In response to:
Goldhagen in Germany from the November 28, 1996 issue
To the Editors:
Josef Joffe’s fine article of November 28, 1996, which mostly analyzes the reception of my book in Germany, does present criticisms of the argument of my book itself. Because of space limitations, I can address briefly only the most important of them.
Whatever it might mean to “indict an entire culture,” the assertion that I do it is wrong. All I do is conclude that German culture before and during the Nazi period was broadly and deeply anti-Semitic. In principle, that is as legitimate as is reporting that the culture of the white antebellum South was broadly and deeply racist against blacks.
I concur fully with Joffe that any explanation of the Holocaust must be multicausal. My book’s analysis is explicitly so, and this is particularly evident in its treatment of the evolution of the eliminationist program and of Jewish “work.” I also specifically discuss the need for a multicausal analysis in the Introduction (p. 9), in the Foreword to the German edition, and in the concluding chapter (p. 416). So it is odd that anyone would criticize me for allegedly holding a different position.
The assertion that my logic suffers from a “confusion of different levels of analysis” and “circular reasoning” (a circular argument is one whose premises assume, namely are logically dependent upon, its conclusion) is wrong. Instead, it proceeds in two ways along standard social scientific lines: hypothesis testing, which is how one can show—contrary to the implication that one cannot in principle demonstrate such a thing—that anti-Semitism was the principal motivation of the perpetrators, and inferring from a representative sample characteristics of a larger population:
(1)By analyzing the evidence regarding anti-Semitism in German society before and during the Nazi period, I concluded that the vast majority of Germans in the 1930s held a set of anti-Semitic beliefs about Jews that included the belief that Jews and Jewish power had somehow to be eliminated from German society.
(2)By investigating the perpetrators of the Holocaust, I established (a) that the perpetration of the Holocaust was carried out by a large number of Germans (at least 100,000), who came from all social backgrounds and all walks of life and (b) that many knew that they could exempt themselves from killing without suffering punishment.
(4)This hypothesis, alongside five other rival hypotheses about the source of the perpetrators’ motivation (e.g., “obedience to authority” and “social psychological pressure”), was then tested systematically against the evidence of the perpetrators’ patterns of actions and against the testimony of the perpetrators and of the victims regarding the perpetrators’ attitudes towards Jews and the program of extermination.
(5)I concluded that the only hypothesis which can account for the diverse actions of the perpetrators is the anti-Semitic hypothesis. The others are resoundingly falsified on many grounds.
(6)Further, since the perpetrators comprised a representative sample of German society, we must infer (using the same rule of inference that is the foundation of survey research) that the many other ordinary Germans who shared this anti-Semitism would likely have acted in similar ways, had they been the ones who, by chance, found themselves working in institutions of killing.
(7)This leads to (a) a recharacterization of the anti-Semitism that was dominant in German society as one which had a specific genocidal potential as a basic trait and not as an accidental, eccentric, or extremist component of it, and (b) the conclusion that this general belief structure was sufficient, when the practical opportunity arose, to motivate those who shared it to brutalize and kill Jews willingly, which however does not mean that it was a sufficient cause (other factors were necessary) for the Holocaust to occur.
So I do not use the perpetrators’ actions as the basis for my initial empirical claim that the vast majority of Germans were eliminationist anti-Semites. That was established by examining German society itself. With their actions, the perpetrators reveal that this eliminationist anti-Semitism had genocidal potential and show what it would lead its bearers to do—namely to torture and kill Jewish men, women, and children by the tens of thousands voluntarily and often with evident enthusiasm—when mobilized in the exterminatory campaign. There is nothing circular about this argument.
As yet, none of my book’s critics who have attacked my allegedly “faulty logic or overblown conclusions” to which Joffe alludes, has yet actually demonstrated that one is faulty and the other overblown, or even bothered to address the vast quantity of evidence buttressing my conclusions that demonstrates the anti-Semitism, voluntarism, zeal, and assent of those who slaughtered European Jewry. 3 That they—including all the German critics, about whom Joffe writes so well—have glaringly failed to do so despite the enormous volume of their attacks, is at the root of the German public’s embrace of my book.
Daniel Jonah Goldhagen
Josef Joffe replies:
Whatever its theoretical or logical flaws, Daniel Goldhagen’s Hitler’s Willing Executioners is an excellent and original contribution to the literature on the Holocaust. His case studies on the police battalions, the death marches, and annihilation by “work” have already changed our understanding of the Shoah and the terms of future research.
