On My debts to Some Heterodox Marxist philosophy of science

One, perhaps intentional, effect of screening off the so-called ‘context of discovery’ as philosophically uninteresting, is to prevent engagement with so-called Marxist philosophy of science. So, for example, despite the significance of Neurath and Left Vienna, there is still no entry for ‘Marxist philosophy of science’ in the Stanford Encylopedia, which is a nice proxy for past trends in analytic philosophy. Boris Hessen (1893— 1936) and Henryk Grossmann (1881-1950) go almost entirely unmentioned in it. Grossman actually was associated with the early phase of the Frankfurt school, but they also tend to ignore him.

A recent, fascinating PhD in the history of economic theory, by Emily Evans (Cambridge), on the German reception of Marx got me interested in Grossmann. Since the fruitfulness of an author can only be judged by using their ideas for one’s ends, consider this post as a kind of assaying as well as an honoring of debts. Luckily for me some key texts are relatively easily available in a volume edited by Gideon Freudenthal & Peter McLaughlin The Social and Economic Roots of the Scientific Revolution: Texts by Boris Hessen and Henryk Grossman (Boston: Springer, 2009).

Now before I continue Hessen and Grossmann are associated with a (rather reductive) thesis that (to paraphrase) science developed in the study of then contemporary technology and that it served technological ends. Freudenthal and McLaughlin have much wise things to say to what degree either Hessen and Grossman subscribed to it. But I leave that aside in what follows.

Grossmann’s (1935) “The Social Foundation of Mechanistic Philosophy and Manufacture” (translated by Gabriella Shalit ) is a response to Franz Borkenau’s (1934) The Transition from the Feudal to the Bourgeois World picture. As Freudenthal and McLaughlin helpfully note, crucial to Grossmann’s response is what we may call a Marxist alertness to a conceptual and material distinction between two kinds of manufacture. In the Marxist narrative, the second kind is associated with Adam Smith, and so for convenience’s sake I will introduce the distinction as one between pre-Smithian and Smithian manufacture.

Pre-Smithian manufacture is characterized by the following features and now I quote Marx’s Capital 1, chapter 14 (as quoted by Freudenthal McLaughlin):

First, the decomposition of a process of production into its various successive steps coincides, here, strictly with the resolution of a handicraft into its successive manual operations. Whether complex or simple, each operation has to be done by hand, retains the character of a handicraft, and is therefore dependent on the strength, skill, quickness, and sureness, of the individual workman in handling his tools. The handicraft continues to be the basis.

So, in pre-Smithian manufacture there is specialization. But even after decomposition of the production process each step remains reliant on the skill of the craftsman. In the limiting case of pre-Smithian manufacture there is no mechanization or algorithmic routine step in the production process. Why this is so, need not concern us (so whether there is tacit knowledge that the craftsman possesses, or whether its too costly to build robots that mimic the craftsman’s skill is irrelevant).

In Smithian manufacture (or Smithian factories), by contrast, the steps in the production process can be done by machines or mechanisms, or by humans following rules without, once they are understood, the need for renewed judgment or skill. As Marx would puts it this involves, “certain simple operations of which everybody is capable.” (Of course, there are further conditions of possibility lurking here.) In the limit, a Smithian manufacture is operated by machines that design themselves and the production process. Smithian manufacture is (and sometimes I will use this as a synonym) an automated or algorithmic factory. Interestingly enough if the production process can be organized in such a way such that it is to a considerable extent deskilled, it can also be more extensive.

There is a question whether the artificial intelligence that is being envisioned in Sillicon valley now will itself be capable of the kind of skill that is presupposed in the distinction. But I set it aside today.

Lurking in the contrast between pre-Smithian manufacture and automatic factories is also a contrast between the qualitative (skill/judgment) and quantitative (precise rules, etc.). At times Grossmann treats (and he purports to be following Marx here) the link between the two distinctions as conceptually rather tight. I quote a characteristic passage:

The most important characteristic of every mechanical labor is its homogeneity; the work done is always identical qualitatively and is only different quantitatively, and these differences can be exactly measured. (Descartes, in the preface to his Traite de la Mecanique (1637) presupposes such homogeneity of performance as a condition for measurability.)— p. 126

I return to this passage below. Okay so much for set up.

