ChapterPDF Available

On Sraffa’s Challenge to Causality in Economics



The opening of Sraffa’s Archives has given the opportunity to unveil what is behind the published works, giving clues to interpret them. Among the contributions which are found in his unpublished papers, there are those related to Sraffa’s challenge to causality in economics. We argue that Sraffa’s entire research project is a struggle to escape from mechanical, i.e. causal theory, and to develop a geometrical representation of the economic structure. While a geometrical theory refers to an instant in time and is concerned with logical relations, a mechanical theory refers to processes that happen in real time, in which causality is involved. The reasons of why he embarked in such a project are complex and possibly related to his early beliefs that the requirements for causal explanations—like those which are met in physics—are too stringent to be applicable in economics, whose causes are often of metaphysical (and therefore ideological) nature. Unlike neoclassical economics, Sraffa held that change in economic realities hardly ever manifested itself in the form of infinitesimal variations in magnitudes that leave the overall structure unchanged. In that approach, change is required to find the marginal product (or utility) upon which supply-and-demand curves are derived and price and quantity are determined both in a given market on a given instant and at different times. The well-known passage in the Preface to Production of Commodities: “The investigation is concerned exclusively with such properties of an economic system as do not depend on changes in the scale of production or in the proportions of ‘factors’” (Sraffa, Production of Commodities by Means of Commodities: Prelude to a Critique of Economic Theory, Cambridge, UK, Cambridge University Press, 1960, v) can be better understood in the light of the above arguments. The issue is the difference between two instants (in which time is absent) and a change that takes place through time. By keeping change out of the scope, he was keeping the notion of causation out of his project of building a geometrical description of the economic system.
91© The Author(s) 2021
A. Sinha (ed.), A Reection on Sraffa’s Revolution in Economic
Theory, Palgrave Studies in the History of Economic Thought,
On Sraffa’s Challenge toCausality
MariaCristinaMarcuzzo andAnnalisaRosselli
The opening of Sraffa’s Archives1 has given us the opportunity to unveil
what lies behind his published works, offering clues to interpret them.
Among the contributions found in Sraffa’s unpublished papers are the
arguments he developed over the years on the role of causality in
In the last few years several scholars have tackled the issue which, how-
ever, remains to be untangled and fully understood. This more recent
1 References to the Sraffa Papers, which are kept at Trinity College Library, Cambridge,
and mostly available online, will be given hereafter following the catalogue classication. In
a few cases, we completed Sraffa’s abbreviations in order to make the text more readable.
M. C. Marcuzzo (*)
Sapienza, Università di Roma, Rome, Italy
A. Rosselli
Università di Roma “Torvergata”, Rome, Italy
literature converges on the interpretation—suggested by Sraffa’s himself
(see below)—that his entire research project is a struggle to escape from
“mechanical”, i.e. causal theory, and to develop a “geometrical” represen-
tation of the economic system. While the geometrical theory refers to an
instant in time and concerns logical relations, the mechanical theory refers
to processes that happen in real time, in which causality is involved. The
question remains open, however, as to the reasons why Sraffa embarked on
such a project; they certainly are complex and possibly related to his beliefs
that the requirements for causal explanations—like those met with in the
natural sciences—are too stringent to be applicable in economics or
because economic theory has often invoked causes that are of a metaphysi-
cal (and therefore ideological) nature. Some have stressed the similarity
with Wittgenstein’s criticism of causality (Davis 2012; Arena 2015), oth-
ers the inuence of his early interest in quantum mechanics (Sinha 2016),
physics, chemistry and philosophy of science (Kurz and Salvadori 2005,
2018), and yet others the inuence of Gramsci (Ginzburg 2013; Davis
2021). Possibly many factors concur to ll out the background picture,
which still needs to be fully discovered. In this paper, we review this recent
literature to raise a few points that may lead to further discussion.
2 Change anddiFFerenCe
In the Preface to Production of Commodities by Means of Commodities, the
reader is warned that the book: “is concerned exclusively with such prop-
erties of an economic system as do not depend on changes in the scale of
production or in the proportions of ‘factors’”. In this respect the work is
presented as an alternative to the “marginal approach” which “requires
attention to be focused on change, for without change either in the scale of
an industry or in the ‘proportions of the factors of productions’ there can
be neither marginal product nor marginal cost” (Sraffa 1960: v, empha-
sis added).
In fact, in neoclassical economics change in the quantities produced (or
consumed) is necessary to nd the marginal cost (or utility) upon which
supply-and-demand curves are built.
Krishna Bharadwaj was perhaps the rst to expand upon Sraffa’s 1960
brief reference to the “requisite kind of change” (ibidem), by arguing that
the role of change in supply-and-demand theories is to allow convergence
towards equilibrium and it must therefore be conceived as following a
well-specied path.2 Thus the direction of change required (i.e. the
assumption that demand and supply are well-behaved functions) imposes
“constraints which operate against an effective handling of problems of
change” (Bharadwaj 1986: 3). On the contrary, changes observed in the
real world are descriptions of events that may occur outside the strictures
of the predictive relation between price and quantity described by supply-
and- demand functions, which are necessary for the stability of equilib-
rium. She writes:
Thus, a statement about one position (at equilibrium) is tied up with a
theory specifying in exact terms the behaviour of the system under ‘change’
away from that position, even though these variations may be considered
‘hypothetical’ or theoretically conned to the innitesimally small (or to
points ‘in the neighbourhood’). […] The prices of production in the classi-
cal theory, in contrast, do not posit and therefore do not depend for their
stipulation upon the postulate of actual or ‘potential’ change.