Unfortunately, the book grew out of a dissertation for Harvard’s Government Department, and so a grand thesis, or at least a compact causal model, was required. This is where the trouble began.
Has he indicted an entire culture in order to explain the Holocaust, as I wrote and he denies? Of course he did—at least the political culture that prevailed in Germany before 1945. The central thesis of the book is that the Holocaust could happen only in Germany because only the Germans—alone among the nations—were the way they were. They were infected with the deadly bacillus of “eliminationist anti-Semitism” that turned “annihilationist” when the time was ripe.
Is this a monocausal theory? That, too, Goldhagen vehemently denies—but more in the manner of an ex post facto concession to his critics. As he states in his letter, he discusses the need for a multicausal analysis in the introduction to the German version of his book. Correct, but this was four months after the US edition had triggered a spate of highly critical reviews around the world whose common denominator was the defects of Goldhagen’s monocausal approach.
Actually, there is nothing wrong per se about a monocausal theory; indeed, such constructs (“X is caused by A, and only by A”) are the most elegant of formulations because you can’t beat them for parsimony. The central problem is: How do you prove it? How do you prove that “A” (German eliminationist anti-Semitism) caused “X” (the Holocaust)?
This is where the “level of analysis” problem begins to creep in, along with the problem of circularity. In order to causally connect a cultural disposition (anti-Semitism) to specific conduct (mass murder), Goldhagen goes through a three-step process.
First, he analyzes the socioeconomic traits of the “ordinary Germans” who largely made up the Ordnungspolizei and the guard battalions of the death marches.
Second, he compares their characteristics with the sociological makeup of the German population as a whole, discovering that the sample perfectly matches the whole.
Third comes the logical leap which, for the sake of simplicity, can be rendered thus: “The killers were ordinary Germans, ergo ordinary Germans were killers.”
That, of course, is a logical no-no. You can’t reason backward from the actions of a sample to the culture as a whole. You can’t conclude from the statement that “Bonnie and Clyde were born gangsters” that “By dint of common socialization and a common lower-class background, their families were born gangsters, too.” This is precisely where the “level of analysis problem” begins to bite.
In formal terms, the “level of analysis problem” states: Behaviorally, the properties of a set are not identical with the properties of its members; therefore, you can’t draw conclusions from the properties of one level (the individual) about another (the group), or from either about the third (the culture or the nation). For example: You cannot describe the behavior of a thermic pressure system by analyzing the behavior of its molecules. You cannot explain the actions of an aggressive crowd by looking at the psychological or sociological traits of its members. You can’t conclude from the correct statement “Sociologically, the order police were a faithful microcosm of German society as a whole” that “German society as a whole consisted of ‘willing executioners.”‘
At the risk of repeating the obvious: The killers of the “order police” apparently were normal folks, but they did not operate in a normal setting. They were trained and indoctrinated; they were beholden to the usual peer group pressures; they committed their unspeakable crimes in a setting far away from their normal petty-bourgeois existence and its moral constraints. Hence, you cannot indict an entire culture by pointing to the behavior of a specific subset of that culture.
Nor can you argue, as does Goldhagen, that an endemic anti-Semitism of German culture explains the mass murder committed by 100,000 of its members (this is the number Goldhagen uses), and then turn around to use the murderous propensities of that subset to prove the “annihilationist anti-Semitism” of the entire people. That, to me, is a circular argument.
But none of these logical and theoretical complaints changes my basic judgment. Hitler’s Willing Executioners is an extraordinary and original contribution to the mountain of literature on the Holocaust. Whatever its weaknesses as causal model, it is first-rate history that has transformed the way we look at the Holocaust.
February 6, 1997
Of course, there is a further assumption of “agency,” namely that, in the broadest sense, people have the capacity to make choices. ↩
The hypothesis of ideal-normative agreement, namely that the actors’ actions were in accord with and followed upon their beliefs, is one that should, in any case, always be tested. ↩
My lengthy reply to my critics appeared in The New Republic, December 23, 1996. Dieter Pohl, in a just-released detailed study of the perpetrators of the Holocaust in East Galicia, Poland, Nationalsozialistische Judenverfolgung in Ostgalizien 1941-1944 (R. Oldenbourg, 1996), comes to conclusions about the prevalence, virulence, and central role of anti-Semitism in motivating the willing killers that support my own. ↩