Sraffa and Wittgenstein

Abstract

This paper documents the long and very intense intellectual friendship between Sraffa and Wittgenstein through their extant correspondence. It argues that most likely Wittgenstein discarded the Tractarian thesis of a ‘picture’ theory of meaning in favour of a ‘language game’ theory of meaning in the Philosophical Investigations because of Sraffa’s consistent criticism of the metaphysical foundations of all theories based on human psychology or ‘mental processes’, since no objective data for such foundations exist. It highlights the similarity between the structural relationship between objective data in Sraffa’s system of basic goods and Wittgenstein’s rules of a language game.

This chapter is a slightly edited version of the fourth chapter of my book, A Revolution in Economic Theory: The Economics of Piero Sraffa, Palgrave Macmillan. Here it is reprinted with the permission of the publisher.

Appendices

Comment

1.1 A Comment on Sinha on the Relationship between Sraffa and Wittgenstein

• John B. Davis 

Ajit Sinha’s discussion of the relationship between Sraffa and Wittgenstein (Sinha in this volume; also, 2016, pp. 92–109) notes the closeness of their friendship, the intensity of their intellectual exchanges, and Wittgenstein’s acknowledgement of his considerable debt to Sraffa for the shift in his thinking from the Tractatus to the Investigations.⁷ Sinha also notes that though there is relatively little evidence regarding the nature and extent of their exchanges, there nonetheless exists something of a consensus among scholars that Wittgenstein’s shift to his later key idea of meaning as use was influenced by what he is said to have called Sraffa’s ‘anthropological’ way of thinking. In part, Sinha shows, this influence appears to have involved Sraffa arguing to Wittgenstein that meaning as use could not be given a psychological interpretation—in effect, the notion that how people thought about how language is used determined how it is used—a view Wittgenstein had explored but whose later rejection became central to his important idea of ‘language-games.’

This elucidation of what Sraffa’s ‘anthropological’ way of thinking may have involved makes sense when we think of Sraffa’s philosophical critique of subjectivism that was motivated by his rejection of the Marshallian marginalist theory of value and effort to rehabilitate a Classical cost of production theory of value based on ‘objectivist’ philosophical foundations.⁸ An ‘anthropological’ way of thinking does not derive how people interact in social settings from how they individually understand their interaction but rather derives from how they understand their interaction directly from their common experience of it. If we think of Sraffa’s friendship with Antonio Gramsci and his knowledge of the class and labour struggles in Turin (see Naldi in this volume), the term ‘anthropological’ could be seen as a mild way of expressing Sraffa’s rejection of subjectivist views of the world. Wittgenstein could not but have been aware of Sraffa’s connection to Gramsci, and ‘anthropological’ may have been a term they used as a general (and innocent) reference for a philosophical view they came to share.

What I wish to focus on, however, is whether Sraffa and Wittgenstein shared a particular, real world instantiation of this ‘anthropological’ way of thinking. We know what that instantiation was for Wittgenstein: his idea of ‘language games.’ A ‘language-game’ (in German, a ‘Sprachspiel’) is a rule-governed form of social interaction in which the rules of use for a term establish its meaning. People learn a language by learning these rules just like they learn how to play games by playing them. Thus one doesn’t begin with a pre-determined set of meanings in one’s mind—a ‘private language’ as he called it—and then apply them in one situation after another. Indeed Wittgenstein thought the idea of a ‘private language’ was incoherent because he thought language is an essentially public activity. The rules we learn in interacting with others in learning language and in learning how to do all sorts of things are intrinsically social.

Thus, a ‘language-game’ or any other rule-governed interaction between people identifies a relatively integrated, distinguishable realm of social activity defined by the set of rules that apply to it. Sets of rules in any setting, in order to give guidance, need to be broadly consistent with one another in regard to what they apply to, and this serves to mark out and distinguish different realms of activity. For example, the ‘language-game’ governing social greetings marks out that realm of language and social activity from others, say, a speech or a lecture where other rules apply. At the same time, however, any given ‘language-game’ is only partially closed in this way because language meanings are transferred across social contexts and across different ‘language-games.’ Wittgenstein’s idea of a ‘language-game’ thus refers to a domain of social activity that is, we can say, both closed yet open at the same time.

