Bentham and Foucault on Biopolitics and Political Epistemology

I published “Once Upon A Time in America: Retelling the Tale of New Federalism” at Isonomia Quarterly (here). Most of the material has previously appeared at Digressionsnimpressions. In addition, together with Sandrine Bergès I published (here) “Against power” on Sophie de Grouchy’s relatively radical republicanism in Aeon.

Today’s post is a long read. The pay-off is that I show that the state’s essential role is the production and diffusion — as a machinery of record — of knowledge for Bentham. In fact, it is the main exception to laissez-faire. And so somewhat surprisingly a certain conception of political epistemology is central to Bentham’s art of government.* Oddly, Foucault seems to have grasped this while we have no reason to believe he read the salient source, while some of the most informed readers of Bentham (Viner, Keynes, Halévy) botch the argument.

The importance of this goes beyond Bentham and Foucault. That the state is conceived as a machinery of record is central to the art of government in Smith and J.S. Mill (neither of whom is a purist about laissez-faire), and this is well understood through Lippmann. Nick Cowen and I have tried to revive the significance of this to liberalism recently in our paper (here), “Novel Externalities.”

So much for brief summary. The details are great fun.

In the first lecture of The Birth of Biopolitics (hereafter: BoB), 10 January 1979, Foucault claims that in the middle of the eighteenth century there is a change in what he calls ‘modern governmental reason.’ This change “consists in establishing a principle of limitation that will no longer be extrinsic to the art of government, as was law in the seventeenth century, [but] intrinsic to it: an internal regulation of governmental rationality.” (p. 10 in the Graham Burchell translation) He lists four characteristics of such an internal regulation.

Fourth, this de facto, general limitation, which is effectuated in terms of governmental practice itself, will establish, of course, a division between what must be done and what it is advisable not to do. It will mark out the limit of a governmental action, but this will not be drawn in the subjects, the individuals-subjects directed by government…but in the very domain of governmental practice, or rather within governmental practice itself, between the operations that can be carried out and those that cannot, between what to do and the means to use on the one hand, and what not to do on the other. The problem, therefore, is not: Where are the basic rights, and how do they separate the domain of fundamental freedom from the domain of possible governmentality? The dividing line is established between two sets of things that Bentham listed in one of his most important texts (to which I will try to return): the division between the agenda and the non-agenda, between what to do and what not to do. (pp. 11-12)

Foucault intimates here a distinction that he develops at much greater length throughout BoB between two streams of liberalism, one concerned with rights (originating with Rousseau and Kant leading toward Ordoliberalism and Hayek) and the Benthamite-radical one that originates in Hume and leads toward Chicago economics (see here). According to the editors of BoB, Foucault is referring to Bentham’s “Method and Leading Features of an Institute of Political Economy (including finance) considered not only as a science but as an art (1800–1804).” This is a text that is usually known as the Institute of Political Economy.*

However, Foucault never tells us what text he has in mind; but he does explicitly return to the division between the agenda and the non-agenda once more on 14 February, in lecture 6, in his discussion of Ordoliberalism. He writes

As you know, broadly speaking the problem of the liberalism of the eighteenth century and the start of the nineteenth century was to distinguish between actions that must be taken and actions that must not be taken, between domains in which one can intervene and domains in which one cannot intervene. This was the distinction between the agenda and the nonagenda.--(p. 133)

The distinction is clear enough. And the distinction lurks throughout his analysis of Ordoliberalism (and its reception by the French in the middle of twentieth century (see, e.g., lecture 8, 7 march, 1979, p. 195 for an explicit echo). But it is also fair to say that Foucault never explains why Bentham’s text is so important, and which one he has in mind..

I happen to think that the text is very important, but before we get there a brief detour. Keynes discusses the very Benthamite distinction between agenda/non-agenda in “The End of Laissez-faire” which is included in his (1926) Essays in Persusasion:

We cannot, therefore, settle on abstract grounds, but must handle on its merits in detail, what Burke termed "one of the finest problems in legislation, namely, to determine what the State ought to take upon itself to direct by the public wisdom, and what it ought to leave, with as little interference as possible, to individual exertion." We have to discriminate between what Bentham, in his forgotten but useful nomenclature, used to term Agenda and Non-Agenda, and to do this without Bentham's prior presumption that interference is, at the same time, "generally needless" and "generally pernicious." Perhaps the chief task of Economists at this hour is to distinguish afresh the Agenda of Government from the Non-Agenda; and the companion task of Politics is to devise forms of Government within a Democracy which shall be capable of accomplishing the Agenda.