As regular readers know, I am quite interested in tracing out the implications of the fact that within the sciences there is an advanced division of cognitive labor (which I often describe, following Millgram, as ‘hyper-specialization.’) In fact, if we treat discovery and justification as at least a part of scientific production, then for much of its history the kind of specialization we find in science exhibits the characteristics of pre-Smithian manufacture. It is the interaction or assemblage of skilled craftspeople, or artisans.

Of course, in practice, there were always attempts to replace some of the work by machine-built instruments and machines (including at first ‘human computers’). Already Tycho’s observatory had non-trivial amount of the work routinized. In Bacon’s New Atlantis, quite a bit of the work is intended to be done by people following an algorithm. So, I don’t want to romanticize earlier science and conceptions of science as a form of artisanal knowledge. In our own time, and we are increasingly seeing modern (big data) science becoming more like Smithian, automated manufacture. But in crucial ways, even the most modern science rests on skilled labor. So, in practice, science is always, locally and globally, a hybrid between pre-Smithian manufacture and the algorithmic factory.

This suggests that in the passage above Grossmann is too quick in treating the conditions of measurability as co-extensive with the use made by the output of measurement in say, discovery or theory development (or teaching). There is a time to ‘Shut up and calculate,’ but even in quantum physics it hasn’t made experimental or theoretical skill dispensable.

Now, in refuting Borkenau, Grossmann wants to argue that Descartes’ natural philosophy (and Hobbes’) is modelled on and shaped by an attempt to understand and improve (different kinds of) machines as such, but not (as Borkenau claims) by reflection on the nature of the division of labor (or manufacturing production). The significance of this is that for Descartes (and Hobbes) the art of government doesn’t become the social organization of different specialized tasks, but the proper integration of and smooth functioning of different machines whose parts are fairly homogeneous (see p. 141 with a reference to preface of Hobbes’ De Cive).

I could end today’s post here.

Freudenthal and McLaughlin also present another, previously unpublished (but conventionally dated to 1946) work by Grossmann “Descartes and the Social Origins of the Mechanistic Concept of the World.” This essay starts by reiterating that given “Descartes’ general tendency to give a mechanistic interpretation of all fields of science, of all phenomena of physical, organic and psychic nature, it is natural to suppose – and this surmise constitutes the basic idea of the present essay – that Descartes also applied the same mechanistic principles to his method, to the very structure and functions of his scientific apparatus, and conceived them upon the model of machines.” (p. 158) But then Grossmann makes explicit the reflexive nature of this point (which is foundational to what I have called Hobbes’ ‘methodological analytic egalitarianism’ (or MAE)), if we are all machines then, for Descartes, we’re fundamentally alike: “The universal science was to be universal not only in that it would be applicable in all the fields of science, but also in that it would be accessible to all the people, including the large masses.” (p. 159; in context, Grossmann explicitly denies that Spinoza adheres to this Cartesian embrace of MAE.) So far so good.

But then Grossmann argues the following (with some references to Bacon):

During the predominance of handicraft and the beginning of [pre-Smithian] manufacture, skilled specialization and the individual worker’s virtuosity in a limited trade set the standard; with the emergence of automatically working machines in industry it became clear that these machines, independently of and without any handicraft training or personal talent of the workers, could perform the work better, in greater quantities and with greater speed; and that this work could be done by anyone who knew how to handle the machine by simple manipulations, by women and children, indeed even by idiots and cripples, because the automatism of the machines simplified the operations so drastically.