In fact, Sraffa employed the same, real world instantiation of this ‘anthropological’ way of thinking in his 1931 unpublished “Surplus Product” manuscript (Sraffa 1931). Noting a problem he found he encountered in his determining the value of surplus product using his cost of production explanation of commodity values, Sraffa explicitly adopted an open-closed interpretation of the relationship between the system of commodity values and how distribution acted on that system (cf. Davis 2012, 2018, and Davis in this volume). Specifically, he maintained that commodity production constituted a ‘closed system’ that was acted upon by causes external to that it associated with the distribution of the surplus. Referring to that ‘closed system’ as the ‘economic field’ he concluded:

There must be a leak at one end or the other: the ‘closed system’ is in communication with the world.

Here, ‘the world’ refers to distribution of surplus product, and reflected the historical and social conflict over the shares of wages and profits in national income. He then added:

When we have defined our ‘economic field’, there are still outside causes which operate in it; and its effects go beyond the boundary. (D3/12/7: 161 (3–5), see Sinha 2016 for the complete note)

Much as for Wittgenstein ‘language-games’ are both relatively closed and yet still open, then, for Sraffa “outside causes” associated with distribution “operate in” the ‘economic field’ and produce “effects that go beyond the boundary” created by ‘the economic field’ (my emphases). His later Production of Commodities by Means of Commodities , does not refer to the open-closed distinction, but it can be argued it retains this same structure since there also the determination of commodity values is influenced by the division of the surplus.

The analogy I am suggesting, then, is between Wittgenstein’s idea of a ‘language-game’ and Sraffa’s ‘economic field’ conception (later his system of commodity values). Both refer to relatively integrated, distinguishable realms of social activity that also communicate with what lies outside them. Both set aside subjectivist accounts of the world for objectivist ‘anthropological’ ones. Both reason in terms of different types of forces operating in the real world—one set that operates in a relatively tight knit way within a set of social boundaries and another set that operates more broadly across social boundaries. Both used this conceptual framework to reject reductionist, subjectivist theories of the world—Wittgenstein in rejecting the idea that terms always refer to and name things we privately recognize, and Sraffa in rejecting the idea that all value is determined in an individualist subjectivist manner. Sinha has alerted us to this connection in emphasizing Wittgenstein’s reference to Sraffa’s ‘anthropological’ way of thinking. My argument is that Sraffa and Wittgenstein shared a particular social instantiation of this way of thinking that made use of the open-closed distinction.

It’s worth commenting that this view gives us a new view of the intellectual relationship between Sraffa and Wittgenstein. Wittgenstein’s acknowledgement of his debt to Sraffa in the ‘Preface’ to his Investigations suggests influence went primarily from Sraffa to Wittgenstein. That may still be true, but it might also be the case that they converged on mutually shared views as a result of their exchange and interaction. It also points us toward a potentially important contribution they may have jointly made, arguably deserving further investigation, to how an ‘anthropological’ way of thinking may combine careful analysis of a particular type of phenomena with its place in a wider social, historical process.

Reply

1.1 A Reflection on John Davis’s Comments

• Ajit Sinha 

I thank Professor Davis for his highly thought provoking comments on my paper (Sinha 2021). There are three conceptual points raised in his comment: (1) the anthropological way of thinking, (2) the question of open and closed systems and (3) the status of ‘surplus’ in economic theory. Davis relates the first and the third points to the second, where he finds the common meeting ground for Sraffa and the later Wittgenstein. He first relates the question of the ‘anthropological way’ with Sraffa’s early critique of Marshall’s subjectivism as well as the real world instantiation of this way of thinking. He argues that Wittgenstein’s conceptual frame of the ‘language game’ is an example of his real world application of the ‘anthropological way of thinking’. A ‘language game’ is a system governed by a set of rules and every new learner of a language learns those rules to make sense of the words used in such language games—in this framework there is no room for the idea that true meaning of a word resides in the psyche of the speaker or the mental picture in the mind of the speaker. Davis argues that though this framework of rules appears to be a closed system, the very fact that ‘language meanings are transferred across social contexts and across different “language-games” makes it open as well—‘Wittgenstein’s idea of a “language-game” thus refers to a domain of social activity that is, we can say, both closed yet open at the same time’ (Davis 2021). He then relates this idea with Sraffa’s 1931 note on ‘Surplus Product’. Davis argues that in this note, Sraffa, like Wittgenstein, argues that the value of the surplus product is determined in a closed system of production that defines the ‘economic field’; however, the economic field of production is acted upon by outside forces such as the distribution of the surplus—‘Much as for Wittgenstein “language-games” are both relatively closed and yet still open, then, for Sraffa “outside causes” associated with distribution “operate in” the “economic field” and produce “effects that go beyond the boundary” created by “the economic field” (author’s emphases)’ (Davis 2021).