(Bharadwaj 1986: 39)
And she adds that the classical system is more general in scope and ver-
satile in dealing with historic-specic factors because
it does not commit itself through its theoretical structure to any rigid form
and direction of change; [on the contrary, neoclassical theory] sought to
explain a single observed situation in terms of potential changes—and as
brought about by the balancing of marginal quantities operating through
the principle of substitution. In order to be consistent with these explanations,
the changes had to be in a direction and of the type postulated by theory.
(ibid.: 63–64)
The question of how change is conceptualised in economics appeared
quite early in the evolution of Sraffa’s thinking.3 In winter 1927–28, while
preparing for the lectures he was to give at Cambridge, he came to the
2 See Marcuzzo (2014), from which the paragraph draws.
3 See Sraffa’s comment in Summer 1927 “[…] the main stream of modern economic
thought proceeds to analyse the ways in which change takes place, without being hindered
by the fact that little is known of the ultimate causes of change” (D3/12/3/14). In fact,
here Sraffa is making a distinction between the notions of ‘ultimate cause’ and ‘mechanical
cause’. He is identifying the difference between the Classical and the Modern by associating
the Classical with the notion of ‘ultimate cause’ and the Modern with the ‘mechanical cause’.
We are indebted to the Editor of this volume for suggesting clarication of this distinction.
important conclusion that classical political economy and neoclassical eco-
nomics rested on alternative theories of value and distribution; only the
latter explains distribution and relative prices by means of the equilibrium
of the two opposing sets of forces, demand for and supply of goods and
factors of production; the classical political economy determines the distri-
bution of the surplus, i.e. of product other than wages.
Another point that Sraffa made in this context was that in economics
there is also the need to distinguish between situations in which change is
necessary to construct the elements of the theory and those in which it is
not required while analysis is conducted in terms of differences that co-
exist in the same instant.
This led Sraffa to raise the related point of the distinction between
“change” and “difference” in economic analysis. In one of the manu-
scripts of the period 1927–28, possibly where mention of the “Difference
(simultaneous) versus Change (succession in time)” rst occurs,
Sraffa wrote:
The general confusion in all theories of value [….] must be explained by the
failure to distinguish between two entirely distinct types of questions and
the universal attempt of solving them both by one single theory.
The two questions are:
1. what determines the [difference in the?] values at which various
commodities are exchanged in a given market on a given instant?
2. what determines the changes in the values of commodities at differ-
ent times? (e.g. of one commodity).” (D3/ 12/ 7/115; emphasis in
the original).
And he continues:
The rst problem gives rise to a geometrical theory, the second to a mechan-
ical one. The rst is so much timeless that it cannot even be called statical.
It does not represent an ideal stationary state in which it is assumed that no
change takes place: but it represents a situation at one instant of time, that
is to say something indistinguishable from a real state of things in such a
short period of time that no visible movement takes place. Its object is, as it
were, the photograph of a market place […] Marshall’s theory of value, with
its increasing and diminishing costs and marg[inal] utility, scissors, pillars
and forces, can only be understood as an attempt to solve the rst question
in terms of the second. (D3/12/7/117)
And in the same set of manuscripts, while working on his price equa-
tions, reecting on the possibility of letting the reader assume constant
returns to scale, Sraffa wrote:
[…] these equations cannot possibly answer as to how or why prices change.
They only explain why, at a given moment (?), prices of different things bear
to one another the proportions which they do. They explain variation (dif-
ference) between individual commodities at one time, not variations of one
commodity at different times.
No system of equations, whether it considers variable returns or not,
could tell this if time does not enter as a variable. (D3/12/ 7/85-6)
The same point, but related to the theory of distribution, is reiterated
more than 30years later in one of the many extant drafts of the Preface to
Production of Commodities:
The fundamental difference is that the extensive (different qualities of land)
is truly a purely timeless, or geometrical representation: all the different
lands exist simultaneously/at one instant, they and their products can be
ascertained, distinguished and measured at one instant, without changing
anything in the present arrangements.
On the contrary, the intensive (successive doses of c[apital] and l[abour]
on a piece of land) dim[inishing] ret[urns] do not exist at any one instant
[…] We can only nd these dim[inishing] ret[urns] by change, or movement:
that is to say, we require time. (D3/12/13/23.2; emphasis in the original).4
In conclusion Sraffa’s point is the distinction between two types of
change, those which involve the passage of time and those which do not.
See the annotation by Sraffa dated October 1929:
4 In a contemporaneous manuscript dated 1–3 Januar y 1958 and inscribed as “Margins
and margins”, which he might have considered including in his book, Sraffa made the same
point, returning to his earlier distinction between difference and change (D3/12/46/50).
Finally, Sraffa reiterated the same point as late as 17 August 1965, when he was still trying to
carry out his project for a comprehensive critique of marginal theor y (D3/12/42/5). On
these points, see Marcuzzo and Rosselli (2011).
[The following] notion of time is important: it really substitutes ‘instanta-
neous photographs’ as opposed to ordinary time. It is only a part of ordi-
nary time, it has only some of its connotations: it includes events, also
different events, but no change of events. It enables us to compare two
simultaneous, but not instantaneous, events—just as if they were ‘things’.
