice Papon introduced into the capital from North Africa, as well as the structures of state violence that made such a tragedy possible. Half of the book is addressed to areas that have been barely touched upon by historians: the processes by which this event was expunged from public visibility and the complex politics that made this possible linked to tensions within and between the mainstream left, the radical anti-colonial movements, and the Front de Libération Nationale (FLN); the ‘underground’ survival of social memories within the Algerian communities and French left; and the conditions under which October 1961 re-surfaced after the 1980s to become a key issue in contemporary French political and cultural life.

Jean-Paul Brunet's reductive analysis of Paris 1961 is emblematic of the narrow focus more generally of his work on October 1961 that concentrates almost exclusively on the French police and the FLN, and ignores these other themes and chronologies of social and political history that also represent key strands to the story of October 1961. Yet Brunet lays claim to a monopoly of historical contextualization and judges his own work to have ‘situate[d] the police repression in a historical perspective. This is in contrast to the majority of authors who have simply denounced the prefect of police Maurice Papon and his subordinates on moral grounds.’Footnote ² Brunet insists on pulling the debate back again, as he did in his attack on Jean-Luc Einaudi, author in 1991 of a ground-breaking work on the events of 17 October, whom he derides as a ‘weekend historian’, ‘hagiographer of the FLN’, and ‘exposes’ as a supporter of repressive regimes, thereby attempting to refuse Einaudi the right to denounce the police violence of 17 October.Footnote ³ Now, for daring to offer a different interpretation of events from that of Brunet, it is our turn for similar polemical treatment, with accusations of bias in the face of ‘objective’, ‘neutral’ scientific enquiry that Brunet claims to represent, indeed symbolize.

We could readily reply in close detail, blow by blow, to Brunet's charge sheet, but do not think that this would be very edifying or informative. Instead, we will try and shift the discussion from the ad hominem mode and address the two substantive issues that lie at the centre of his polemic: the interesting methodological challenge faced by the historian in relation to the kinds of evidence that might be used to address such a complex and contested field as police mass killings in a European capital, and the problems of interpretation that they raise. Secondly, and unavoidably, the much contested issue of the number of fatalities.

The H series (Algerian War) of the Archives of the Paris Prefecture of Police (APP) provides a huge and complex body of material in relation to the events of October 1961 which present a particular challenge to the historian. Brunet expresses surprise at our supposed claim that these police sources are so radically falsified that they present, to quote us, an ‘impossible task of trying to uncover the truth’, so that we are compelled to fall back on ‘inferior’ oral evidence.Footnote ⁴ But this is not what we wrote, and by lifting a phrase from its context, Brunet deploys that age-old method of attack which consists of falsifying an author's meaning, and then subjecting this Aunt Sally to an-all-too easy ‘demolition’. What we actually wrote was as follows: ‘Brunet notes, quite correctly, that the [APP] archives are of great interest to the historian since many of the documents were for internal use and reveal a surprising amount of reliable information on the concealed workings of the Prefecture apparatus.’ Then the passage, from which Brunet lifted the phrase which is here italicized: ‘In this world of mirrors and obfuscation, of propaganda and counter-propaganda, the historian would seem to be faced with the impossible task of trying to uncover the truth through police archives that were falsified in their very constitution.’Footnote ⁵

This is the first of a sequence of misleading or distorted references to our book. Space does not allow a detailing of all these misrepresentations, but quite typical is Brunet's rebuttal of our claim that he rushed into print since we incorrectly claim that his permission (dérogation) for access to the police archives was granted on 29 March 1999, only six months prior to the publication of his book.Footnote ⁶ However, we correctly state the APP dérogation as 26 May 1998, and Brunet has deftly switched this with the date of a second dérogation granted for the judicial archives on 29 March 1999.Footnote ⁷ It is somewhat ironic that Brunet, in a long homily on the need for professional skill in the close analysis of documentary sources and evidence, should succumb to such simple errors, and it casts a shadow over the reliability of his analysis of the far more complex materials contained in the archives.