Without contradicting Davis, I would like to take this opportunity to think further about the idea of open/closed systems and to what extent this could be made a central or common ground for Sraffa and the later Wittgenstein’s way of thinking. Strictly speaking, there is no system that can be characterized as completely closed or isolated except for the universe itself, which has no ‘outside’. But still, in many fields of pure sciences the distinction between open and closed systems have precise meaning. For example, ‘in thermodynamics, a closed system can exchange energy (as heat or work) but not matter, with its surroundings’, ‘in chemistry, a closed system is where no reactants or products can escape, only heat can be exchanged freely’ (Wikipedia). Unfortunately, in economics, or social sciences in general, we have not yet given it a precise meaning—therefore, when we say ‘closed yet open system at the same time’, it, at best, remains vague as to what we want to say. Let us take an example from economics. Many heterodox economists consider the orthodox general equilibrium theory to be a ‘closed system’ since it determines all prices as well as income distribution from the closed set of information it admits at the outset. But it cannot be denied that the information on ‘given endowments’ and ‘preference orderings’ are the data that connect it to the outside of the system. One can always argue that the initial endowments and preference orderings are determined in a larger socio-economic context and changes in those have effects on the variables determined inside the system and these variables in turn have effect on the larger socio-economic context. So how is this any different from Sraffa’s ‘economic field’ of his 1931 note and the problem of the determination of the ‘surplus’ as explained by Davis? However, if we define ‘economic field’ as a ‘system’ that determines primarily prices of commodities and the variables of income distribution such as rates of wages, profits and rents then one can legitimately classify the general equilibrium system as a relatively closed system since it determines all economic variables within the system. On the other hand, if we compare this with Sraffa’s ‘system’ in Production of Commodities by Means of Commodities then we find that the two nodes of the ‘general equilibrium system’ that connect it to the outside of ‘economic field’, namely, the initial endowments and the preference orderings are made irrelevant. In this case, we find the economic variables of income distribution such as rates of wages, profits and rents are placed in the open space—hence the ‘economic field’ as such is no longer ‘closed’ but remains open.

On the question of Wittgenstein learning from Sraffa the ‘anthropological way’ of looking at philosophical problems, my sense is that it is not much concerned with the question of open and closed systems. Wittgenstein’s philosophical questions were mainly related to the problem of whence a word or a proposition gets its meaning or what criteria can distinguish sense from non-sense? Now, as I have argued in my paper (Sinha 2021), the young Wittgenstein of the Tractatus thought that every proposition can be broken down to or analysed into independent elementary propositions. These elementary propositions are made of string of words that represent ‘simples’ or things. If these strings of words or things that they represent are related in a ‘logical form’ then the proposition becomes meaningful or sensible otherwise it is non-sense. I think Sraffa’s attack must have been on the metaphysical or the unscientific foundation of this thesis. This way of thinking implies that the true meaning of a proposition resides in the picturing of the things and the relations in which they stand in the mind of the speaker. Now we have enough evidence to be confident that Sraffa must have argued that since the analyst cannot have access to the speaker’s ‘mental images’, Wittgenstein’s criterion to distinguish ‘sense’ from ‘non-sense’ becomes metaphysical. The ‘anthropological way’ in this context must have meant a pointer to look for observable data. Thus, instead of analysing ‘propositions’ or ‘words’ in abstract or in its generality, look for particular cases or concrete examples in which a proposition is made. It was the genius of Wittgenstein to develop the idea of ‘language games’ to contextualize such particular cases or utterances and then show that most of the words in a language do not represent simples or things—the same word in different language games can convey different meanings; furthermore, these ‘language games’ themselves cannot be generalized as having some common property—they only resemble each other as faces in a family. Thus meaning of a word or a proposition cannot be separated from a way of life.