Keynes, who was a great historian of economics, treats the distinction as out of date (but still useful). He does not explain Bentham’s argument against interference. But goes on to illustrate how the distinction can be made useful. He does tell us the source of the distinction in a note: Bentham’s Manual of Political Economy (hereafter: Manual) published posthumously, in Bowring's edition (1843; see here), probably written in the early 1790s when Bentham was knee deep in engaging with Adam Smith. As Stark notes there are important differences between the Institute of Political Economy and the Manual, but the distinction between agenda/non-agenda is in both. Somewhat frustratingly there is no direct evidence that Foucault annotated either so we can’t be sure he read Bentham directly; see here (thank you Vincent Carret—I love checking these). However, Foucault did read and make annotations on Keynes’ essay (see here). I am happy to pay anyone who can transcribe this for me!

Jacob Viner, who was also a great historian of economics and who had by then left The University of Chicago for Princeton, does explain Bentham’s argument against interference accurately in "Bentham and JS Mill: The utilitarian background," The American economic review 39.2 (1949). All of Viner’s citations are from a Manual. (Sadly no annotations from Foucault on Viner.) I quote the relevant passage.

Bentham presents three grounds for the general rule against governmental activity of a positive kind in the economic field: (1) in this field, individuals know their own interest better than government can; (2) individuals operate more ardently and more skillfully in pursuit of their own interests than government can or will operate on their behalf; (3) governmental intervention means coercion, either directly or indirectly through taxation, and coercion involves "pain" and therefore is an evil.—p. 370

It’s worth noting that (1-2) are quite indebted to Adam Smith. I doubtSmith would claim anything like (3) although Smith did think that government intervention involved violence of some sort (see my paper with Pack).

Viner’s wording (centered on (1-3)) also overlooks and so contrasts with (4) Smith’s moral-inflected argument (articulated against Colbert’s mercantilism) for “allowing every man to pursue his own interest his own way, upon the liberal plan of equality, liberty and justice.” (WN 4.9.3) The contrast between Viner’s Bentham and my Smith is sufficiently large that I wanted to check out Bentham’s own words.

Interestingly enough, when Milton Friedman (recall) addressed this very issue (in a rare paper on Marx and Mill) in the same year — Viner and Friedman were both celebrating a Mill centenary — for Friedman Bentham’s choice for laissez-faire was an effect of the state in “Bentham’s time” being “a corrupt, inefficient instrument whose enactments were widely held in low repute.”

Viner introduces his analysis with the following claim: “The general lines of Bentham's thought were wholly of the eighteenth century, as I could demonstrate if there were time.” [We could call this the Fermat move.] Viner continues, “Of English intellectuals who have had great influence, Bentham was perhaps the least original in his stock of general ideas, but clearly the most original in finding means and devices for putting his philosophy to practical use.” This judgment on Bentham echoes (recall) Oakeshott’s (1932) “A New Bentham:” “Bentham was, in all respects, a typical eighteenth century philosophe.” And it anticipates Schumpeter’s famous treatment of Adam Smith. Somebody should explore this mid twentieth century eagerness to insist that the dead folk we know by name were the un-originals. The common source to the judgments (of Viner and Oakeshott on Bentham) is also pretty clear. It’s Halévy’s The Growth of Philosophic Radicalism (which Foucault clearly did read, and I think is rather important to Foucault). I found the relevant passage in the English translation:

Political economy, Bentham tells us, includes a science and an art; and the science must be taken to be closely subordinated to the art. In conformity with the principle of utility, the end to be, pursued in all branches of the art of legislation should be to produce the maximum of happiness, during a given time, in the society in question. In other words, Bentham defines political economy as Adam Smith had defined it, as a ' branch of the ‘art of legislation', the knowledge of the best use to be made of the national wealth, and of the means necessary to produce ‘that maximum of happiness, in so far as this more general end is promoted by the production of the maximum of wealth and the maximum of population.’ But Adam Smith had prefaced his researches in political economy by three books of purely theoretical research into the conditions of the production and distribution of wealth. Bentham, on the contrary, completely neglects this preliminary part of the task of the economist. His one anxiety is to find a useful application of the theories, and so it is the solution of a practical problem which he demands of Adam Smith. He adopts his liberalism.—pp. 107-108 in Mary Morris’ English (1928) translation.

I quote the passage in full because of its emphasis on the art of economics. This turns out to be the art of legislation. And this identification clearly shapes Foucault’s account of the art of government. In context Halévy cites Bentham’s Manual (and it is just about the only place where Halévy cites it.) Bentham does use ‘art of government’ elsewhere but then he means something different than Foucault emphasizes.

To be sure, Halévy isn’t wrong about the contents of Bentham’s position here (although unfair about Bentham’s neglect of theory). In particular, he conveys accurately how for Bentham wealth and population growth are intimately linked. This connection clearly motivated Foucault to conceive of his lectures in terms of liberalism’s ‘art of government’ and political epistemology of populations. (That’s part of the ‘biopolitics’ of the title.) This is very clear in Foucault’s “Course Summary” (e.g, “How can the phenomena of “population,” with its specific effects and problems, be taken into account in a system concerned about respect for legal subjects and individual free enterprise?” (p. 336)) But the lectures he delivered actually don’t make the connection between the political epistemology and population all that clear. About that some other time more.