However, Brunet's preliminary intervention does introduce the key issue of the nature of the Prefecture record and how this can be ‘read’ by the historian. The Paris police force indubitably engaged in a massive and systematic cover-up which operated at every level, just as was the case within the judicial and governmental hierarchies and in the armed forces in Algeria which ‘gangrened’ metropolitan practices and the rule of law.Footnote ⁸ The APP archive is intriguing, since close analysis of the huge volume of telegrams, directives, and reports is revealing of how these mechanisms of deceit operated, including the falsification of reports relating to individuals who were shot or beaten to death by police officers.

Papon, as we know from his prosecution at Bordeaux in 1997–8 for crimes against humanity, was a past-master in the techniques of bureaucratic deceit. One minor but typical example: on the night of 17–18 October 1961 the riot police near the Bonne-Nouvelle Metro killed several Algerian demonstrators, but made the bavure or ‘mistake’ of also clubbing to death an ‘innocent’ French bystander, Guy Chevalier, as he entered the Rex Cinema. Within hours the head of the Police Judiciaire informed Papon's cabinet by telex that the Cochin Hospital had certified death by severe head-wounds caused by police batons, ‘trauma of the neck and vertebra’. Papon later referred to Chevalier's death as that of a ‘European civilian whose heart – it would appear – gave way under the shock’.Footnote ⁹ That Papon used this cynical formulation was no accident, since examination of the APP archive relating to the violent death of Algerians shows how a quite standard procedure at the police-run morgue (IML) was to make a report, in instances where the victim had been beaten or crushed to death by the police, of ‘death by natural causes’ which then led to an official decision to file away the case without further police or judicial investigation.Footnote ¹⁰

But ‘death by natural causes’ was just one of a whole gamut of techniques of concealment, and the police archives enable the historian to locate the standardized procedures that operated at every level of the Prefecture of Police – a huge and complex force that employed some 29,000 people – with the open or tacit support of Papon and his officers. Progressing up the paper-trail from the act of murder itself (officers killing arrested Algerians at night, on isolated waste ground, removing identifying documents, tipping them into rivers and canals so as to destroy forensic evidence), via the planting (by specific order of Papon) of weapons on the bodies of Algerians gunned down on the street, these practices were complemented by the various reports on the later recovery of corpses (cause of death falsely entered as victim of a fratricidal killing, death by ‘accident’ or ‘natural causes’), and were reinforced by magistrates entrusting further judicial investigation back into the hands of the police, resulting in the latter's ‘inability’ to locate identifiable accused officers, year-long delays in paperwork, and false versions or alibis concerted by groups of police. Such techniques continued up through the higher echelons of the establishment, including the issue of misleading or false press releases, seizure of publications that included incriminating evidence, the blocking of parliamentary commissions of investigation, the opening of bogus legal procedures against ‘X’ (person(s) unknown) so as to prevent political debate through the formulaic ‘the matter is now in the hands of the courts’, to the eventual amnesties post-1962 that immediately stopped all further judicial investigation.

Working independently from us, Alain Dewerpe, in his monumental study of the Charonne massacre of 8 February 1962 when the police bludgeoned or crushed to death nine French anti-war protestors, has arrived at similar conclusions. Dewerpe has been able to identify not only the same individuals and police units at work as on 17 October, but an identical and well-practised modus operandi of systematic cover-up, from ministerial rank down through Papon to the lower levels of the force. Dewerpe shows not only how brutal violence was built into police training and procedures, but also how the police and judicial archives reveal ‘a system of dissimulation’, the fabrication of an official ‘truth’ or ‘mensonge d’État', and how the complex battery of legal procedures was deployed to cover up the murders so that the cases could be eventually filed away as lacking in evidence (sans suite) without any police suspects (although clearly identifiable) ever being brought to justice.Footnote ¹¹

Jean-Paul Brunet has to agree that such processes of cover-up existed. Where we differ from him is in his conviction that the falsified record can still be used alone to arrive at an accurate reconstruction of the events and a true count of the number of individual Algerian fatalities. Most surprising in Brunet's overall work (and he repeats this once again in this present critique) is his failure or refusal to extend his investigation beyond police and official or state sources.Footnote ¹² If we want to know what happened to the hundreds, indeed thousands, of Algerians who were subjected to extreme police violence it might be useful to include the victims' testimony, as has been the case for investigations of countless massacres or genocidal acts over the last century, from the Holocaust to ‘Bloody Sunday’ in Londonderry. For Paris in 1961 there exists not only a rich and voluminous oral evidence, but also several hundred contemporary accounts by Algerian participants written, often in biro in school text-books, within days of 17 October, as well as eye-witness accounts by many dozens of French observers, sources which we have duly used to build up a multi-faceted interpretation of the events.