(D3/12/13/1.3 emphasis added)
In a different context, in the Introduction to Ricardo’s Principles, we nd
Sraffa explicitly referring to “the two points of view of difference and of
change” (Sraffa 1951: xlix) to distinguish between real and apparent
change. The distinction lies behind the question of how to measure “the
magnitude of aggregate of commodities” in order to study the distribu-
tion of surplus: there seems to be a change in the quantity of output to be
distributed whenever there is a change in its value due to a change either
in wages or in prots. However, this is an apparent change,5 since the
conditions of production of the commodities and the quantities produced
remain the same. It is the kind of change studied in Production of
Commodities when analysing the effects on prices of a change in distribu-
tion, not to be confused with the effects deriving from the introduction of
the time factor.
3 determinism
“Real” changes imply a process which occurs in time. Under what circum-
stances can we consider the effect of a change that implies comparison of
the same object at two different instants, equivalent to the observation of
two distinct objects at the same time? There are two possibilities and Sraffa
seems to have considered both. The rst is that we assume that the time
interval is so short that the assumption of ceteris paribus holds because of
lack of time for modications of the existing situation other than that in
5 Finally, the distinction between difference and change is employed by Sraffa to distin-
guish two different aspects of the effects on the relative value of two commodities of different
proportions or durabilities of capital employed in their production. Sraffa writes: “[the rst
aspect] is that of occasioning a difference in the relative values of two commodities which are
produced by equal quantities of labour. Second, that of the effect which a rise of wages has
in producing a change in their relative value” (Sraffa 1951: xlvii). This testies to the impor-
tance Sraffa attached to the distinction between change and difference in economics.
the independent variable.6 The second is that once the initial position is
known, any other position is known with absolute certainty, in the same
way in which, studying the motion of a physical body, the knowledge of its
initial position and of the forces applied to it allows us to determine any
successive position. As Sraffa wonders, discussing marginal productivity
“are the potential, hypothetical returns which would be obtained by addi-
tional doses part of the existing situation? (D3/12/49.5 recto; empha-
sis added).
To appreciate the relevance of this question fully, let us recall that in
1929 Sraffa had written that solely to be included in the “data” of the
theory were “things that actually happen”. Marginal magnitudes would
appear to be excluded from this condition by denition, being called into
existence by thought experiments. But now Sraffa seems to be considering
the possibility that marginal magnitudes might be conceived as “part of
the existing situation” as if they were “things that actually happen”.7 The
question Sraffa poses is, then, a crucial one. For if the answer were to be
afrmative, then marginal magnitudes could after all be taken as a basis for
economic theory to determine price and distribution. We hear in this
question an echo of Sraffa’s reections on the developments in physics in
the early decades of the twentieth century that challenged the determin-
ism of Newtonian mechanics, which had been the source of inspiration for
Marshall. We nd in his notes several passages copied from an essay of
1931 by Erwin Schroedinger with the title “Indeterminism in Physics”. In
one of these passages Sraffa summarises the “so-called problem of causal-
ity” in physics (D1/91/66 recto), quoting from Schroedinger: “Given
any physical system, is it possible, at any rate in theory, to make an exact
prediction of its future behaviour, provided that its nature and condition
at one given point of time are exactly known?” And Sraffa seems interested
in the conclusions that identical initial conditions do not lead to identical
results, according to the most recent experiments in physics of his time.
6 Referring to Russell’s interpretation of differential calculus in physics, Sraffa seems to be
thinking of differential calculus as a model because it considers “time in which effects follow
causes, but so closely that there is no room either for dispersion or for entering of foreign
inuences: it does this by differentiation (making time so short as actually to leave no room
for change in circumstances: the cause and effect are perfectly contiguous—nothing happens
in between” (quoted in Martins 2013: 44).
7 We interpret the existing situation as also including alternative techniques as long as they
are comparable because, even if not used, they “exist” at the same time. We are grateful to
the Editor of this volume for raising this point.
But under precisely what conditions would any future development be
included into the initial condition?8 To answer this question, in what
might be a draft of the Preface to Production of Commodities, Sraffa makes
reference to a passage in Wicksteed’s Alphabet of Economic Science on the
existence of a labour supply curve. Wicksteed’s passage runs as follows:
But even if he [the labourer] cannot tell what amount of work he will be
willing to do under the varying circumstances, obviously there is a given
amount which, as a matter of fact, he would be willing to do under any
given circumstances. Thus the curve really exists whether he is able to trace
it or not. (Wicksteed 1888: 55)
Variation in the labour supply “under varying circumstances” is clearly a
process of change. The point is, can the effects of this change be included
within the “existing situation”? “Wicksteed”, Sraffa observes, “with great
and indeed reckless consistency, thinks that they do.” (D3/12/49.5 recto).
In fact, according to Wicksteed such change appears to be determined in its
entirety by the initial position as if it were known a priori. To this Sraffa
replies:9 “Wicksteed considers that the path to be followed when one of the
quantities is changed is prescribed a priori like the rails prescribe the path of
a tramcar”. (D3/12/46.43b recto). And yet again, in another draft:
This sounds like a declaration of faith in universal determinism (‘not a bus,
not a bus, but a tram’): however Wicksteed considers that when one of the
quantities is changed, the dependent variable will follow a path which is
prescribed a priori although no rails or other circumstances capable of
directing its motion are visible. But of course, saying that it will follow ‘one’
path at a time does not prove that it will follow one particular prescribed
path rather than another. (D3/12/46.43b recto)
8 The following paragraphs draw on Rosselli and Trabucchi (2019).
9 For the vivid contrast of the image of a tram vs. that of a bus, we have preferred to quote
this version, which Sraffa crossed out, to the one he kept in the manuscript which runs as
This is nothing less than a declaration of faith in universal determinism, for nothing
less can support the belief in the actual existence of a prescribed path which must
inevitably be followed, whether by the consumer or by the producer, such as is
described by the demand- and supply-curves: for no observation, however minute, of
the existing situation (in our case, of the existing methods of production) can bring
out the path along which they must move in any given circumstances.