In addition, in wider context Élie Halévy also ignores (4), Smith’s moral argument for Liberalism’s individualism (despite it being signaled in Wealth of Nations). Halévy goes on to make a disastrous claim: “But the liberal thesis seems to imply the natural identity of interests as its necessary principle, while the examination of the distribution of wealth in a society composed of labourers, capitalists and landowners reveals natural divergences of interest.” In Smith the liberal thesis implies no such thing. Unfortunately passages like this (which express late nineteenth century banalities) seems to have influenced Rawls (recall here.)

But Halévy goes on to summarize the key bit in Bentham’s Manual:

There remains knowledge. 'There are cases in which, for the benefit of the public at large, it may be in the power of government to cause this or that portion of knowledge to be produced and diffused, which, without the demand for it produced by government, would either not have been produced, or would not have been diffused.' What is the use, for example, of advancing capital to men in industry? Is it to encourage them to turn their industry to the best advantage? But what except their ignorance prevents them from doing this? The government therefore should give them information, not capital; its function is to instruct and not to lend. The activity of the government should be confined to encouraging the study of the useful sciences; it should start prizes for discoveries and experiments; it should see to the publication of the processes used in every branch of industry, and of the prices of the various products; it should protect inventors against theft and imitation: this should be the limit of governmental activity. Beyond this it should lie low, and practise what Bentham calls quietism. At the outset, Bentham had professed that in political economy he was studying not the science but the art. Now, in terms which at least seem to contradict his original declaration, he affirms that, in political economy, the science is almost everything and the art almost nothing: for the art in such matters is to know how to do nothing. p. 109

As I noted at the start of the post, the significance of Bentham’s position is that the production and distribution of knowledge is treated as an exception to laissez-faire. Halévy is right to claim that for Bentham the government’s function “is to instruct.” (Elsewhere Bentham says that the art of government consist in education.)

What’s odd about Halévy’s summary is that he subtly mispresents Bentham’s position. While it’s undoubtedly true that Bentham justifies the exception in terms of “useful sciences,” what he means by this is rather capacious:

The gain to be produced in this article by the interposition of government, respects principally the head of knowledge. There are cases in which, for the benefit of the public at large, it may be in the power of government to cause this or that portion of knowledge to be produced and diffused, which, without the demand for it produced by government, would either not have been produced, or would not have been diffused.

There is an interesting technical question lurking here how on Bentham’s view the government would be in the position to know these last two counterfactuals.

One may defend Halévy against my charge by claiming that the second sentence with its emphasis on the “benefit of the public at large” clarifies Bentham’s intentions and restricts what knowledge should be produced and distributed. And it is true that Bentham has a tendency to focus on useful technical knowledge or skills that individuals are unlikely to find profitable to pursue themselves, but do benefit society at large. But later in a Manual Bentham explicitly also writes: “we see that government must rely upon the intelligence and inclination of individuals for putting them in operation, and that nothing is necessary to be done on its part but to leave them in possession of the power, to insure to them the right of enjoyment, and to hasten the development of general knowledge.” (emphasis added) The argument here is, of course, consequentialist in character — such knowledge is, of course, indirectly also useful in the development of human capital and citizenry — but it is pursued under the heading of general knowledge. (After all, we can’t always know in advance when knowledge will be useful.)

It’s natural to say about liberalism that it treats the state as machinery of record merely as a coordination device with socially beneficial consequences. (And surely this is an important motive in Bentham.) But this misunderstands the underlying argument. In different ways, Smith, Bentham Mill, and Lippmann explicitly allow that the state teaches the truth or can impose the curriculum by which the population is educated by non-state agents even if they have slightly different motives and arguments to advance this. From the start, the liberal tradition insists that an educated public is necessary. (In this sense, it really is an Enlightenment doctrine.) That is, from the beginning, liberalism state and science lend each other and mutually enforce each other’s authority without collapsing into each other and this is central, as Foucault discerns, to its art of argument.

There is another, final oddity in Halévy’s presentation. He never mentions the distinction between agenda/non-agenda! So until we find another source, my hypothesis is that alerted by Halévy and Keynes, Foucault consulted Bowring’s edition of a Manual. This relatively short work made him see the significant link of political epistemology to biopolitics within liberalism’s self-understanding. It’s completely explicit there. (Of course, Foucault had already establish this link among non-liberals like Bacon and Petty.) But when Foucault came to the Ordoliberals (who contemplated the utility of population stability even decline) and Chicago economics (which assumes a homogeneous population and keeps it distance from Malthus), he stopped emphasizing the link. Perhaps he oughtn’t have. As Melinda Cooper’s Family Values: Between Neoliberalism and the New Social Conservatism shows the homogeneity assumption of individuals is dropped when it comes to family policy.

---

*They cite the first edition of W. Stark’s critical edition (1954) Jeremy Bentham’s Economic Writings, where the text is usually referred to in the shortened version. Unfortunately my copy of that edition (the third one) has different page numbers (see here for it).