On the question of oral sources, Brunet, however, is highly inconsistent regarding our work, accusing us of both their over and under-use.Footnote ¹³ Brunet fails to see that methodologically we use oral sources as but one source amongst others:Footnote ¹⁴Paris 1961 accords oral sources no lesser or greater reliability than written sources, whereas for Brunet, oral sources remain evidence that one uses faute de mieux, even if he appears to change position from his two books by now recognizing that oral evidence is of ‘central importance’. However, Brunet then immediately backs this up by still further references to interviews with police officers, not with Algerians.Footnote ¹⁵ His altogether unconvincing reason for not interviewing any Algerian demonstrators was, Brunet states, because Einaudi had told them not to co-operate with him.Footnote ¹⁶ But now he claims it is because the methodology of oral history is too problematic such that his carrying out such a project would have encountered great difficulties, with witnesses clamming up in relation to such a touchy subject.Footnote ¹⁷ This is a difficulty that does not seem to apply to police witnesses.

But there is more than one way of skinning a cat, and Brunet then proceeds to devalue all Algerian testimony (written and oral) by showing its supposed unreliability and pro-FLN bias, or fantastic delusional structures (the terror of being ‘gassed’).Footnote ¹⁸ We accept Brunet's criticism of our claim that he failed to make use of the eighty individual reports by Algerian demonstrators in the APP archive.Footnote ¹⁹ He does indeed quote at length from one of these,Footnote ²⁰ but for his purposes he has selected the one statement that is evidently delusional and problematic, and ignored the other seventy-nine deposited in the APP as well as the hundreds of other witness statements filed in the archives of the FLN Fédération de France, all of which tell a rather different story.

Brunet, however, divides in a Manichaean way witnesses into two groups, police and state officials on the one hand (ultimately reliable), and all Algerian nationalists and French anti-war opponents of police repression on the other (ultimately false). In the case of the former, in spite of his acceptance of cover-up and fabrication, for Brunet their evidence can, in the final analysis, be trusted because of their status and professional ethic. Oral evidence from his interview with the former army minister Pierre Messmer is credible since ‘he is an honest gentleman and trustworthy’; autopsy reports cannot be falsified since this would constitute an inconceivable, unprofessional act; Papon's revelations to the Conseil général de la Seine are reliable since ‘it was not his manner to purely and simply invent facts’;Footnote ²¹ the inspecteur général Roger Wuillaume who, in his investigation and report on 17 October, provided a classic cover-up, for Brunet ‘undertook his work in a conscientious way’;Footnote ²² and now, in Brunet's present contribution, we have the further oral testimony of senior ex-police officers, Jacques Delarue and Pierre Roux, ‘whose word is unimpeachable’. Delarue and Roux note that the police could not have concealed the true number of Algerian deaths since this was going against ‘long established police culture’, telling Brunet that ‘a police officer would never have done such a thing’.Footnote ²³ Where this ultimately leads Brunet is to the claim that official police sources, leaving aside the few ‘rotten apples in the barrel’, do accurately estimate the number of Algerian victims of police violence and that we can accept ‘with little hesitation’ the figures put forward as true.Footnote ²⁴

Brunet's over-reliance on police reports, and the marginalization of alternative evidence from Algerian and French demonstrators, leads to some serious problems of misinterpretation of key events. Let us take just one example. Brunet places particular emphasis in his two books and present contribution, on the events at the Pont de Neuilly, the location of one of the most violent incidents between demonstrators and police on 17 October 1961. He argues that the police resorted to understandable violence (rather than calculated, and cold-blooded killing) in a chaotic situation at the bridge where only sixty-five officers were hugely outnumbered by thousands of Algerians and forced to use their guns in self-defence.Footnote ²⁵ Brunet conveniently leaves out the numerous other flash-points around Paris which disconfirm this hypothesis during which the police initiated violent paramilitary baton charges to disrupt peaceful marching columns of men, women, and children and inflicted numerous deaths by clubbing and gun-fire (Bonne-Nouvelle, Boulevard Saint-Germain, Saint-Sulpice …).Footnote ²⁶