(D3/12/46.43b recto)
In the light of the distinction between difference and change we can
appreciate the sense of Sraffa’s forthright reference to a “faith in universal
determinism” as the only conceivable foundation for the “marginist”
method (as Sraffa used to call it). Variations in the labour supply “under
varying circumstances”—and the same obviously applies to variations in
the cost of production of a commodity attending variations in the quantity
produced or, again, to variations in output due to variations in the propor-
tions in which different means of production are used—are all processes of
change. As such they should be open to a wide range of outcomes.
However, this would preclude them from serving as a basis for an explana-
tion of value and distribution. Or, as Sraffa notes analysing the change in
output as a result of the increase in the use of a factor of production: “The
‘change’ or historical method would in practice involve trial and error but
the theory assumes that ‘the best’ or at any rate prescribed method is
known a priori” (D3/12/49.6 recto).
Thus the only way to support the marginist method would appear to lie
in postulating a certain determinism: in the form, for example, of a marked
degree of regularity in the relation between cost and quantity produced—
which, we may note, is precisely what would happen if the product
obtained with n units of a factor and the product obtained with n+1 units
were two different magnitudes existing side by side.
In fact, Sraffa speaks of faith in such regularity, because no actual expe-
rience can conrm it. Here it is worth noting that the role of the experi-
ments that should lead to denition of the marginal magnitudes is, in fact,
completely different from the part experiments normally have to play in
the physical sciences. Unlike physics, experiments for marginalist theory
do not consist in application of the conceptual tools of the theory to inves-
tigate certain properties of the economic systems under study. Rather, the
very creation of these tools is made to depend upon one and only one
possible result of the experiments. If the experiments were to yield differ-
ent outcomes, we would not be dealing with a different theory, i.e. a
theory yielding different results, for there would no longer be any theory
from which results could be derived.
This criticism of “marginism” (again, Sraffa’s wording), focussed on
the regularity of the outcome of changes that must be assumed in order to
build supply-and-demand curves, has more general implications for the
study of causal relations in economics.
4 Continuity andstruCtural Change
Let us now look at the other possibility that we can take into account to
consider the effect of a change, which implies comparison of the same
object at two different instants, equivalent to the observation of two dis-
tinct objects at the same time. Could we consider variations so small that
the assumption that nothing else changes is plausible? Sraffa seems to
argue that in economics it is not possible to resort to innitesimal calculus.
First, variations are highly unlikely to come about in a continuous manner
but rather occur in discrete form. Moreover, change hardly ever takes the
form of variations in magnitudes that leave the overall structure unchanged.
On various occasions over the years Sraffa objected to Marshall’s case of
an “alert” railway manager who deals with the increasing number of pas-
sengers by altering the composition and size of carriages in a train, “con-
stantly weighing the net product in saving of time and of annoyance to
passengers, that will accrue from the aid of a second guard on an impor-
tant train and considering whether it will be worth its costs” (Marshall
1920: 427).
On 29 March 1963, Sraffa again wrote:
This suggests that his [the alert manager’s] main task is to sack a porter here,
add a coach to a train there, or shorten a platform elsewhere. The idea is that
the process of change can be reduced to a continuous process, like shorten-
ing platforms: ‘a penny is the basis of a million’, and so a process of shorten-
ing, adding [,] sacking in detail is the route from one position to another. In
Marshall’s view the ‘alert’ Dr. B…. never needs to take bird’s eye view of his
enterprise. (D3/12/42.12)
This awareness that in economics changes in one variable affect the
whole structure connects with Sraffa’s repeated observation that the mar-
ginal productivity is always measured in the case of two factors, which is
the only case in which changing the quantity of one factor and changing
the proportions in which factors combine amount to the same thing.
However, when we have more than two factors, can we still maintain that
the proportion between the unchanged factors remains constant in the
presence of a change in the quantity of one factor?
We have several passages showing that Sraffa was critical of the assump-
tion that one could deal with change in real time as if the ceteris paribus
assumption held.
M[arshall] is constantly on the defensive against objections to continuity
based on facts […] This is not the basis on which this [my argument] is
based. That can be granted in detail, wherever it is possible. It is against the
logical possibility of the type of continuity assumed. That type is only pos-
sible with two factors. (D3/12/42.12)
And again:
Where marginism goes astray is in (falsely) assuming […] that it has general
applicability whereas in fact it only applies exceptionally (in cases where par-
tial change is feasible, there is independence, the whole is not affected).
However, even if supply-and-demand functions (for goods and factors)
were continuous so that marginal calculus were applicable, there would
still be one obstacle to adopting the methods of mechanics in economic
analysis. Sraffa wrote:
[…] in mechanics if the experiment is repeated in similar circumstances (say,
on the elasticity of a metal) the same results will be obtained. But with sup-
ply and demand, even if the external circumstances were the same, the result
would be different because man learns from experience, or at any rate is
changed by it, forms and transforms habits, etc. (D3/12/42/11recto)
In conclusion, differential calculus and demand and supply functions
are not tools applicable to economic analysis borrowing the method of
classical physics—i.e. mechanical causality does not work in economics.
Does this mean that Sraffa gave up the possibility of establishing causal
relationships to explain change in economics? In the next section we look
at what the literature had to say on the matter.