Under the influence of this fundamental technological and social revolution – the transition to large-scale industry and machine production – the conviction grew in the most advanced minds that a new epoch had begun; that specialization was a thing of the past that was indispensable only so long as no adequate instruments were available, but that it was no longer necessary in the machine age. This conclusion, drawn from the experience of industrial production, was then extended by a process of generalization to intellectual production, to the sciences, and the idea asserted itself that specialized individual talent was of decisive importance only at the lower level of development of human society when mankind still lacked adequate intellectual instruments, but that later, at higher social levels, even average human intelligence would enable every man actively to participate in the intellectual work of the nation and to attain the knowledge of highest truths if he only knew how to use the proper “instruments.” For the intellectual auxiliary means of a systematically conceived procedure, the method, was assigned the same part in the intellectual production that the machine, that technical auxiliary means played in industrial production. p. 163.

Not unlike Heideggerians and Straussians of the age, there is a tendency here to assimilate Bacon and Descartes into an integrated project. So, Grossman attributes this vision (somewhat implausibly) to Bacon and then the development of Descartes’ philosophy through Descartes’ life is used (in much more detail) as a way of illustrating this historical thesis. As a way of organizing reflection on the significance of Descartes’ philosophy for social theory this is with considerable interest. (So, I return to the details another time.)

What’s neat about this passage is that Grossmann develops a stadial account of social development itself organized around the distinction between pre-Smithian manufacture and Smithian factories. On this view it’s the growing significance of Smithian factories — in which human nature (qua worker) is treated as fundamentally homogeneous — that provides the template for a new kind of image of science. (Somewhat amusingly, Grossmann here draws on features of Borkenau’s project he had criticized before.) Rather than treating science as a highly differentiated pre-Smithian manufacture populated with highly unequal artisans, it can become like the newly developing Smithian factories and be populated by undifferentiated and homogenous human capital. And a universal method and a universal language are the instruments that will make human capital do its work in Smithian factories of science as developed by Descartes.

As an aside, Grossmann could have strengthened his argument if he had emphasized that the underlying matter that composed the machinery of nature (and the workers that investigated and shaped nature) is itself surprisingly homogeneous in Descartes (and Boyle). (Newton’s abstract quantity, mass, is a thinner version of the same aspiration.) There is, thus, a considerable standardization presupposed in mechanical science.

The point of a universal language (of ideas) is really three-fold: (i) economize on the time needed to grasp/learn/teach knowledge “So that it could be taught in a very little time” (Descartes to Mersenne, Nov 20, 1629); (ii) provide order “according to definite classifications” and so make it easily available to others (p. 165) — this also overcomes the natural tendency toward secrecy and esotericism among artisans —, that is to facilitate uptake of known truths; (iii) and, thereby, undermine intellectual (and perhaps political) elitism: “I hold that this language is possible and that one can find the science on which it depends, with whose help the peasants would be able to judge of the truth of things better than the philosophers do now.” (p. 166; Grossmann quoting Descartes)

Some other time I will have more to say about all of this as an interpretation of Descartes. (There is more in Grossmann.) But I want to tie some threads together. For Grossmann helps explain something what might otherwise be thought puzzling about Descartes. If one looks at Descartes’ tree of knowledge, one finds that “Unlike the Thomists and Jesuits who think about the sciences as different disciplines with separate subjects, Descartes attempts to extend his successes in mathematics and the middle sciences into all of philosophy.” (I am quoting Ariew’s (1992) useful summary (pp. 110-111).) So, Descartes does not have to create all kinds of mechanisms to handle the negative externalities and challenges (as diagnosed already by Plato in the Republic) that follow from hyper-specialization. The method and language of science will maintain unity.

As Ariew makes clear by the time of the philosophes and their Encyclopedia Descartes’ project (which confusingly for us they called “a genealogical order”) was explicitly rejected (p. 103). (They were not so willing to acknowledge that this was a kind of vindication of the Thomists and Jesuits of sorts.) In fact, the embrace of Encyclopedism, which is the collecting, editing, repackaging of existing knowledge and disseminating it in an organized fashion, is a recurring temptation when Cartesian-like attempt at eliminating the effects of hyper-specialization fails.