That aside, we may agree that the force of CRS, riot police (Compagnies d'intervention) and harkis at the Pont de Neuilly may appear to have been numerically small. However, as the column of Algerians, including women and children, moved towards the bridge they were met, not by a thin blue line that fragmented in disarray, but by a well-positioned barrage of experienced riot police armed with deadly clubs (bidules), side-arms, and machine-guns. Brunet claims that the idea of police opening a deadly fire that stampeded the crowd into retreat does not appear to be true because the researcher Rémy Valat has now located the harki – the very man who was seen by witnesses to first open fire on the approaching crowd with a machine gun – ‘an allegation today strenuously denied by the man in question’.Footnote ²⁷ It could be added for good measure that the commander of the harkis at Pont de Neuilly, Lieutenant-Colonel Montaner, who was notoriously accused of heading the torture teams in the cellars of the Goutte d'Or, has left his own account in the extreme-right wing and negationist journal of Bernard Lugan, L'Afrique réelle.Footnote ²⁸ According to Montaner only fifteen police, surrounded by thousands of Algerians, were ‘attacked from all sides making it impossible to manoeuvre’, but showed great discipline and ‘despite the situation did not make use of their fire-arms, so avoiding a disaster, a blood-bath’.Footnote ²⁹

Whether these police witnesses are credible or not, why is it that Brunet chooses not to mention the testimony of dozens of demonstrators in order to arrive at a more balanced picture? Here we quote a few alternative perspectives: first, Khaled Benaïssa, then a young boy, found himself near the head of the march: ‘It was at this moment that I could make out, facing us, the police that had formed a barrier right across the boulevard. Some of them had a machine-gun positioned on a tripod. Shots rang out and a general panic ensued.’ Benaïssa fell and was nearly crushed to death in the fleeing crowd.Footnote ³⁰ The humanitarian worker Monique Hervo, one of the very few metropolitan French marchers, advanced with a group of Algerian mothers, carrying one of their babies. From a higher point, ‘I could clearly see lower down a barrier of dark police uniforms. The barrels of rifles or machine-guns moved in a raking action, aimed at the front line of demonstrators. Bullets were fired towards our procession. The shots increased … Panic set in among the ranks of women and children.’ This was followed by a stampede back towards Nanterre and the safety of side streets.Footnote ³¹ These sources do not suggest a police force desperately overwhelmed by Algerians.

Another witness, François Lefort, watching the Algerian column from his house, described a classic police offensive manoeuvre, a pincer movement involving a running charge and baton assault to split and panic the demonstrators.Footnote ³² Why does Brunet, who prefers police evidence, remain mute on the eye-witness account of Michel Gibier present at the Pont de Neuilly? This officer recounts how when the column, which was in no way hostile or threatening, arrived within fifty metres, the police without warning opened fire: ‘In front of him he [Gibier] saw the harkis firing on the demonstrators with automatic weapons. One of them was on one knee and fired his machine-pistol as at a shooting range. The other harkis fired from a standing position … The crowd of demonstrators surged back.’Footnote ³³ If the police were outnumbered and lost control, why did commanders decide at this juncture to pull back two CRS companies from Neuilly to central Paris?Footnote ³⁴ As a general rule Brunet always accepts police and official accounts of events, when they support his interpretation, as credible, but never the evidence of Algerians or left-wingers, which is either excluded or discredited.