5 interpreting meChaniCal Vs.
geometriCal Causality
On Sraffa’s challenge to the notion of mechanical causality in economics
there have recently appeared a fair number of interesting contributions
which, although not differing greatly, nevertheless highlight different
aspects of the matter. In common they offer a defence of Sraffa’s theory
against the criticism of lacking generality (for example, valid only by
assuming constant returns to scale) or of being narrow in scope (for exam-
ple, it allegedly describes an equilibrium situation without specifying how
the system reaches it). Moreover, investigation into Sraffa’s research proj-
ect has brought to light the fact that his theory of prices and distribution
is not subject to the limitations that constrain the marginalist approach
which determines prices through the balancing of the forces of supply and
demand. This is the view that Sraffa expressed at the beginning of his work
on Production of Commodities in the famous parable of the “Man from the
moon” (D3/12/7.87) and reiterated in 1942:
The problem is that of ascertaining the conditions of equilibrium of a system
of prices and the rate of prots, independently of the study of the forces
which may bring about such a state of equilibrium. Since the solution of the
second problem carries with it a solution of the rst, that is the course usu-
ally adopted in modern theory. The rst problem however is susceptible of
a more general treatment, independent of the particular forces assumed for
the second; and in view of the unsatisfactory character of the latter, there is
advantage in maintaining its independence. (D3/12/15.2)
Ginzburg (2013) provides an interesting interpretation of this point,
highlighting the distinction between causal representations, which answer
the question “why” and non-causal representations which answer the
question “how”.
The distinction is between representations which investigate the causes
of phenomena (and are therefore presumably able to predict them) and
merely descriptive representations, like “snapshots”,10 which capture the
system as it actually is at a given instant in time. Ginzburg claims that this
“metaphor, if taken literally as a photograph of a real economic system, is
misleading. It can be considered a snapshot only in the sense that all the
phenomena taken into account relate to the same period” (Ginzburg
2013: 113). Ginzburg argues that non-causal representations are not real-
istic descriptions of a situation, but require a mode of abstraction: for
example, the methods of production under consideration are those pre-
vailing in the system and not those actually adopted under specic circum-
stances. Non-causal representations capture meaningful features of the
10 The rst to employ the “snapshot” metaphor was Roncaglia in his 1975 Italian book,
translated into English as Roncaglia (1978). For a more recent restatement of his interpreta-
tion of Sraffa’s approach, see Roncaglia (2009). On the evidence of Sraffa’s reference to the
snapshot, see Kurz and Salvadori (2018).
object of analysis and the properties of the structure without the need to
bring in specications of their developments over time, which remain
open to a variety of possibilities.
The idea that equilibrium prices can be dened independently of the
forces which guarantee that equilibrium is reached does not imply that
equilibrium is independent of other factors which call for a different kind
of analysis. The argument that social analysis, and economic theory in
particular, must address questions which require different levels of abstrac-
tion and generality had been put forward by Garegnani in several contri-
butions (1984, 1990a, b) as well as various other scholars who adopted his
This view—which is known as the “core” interpretation of Sraffa’s the-
ory—has found support in a number of studies by Kurz and Salvadori
(2004, 2005, 2018). In particular, they have introduced two elements
which clarify Sraffa’s research project through a careful reconstruction of
the complex work underlying the nal text of Production of Commodities.
The rst element is the inuence on Sraffa of his perusal of a great many
books on the natural sciences and how this led to a search for “objectiv-
ism” which, at an early stage11 consisted in an attempt to construct a
theory of exchange values whose elements would be measurable quanti-
ties, thus excluding any reference to motivations. The economic system
could then be represented as production of commodities by commodities,
since any output would be the result of the destruction of all the inputs
used up in its production.
The second element highlighted by Kurz and Salvadori and, with his
alternative interpretation, by Sinha (2016)—and recently by Davis
(2020)—is Sraffa’s important note on “Surplus Product” written in
August 1931. It marks the abandonment of the “natural science point of
view” which links one effect to one cause.
Sraffa realised that the distinction between surplus and necessities
brought him up against an apparently insoluble dilemma: either all the
11 Sraffa had endorsed Francis Bacon’s principle of “efcient causes” as against “nal
causes”. He wrote in the December 1927–December 1928 period: “‘Efcient causes’ are
facts of the past that act on the present: ‘nal causes’ are facts of the future that act on the
present. The existence of the latter is at best dubious and they are better called ‘illusions’.
The classical P[olitical] E[conomy] dealt only with the rst sort of causes, i.e. of ‘material
things’ that have existed in the past. Modern economics deals with the second class, i.e.
hopes for the future, such as utility, abstinence, disutility, etc.; these things, it must be
noticed, refer only to the foreseeing of future acts.” (D3/12/10/ 61.1recto).
surplus is necessities, i.e. all that is produced is necessary for the reiteration
of the production and the surplus vanishes, or the components of the sur-
plus, like prots and rents, must be represented as inducements and then
the distinction between objective and subjective disappears. This led Sraffa
to conclude that “The surplus may be the effect of the outside causes; and
the effects of the distribution of the surplus may lie outside.”
In the search for the inuences which might have led Sraffa along the
path to abandoning the notion of mechanical causality in economics, it has
been argued in the literature that the exchanges between Sraffa and
Wittgenstein from 1929 to 1946 provide evidence that Sraffa had an inu-
ence on the preparation of Wittgenstein’s Philosophical Investigations and
in turn Wittgenstein had an inuence on Sraffa’s Production of Commodities.
Arena (2015) in particular has argued that the letters between them show
that both Sraffa and Wittgenstein criticised various notions of causality,
and specically, the notion of mechanical causality.