In this context, it does not appear to me that the conceptual distinction between open and closed systems is particularly helpful in understanding the distinction between the Tractarian thesis and the thesis proposed in the Investigations by Wittgenstein. As I have pointed out in my paper, ‘the fundamental concerns of both the Tractatus and the Investigations was to draw a line between sense and non-sense. In the Tractatus Wittgenstein drew an outer limit of language, which pushed all other propositions except the factual propositions to the other side of the limit. In the Investigations, however, there is no single outer limit of language but several internal limits are drawn. The “language games” represent those internal limits. A word may have sense within a particular language game but an attempt to drag it from one language game to another produces non-sense. For example, words such as God or Soul may have well-understood meanings in a religious language game but produce non-sense in a scientific language game’ (Sinha 2021). One can draw a parallel to Wittgenstein’s ‘language game’ with Sraffa’s Standard system and the Standard commodity. Every Sraffa’s system of ‘basic goods’ has a unique Standard system and a Standard commodity. It is within this context only that changes in prices with respect to variations in the distribution of income can be meaningfully interpreted since the Standard commodity as the standard of measure of prices is the only commodity that is not itself affected by the variations in the distribution of income; but once the ‘system of basic goods’ changes even slightly, ‘a comparison of the prices by the two methods become meaningless since its result appears to depend on which commodity is chosen as the standard of prices’ (Sraffa 1960, p. 82).

Now coming to Sraffa’s note of 1931 on ‘Surplus Product’, it seems to me that Professor Davis has mistaken Sraffa’s characterization of the Physiocratic interpretation of the ‘surplus product’ as Sraffa’s own, since in his interpretation of Sraffa’s quotation the crucial reference to it by Sraffa, expressed as, ‘Consider, e.g., the so-called “natural causes” of rent’ is missing (see Sinha 2016, pp. 82–89, for the complete note of Sraffa and my detailed commentary on it). This interesting note by Sraffa is yet again not primarily concerned with the question of open and closed systems . Its problem is: how to explain the existence of a ‘surplus product’ in economics if one relies on the ‘principle of sufficient reason’ or takes a deterministic scientific position. The problem is that conceptually the ‘surplus product’ is something that has no cost associated with it but if we take the deterministic scientific point of view then its existence must be explained by the sufficient cause that gave rise to it and in economic terms that sufficient cause must turn into a ‘cost’, negating the very concept of surplus. It was in this context, that Sraffa suggested that one way out of it would be to limit the economic field in such a way that the cause of ‘surplus product’ has no economic value and hence zero cost—the Physiocratic explanation of the ‘surplus product’ as a ‘gift of nature’ fits this way out; that is, the surplus product being the free energy of nature converted into economic goods in agricultural production. But this was not Sraffa’s own position as it is evident from the position adopted by him in the Production of Commodities by Means of Commodities. In this case, the surplus product is defined as the output over and above all the commodity inputs used in the production process—its origin or cause is not sought in anything freely available. The problem with the Physiocratic perspective lies in strictly measuring the cost of labour input. Is the contribution of energy by the labourer in the production process exactly equal to the calorie or the total energy content of the wages? If the energy content of wage is more than the energy contributed by the labourer in production then how does one account for the ‘surplus product’ appropriated by the labourer? Even if we suppose that one can identify exactly all the commodities that determine the minimum subsistence of the worker, once it is accepted that the wages are higher than the specified minimum subsistence it is no longer possible to draw a clear line between the subsistence wage and the surplus; as Joan Robinson had nicely put it, ‘we could hardly imagine that, when the workers had a surplus to spend on beef, their physical need for wheat was unchanged’ (1961, p. 54). This creates an indeterminacy in the measurement of either the necessity or the surplus. Sraffa’s way out of it was to adopt a technocratic subject’s standpoint who observes a reproducing system going round and round where it no longer is necessary to treat wages in a linear fashion as an input that produces human energy that in turn goes into the production process—the entire wage could be treated as a payment made after the production or a deduction from the ‘surplus’ produced, which can be determined by deducting all the commodity inputs from the commodity outputs. This is a measure that is determined from the standpoint of an observer just as the position of an electron is determined only in relation to an observer. This perspective rejects the Newtonian determinism in favour of indeterministic perspective of the quantum physics.