Turning, finally, to the numbers debate, Brunet repeats his well-worn argument that police and judicial records alone allow a close approximation to the number of deaths caused by police violence (thirteen individual cases proven, eight likely, and four possible).Footnote ³⁵ As we point out in Paris 1961, Brunet is able to create even lower statistics by excluding from his calculation the wave of police killings that had begun during the six weeks proceeding 17 October. However, it is simply not possible to arrive at even approximate figures of fatalities using the approach to police sources adopted by Brunet: police death squads stripped many dozens of Algerians of identification, tipped them into rivers and canals so that by the time of eventual recovery of corpses individuals were no longer recognizable. These deaths were then falsely reported as being the result of internecine violence. It is therefore impossible to provide conclusive evidence of police killings, the date and time, or their location. That the police deliberately resorted to laying a false trail to disguise killings is evident from various sources: the FLN militant Amirat Slimane told the investigating magistrate Braunschweig how on his arrest in January 1961 he was taken to the harki unit at 28 Rue de la Goutte-d'Or: ‘I was first of all interrogated for half an hour by Captain Montaner. He said to me: “At the Dépôt [Prefecture cells] you are listed as having been freed and I can lodge a bullet in your brain and put you in a sack with the words ‘traitor to the FLN’.”’Footnote ³⁶

No historian would deny that there were internecine Algerian killings: indeed more Algerians – an estimated 4,000Footnote ³⁷ – died in France during the war of independence at the hands of fellow countrymen than of the police. However, by 1959–60 the FLN had nearly eliminated the rival Mouvement national algérien in Paris which was reduced to a tiny number living in fortress enclaves. On a monthly basis through 1961 there continued to be a very small number of internecine assassinations, probably of the order of ten to fifteen per month. It is not surprising that Brunet is able to discover a number of individual cases to prove that the FLN killed recalcitrant Algerians, but he makes a mistake in extrapolating from an insignificant number of exceptional cases to conclude, in agreement with the Prefecture, that the dramatic surge in killings during September–October was due to the FLN.Footnote ³⁸ The head of the government joint intelligence unit (Service de coordination des informations nord-africaines, SCINA), as well as the senior civil servant Michel Massenet, the Ministry of Justice (which received the legal reports on violent deaths), and the main police trade union, the Syndicat général de la police (SGP) were in no doubt that the surge was due to police actions, not a remarkable spate of internecine killings.Footnote ³⁹ None of the key intelligence agencies that provided weekly and monthly bulletins on FLN activity – the Service d'assistance technique aux Français-Musulmans d'Algérie (SAT-FMA), the Service de coordination des affaires algériennes (SCAA), the SCINA, or the Renseignements généraux – reported an increase in internecine killings during September to October 1961.

But, as we argue, in a simple but elegant solution that Brunet studiously avoids, the statistical surge of between 108 and 121 violent deaths in September–October, over and above the longer-term monthly average of 10 to 16 homicides (including FLN internecine killings), can be attributed to police actions.Footnote ⁴⁰ This method, which does not claim to do what is impossible – to provide a totally accurate count of individual homicides – does at least enable us to arrive at a meaningful order of scale. It also avoids the insurmountable mine-field that Brunet enters of trying to identify all causes of individual deaths from multilayered corrupt or falsified police reports. Yet Brunet is keen to equate our work with that of Jean-Luc Einaudi, whereas Paris 1961 also reserves criticism for the at times imprecise and inflationist methodology that Einaudi uses to assess the number of Algerian deaths.Footnote ⁴¹

It is surely time to move the debate on 17 October forward beyond the dispute over numbers and the issue of individual culpability (Papon as villain), rather overplayed issues that have to date dominated the controversy and blocked more fruitful lines of historical enquiry. Indeed, there are encouraging signs of studies that escape from the memory ‘battles’ linked to 17 October stemming from fundamentally opposed ways of dealing with and mobilizing the past. These studies move beyond narrow spatial or temporal aspects towards a deeper historical contextualization,Footnote ⁴² and shift the focus from fixed positions to multiple new questions or fields: these include the deeper, long-term roots of policing and state violence;Footnote ⁴³ the impact of decolonization on metropolitan France as colonial personnel ‘came home’ to influence state institutions, from housing to welfare policy;Footnote ⁴⁴ the cultural and identity politics of Algerian immigration;Footnote ⁴⁵ a widening comparative history of French end-of-empire (particularly rich in potential in relation to Britain);Footnote ⁴⁶ the influence of the Algerian War on contemporary French cultural production;Footnote ⁴⁷ and the ongoing tensions within French society as it comes to terms with the darker side of a recent colonial past.Footnote ⁴⁸