Wittgenstein abandoned the idea that there is only one type of causality
that we could discover by observation and experiments, interpreting “cau-
sality as one grammar rule amongst others” (Arena 2015: 1097). He also
criticises the notion of causes based on individual motivations,
such as beliefs, intentions or desires, arguing that if agent motivations do
exist, they do not necessarily produce and therefore explain their actions:
motivations (even social ones) are not causes. (ibid.)
Accordingly, Arena nds a clear parallelism with what Sraffa had been
arguing since his early period:
Economists who do not take this objective test as the standard of what is the
cause of an event, are always driven back to trace the ‘ultimate causes, causae
causantes, etc.’ to the wants, desires, aversions, decisions, volitions and
intentions (or inducements and rewards) of individuals. In fact, if we do not
use an objective standard and rely simply upon an unconscious ‘feeling that
this must be the real cause’ … we are bound to base the conclusions on our
own individual experience, from which it appears that we do what we want,
or what we like, etc., and this seems the only convincing nal conclusion”.
(D 1/9/6, pre-192812)
12 According to Nerio Naldi in a private communication, the papers in D1/9—in the cata-
logue dated pre-1928—in fact belong to the 1928–1931 period, as revealed by the annota-
tion “next lecture” on one sheet.
Wittgenstein introduced a new type of representation, that can be
dened as surveyable or perspicuous representation, which aims at bringing
out “mutual links and connections in the data and corresponds to the
synthesis or the synopsis of the grammatical rules of a form of life, but
without making any assumptions about their evolution over time” (Arena
2015: 1099).
So Arena also agrees that Sraffa must have come to the conclusion that
a “snapshot” was best suited to escape from causality in providing a repre-
sentation of the economic system.
Davis (2012) also re-assessed the nature of the inuence on and by
Wittgenstein, but in Davis (2020) he adds a further element to the present
discussion, i.e. the distinction between closed and open systems. A “closed
system” is isolated from the environment and is subject to principles which
are invariant relative to it; on the contrary the functioning of an open sys-
tem is inuenced by the environment.
Davis argues that Sraffa conned the objectivist, sufcient cause prin-
ciple to the eld of commodity prices, but also supposed that this “closed
system” was operated upon as a whole by causes that lay outside of it.
Davis contends that Sraffa both maintained and modied his objectivism
by allowing for different types of causal factors associated respectively with
production and distribution. In conclusion, Sraffa is said to allow for two
types of causal forces, one associated with the natural world, the other
with the social world, and these two types of causal forces are assumed to
be interacting “in a complex manner” (Davis 2012: 9).
The distinction between an open and a closed system seems to nd cor-
roboration in the following passage dating to August 1931:
Thus there must be a leak at one end or the other: the ‘closed system’ is in
communication with the world.
When we have dened our ‘economic eld’, there are still outside causes
which operate in it; and its effects go beyond the boundary. This must hap-
pen in any concrete case. (D3/12/7/161.5recto)
Sinha (2016) places at the centre of his interpretation Sraffa’s insistence
on the difference between two types of explanations, meant to answer two
types of questions:
why ‘x’ has increased in price from $2 to $3 […] why ‘x’ is sold at $3. The
rst question needs an answer in terms of a discovery of a cause that explains
the change… the second question […] requires an answer in terms of a rela-
tion (a logical relation), that is, why ‘x’ must relate to $ or other commodi-
ties ‘y’, ‘z’ and so on in a precise quantitative association at one moment.
The rst case is mechanical in nature whereas the second is geometrical. In
the second case, the problem of causation does not arise because time is
absent from the problem—hence the explanation must be a description of
what exists”. (Sinha 2016: xii)
Sinha has claried the core point of the discussion on the nature of
Sraffa’s objections to the notion of causality in economics, or at least the
notion of causality borrowed from classical mechanics, namely the nexus
time-change-cause. Observed phenomena could be traced back to a given
cause only if we could observe the “forces” that produced that effect, but
this presupposes going back in time to produce a change in the existing
situation. This, however, does not mean that we cannot study situations
following a change, for instance in a given variable (wage or prot),
although we can only do so as long as these changes have no time dimen-
sion, and are merely a logical derivation.
6 the Boundaries oFthe“eConomiC Field
As in the case of objectivism, which has been interpreted as if Sraffa were
a crude positivist, unable to see the importance of subjective magnitudes,
like expectations, which although unobservable play an important role in
economics, his rejection of mechanical causality should not be interpreted
in a restrictive way. Obviously, Sraffa set out to understand economic real-
ity and therefore how economic phenomena could best be explained; his
concern seems to have been primarily to avoid the employment of tools or
concepts that appeared from the outset to be of limited generality or dubi-
ous existence (utility), or indeed ideologically vitiated (abstinence). The
pursuit of a scientic method to apply to economics required the utmost
care and attention to avoid metaphysics, ideology and ad hoc assumptions.
The way Sraffa went about it was a very drastic delimitation of the
“economic eld”, as he called it, leaving out “causes” from the investiga-
tion. It is argued by many interpreters, as we have seen, that Sraffa makes
a distinction between the “economic eld”—the plane to which only a
geometrical theory can be applied, i.e. satisfying the criterion of using only
logical and timeless relationships, without dealing with “ultimate causes”—
and the outside realm of circumstances and historically given context, as
was clear to him from the very beginning of his research programme:
Clearly, we must reduce all the data to things that actually happen, exclud-
ing inexistent possibilities. Only such things are measurable, and can enter
the theory as ‘knowns’, or ‘constants’; and, in reality, only really happening
things can be real causes and determine effects. (D3/ 12/ 13/ 1.2)
What this implies is that institutions, technology, social norms and lev-
els of activity cannot be approached with the same level of abstraction as
within the “economic eld”. However, this leaves open the question of
the nature of the relationships prevailing outside. The question arises if we
can take them as causal forces even though we cannot represent them in a
functional deterministic form.
Any attempt to assess how Sraffa would have answered this question
takes us onto slippery ground, because direct evidence is lacking. Seeking
out Sraffa’s motivations and objectives—e.g. demolition of the marginal
productivity theory, rehabilitation of classical political economy—and/or
searching for inuences by the contemporary scientic discussion (quan-
tum physics) or personal interactions with particular individuals
(Wittgenstein, Gramsci) may well prove illuminating, but hardly conclu-
sive on this point.
What is clear is that we are left with alternative options to continue
Sraffa’s research programme and accept his legacy. We could attempt to
expand the “economic eld” beyond the boundaries marked by Sraffa,
introducing further relationships of the wage-prot type, namely those
that have the property of bringing in an “if and only if” clause. Another
approach could be to enlarge the realm of the “economic eld”, applying
price equations to problems that imply changes, for instance, in the quan-
tities produced, but introducing assumptions (such as constant returns to
scale) that restrict the level of generality Sraffa aimed at.
In any case, we have to continue looking into the causes at work out-
side the economic eld and build a bridge between the core and the out-
side world. This might lead us to accept that the level of abstraction and
therefore of generality will be lower, gaining in relevance what is lost
in rigour.
Only by keeping the uses we can make of Sraffa’s system of equations
separate from the reasons which may have led Sraffa to travel the road of
the geometrical representation and, above all, of what is really meant by it,
can we make any further headway, without being bogged down by dis-
putes on Sraffa’s motivations or being unfaithful to his text.
What we are left with is a powerful warning against invoking spurious
causes to explain economic phenomena; we are also left with means to
prevent us from mistaking changes for differences and above all a fence
against false claims of rigour and generality.
Arena R. (2015), Order, process and morphology: Sraffa and Wittgenstein,
Cambridge Journal of Economics 39: 1087–1108.
Bharadwaj K. (1986), Classical Political Economy and the Rise to Dominance of
Supply and Demand Theories (2nd ed), Calcutta: Orient Longmans.
Davis J. (2012), The change in Sraffa’s philosophical thinking, Cambridge Journal
of Economics 36: 1341–56.
Davis J. (2021), ‘Openness’ as a methodological principle of Sraffa’s economic
thinking; this volume.
Garegnani P. (1984), Value and distribution in the classical economists and Marx,
Oxford Economic Papers 36: 291–325.
Garegnani P. (1990a), On a supposed obstacle to the tendency of market prices
towards natural prices, Political Economy: Studies in the Surplus Approach
6: 329-59.
Garegnani P. (1990b), Sraffa’s Production of Commodities: classical versus mar-
ginalist analysis in K. Bharadwaj and B. Schefold (eds), Essays on Piero Sraffa,
London: Unwin Hyman.
Ginzburg A. (2013), Sraffa, Sen and Non-Causal Representations in Economic
Analysis. In E.S. Levrero, A. Palumbo, A. Stirati (eds), Sraffa and the
Reconstruction of Economic Theory, Volume III: Sraffa’s Legacy: Interpretations
and Historical Perspectives, Basingstoke: Palgrave Macmillan.
Kurz H. and Salvadori N. (2004), Man from the Moon: on Sraffa’s objectivism,
Économies et Sociétés 35: 1545–57.
Kurz H. and Salvadori N. (2005), Representing the production and circulation of
commodities in material terms: on Sraffa’s objectivism, Review of Political
Economy 17: 69–97.
Kurz H. and Salvadori N. (2018), On The “Photograph” Interpretation of Piero
Sraffa’s Production Equations: A View From The Sraffa Archive in J.Kregel,
M. Corsi and C. D’Ippoliti (eds), Classical Economics Today, London and
NewYork: Anthem Press.
Marcuzzo, M.C. (2014), On alternative notions of change and choice. Krishna
Bharadwaj’s legacy, Cambridge Journal of Economics 38: 49–62.
Marcuzzo, M.C. and Rosselli, A. (2011), Sraffa and his arguments against mar-
ginism, Cambridge Journal of Economics 35: 219–31.
Marshall A. 1920 [1890], Principles of Economics. London: Macmillan.
Martins N. (2013), The Cambridge Revival of Political Economy, London:
Roncaglia A. (1975), Sraffa e la teoria dei prezzi, Bari: Laterza.
Roncaglia A. (1978), Sraffa and the Theory of Prices, Chichester: Wiley.
Roncaglia A. (2009), Piero Sraffa and his life, thought and cultural heritage.
London: Routledge.
Rosselli A. and Trabucchi P. (2019), Sraffa, the ‘marginal’ method and change,
Structural Change and Economic Dynamics 51: 334–340.
Sinha A. (2016), A Revolution in Economic Theory: The Economics of Piero Sraffa,
London: Palgrave Macmillan.
Sraffa P. (1951), Introduction in D.Ricardo, Works and Correspondence, edited by
P.Sraffa with the collaboration of M. Dobb, vol. I, Cambridge: Cambridge
University Press.
Sraffa P. (1960). Production of Commodities by Means of Commodities, Cambridge:
Cambridge University Press.
Wicksteed P. (1888), The Alphabet of Economic Science, London: Macmillan.
Recent economic and financial crises have exposed mainstream economics to severe criticism, bringing present research and teaching styles into question. Building on a solid and vivid tradition of economic thought, this book challenges conventional thinking in the field of economics. The authors turn to the work of Luigi Pasinetti, who proposed a list of nine methodological and theoretical ideas that characterize the Classical Keynesian School. Drawing inspiration from both Keynes and Sraffa, this school has forged a long-standing and ambitious research programme often advocated as a competing paradigm to mainstream economics. Overall, the Classical Keynesian School provides a comprehensive analytical framework into which most non-mainstream schools of thought can be integrated. In this collection, a group of leading scholars critically assess the nine main ideas that, in Pasinetti's view, characterize the Classical-Keynesian approach, evaluating their relevance for both the history of economics and for present economic research.
Full-text available
Piero Sraffa’s Production of Commodities by Means of Commodities (1960) was path-breaking as a contribution to political economy and penetrating as a critique of the orthodox economics of the twentieth century. As Ajit Sinha recently put it, the book produced a ‘revolution in economic theory’ (Sinha, A Revolution in Economic Theory: The Economics of Piero Sraffa, London: Palgrave Macmillan, 2016; cf. Martins, “The Sraffian Methodenstreit and the Revolution in Economic Theory,” Cambridge Journal of Economics 43 (2): 507–525, 2019), the impact and significance of which continues to be investigated. At the same time, despite the depth and far-reaching implications of Sraffa’s critique of orthodox economics it was ignored by the great majority of economists, and this not only complicates our understanding of its impact on economic theory, but also creates a paradox regarding our interpretation of ‘the’ history of economics. For Sraffa, ‘the’ history of economics dates back at least to Adam Smith and David Ricardo as founders of a subject specifically understood as political economy. Yet economics today is no longer identified as political economy by most people in the field, but is conventionally said to be a science independent of history, politics, and social values, and thus makes little reference to how the social organization of the economy was a distinctive characteristic of the thinking of Smith, Ricardo, Marx, and others in the history of political economy.
Full-text available
This paper discusses the impact of Sraffa's thinking on economics. It argues increasing specialization in research is producing an 'all trees, no forest' fragmentation of economics that creates opportunities for a return to concerns that motivated classical political economy. It associates this with a methodological conception of what a more pluralistic economics involves, and applies this to relationships between production and distribution. A methodological conception of 'openness' is traced to a 1931 turning point in Sraffa's thinking when he used an open-closed distinction to explain the relationship between production and distribution, and engaged in a philosophy of science reasoning reminiscent of systems theory. The paper argues there are important parallels between Sraffa and Gramsci's thinking regarding the open-closed distinction.
Full-text available
In the preface to Production of Commodities by Means of Commodities Sraffa emphasizes that his book does not make use of the method of marginal magnitudes. This paper, based mainly on the notes that Sraffa wrote over nearly forty years, shows that Sraffa's rejection of this method, which he calls ‘marginism’ is not due to some aprioristic methodological preconception, but is part of his views on the appropriate method to deal with actual economic phenomena. Indeed, ‘marginism’ deals with changes, which occur in time, as if changes were always amenable to the difference between two situations which exist side by side. But change, outside the world of mechanics and in that of social phenomena, does not follow predetermined paths which are known a priori. Therefore, the marginist method appears to Sraffa as constraining economic analysis within particularly rigid patterns inadequate for the study of economics. In the light of this criticism, the paper sheds new light on Sraffa's attention, which he never relented, for some passages of Marshall's Principles.
'Drawing extensively on Sraffa's Cambridge archives, particularly his unpublished notes, Ajit Sinha reveals Sraffa's struggle – philosophical, methodological and technical – on the road towards the brief formulation in Production of Commodities by Means of Commodities (1960). Sinha puts paid thereby to alternative readings of that enigmatic work, including that of the late Pierangelo Garegnani, Sraffa's student and the leading Sraffian commentator. Sinha's concern with the philosophical underpinnings of the discipline and his dissatisfaction with technicality for its own sake reflect a broadness of intellectual vision that would serve well any economics program taking the history of thought seriously.' – Samuel Hollander, Professor Emeritus, University of Toronto, Canada, and Officer of the Order of Canada 'The two main features of Ajit Sinha's book show the fundamental nature of its contribution to economic analysis and history of economic thought. On the one hand, the author offers a complete and thorough reconstruction and interpretation of Sraffa's works, including his unpublished papers. On the other hand, it strongly contributes to the recent revival of the economists' debates on the meaning of Sraffa's contribution to economics, especially and unusually stressing the importance of its philosophical underpinnings.' – Richard Arena, Professor of Economics, University of Nice Sophia Antipolis, France This book draws on the work of one of the sharpest minds of the 20th century, Piero Sraffa. Ludwig Wittgenstein credited him for 'the most consequential ideas' of the Philosophical Investigations (1953) and put him high on his short list of geniuses. Sraffa's revolutionary contribution to economics was, however, lost to the world because economists did not pay attention to the philosophical underpinnings of his economics. Based on exhaustive archival research, Sinha presents an exciting new thesis that shows how Sraffa challenged the usual mode of theorizing in terms of essential and mechanical causation and, instead, argued for a descriptive or geometrical theory based on simultaneous relations. A consequence of this approach was a complete removal of 'agent's subjectivity' and 'marginal method' or counterfactual reasoning from economic analysis – the two fundamental pillars of orthodox economic theory.
Amartya Sen (2003 and 2004) has drawn attention to two aspects that are important for an understanding of Sraffa’s (1960) Production of Commodities by Means of Commodities. The first (mostly at the level of enunciation) addresses the need not to separate Sraffa’s contribution as an economist from his philosophical stance, in its broadest sense, and in particular from the context of intense cultural exchange entertained with Gramsci on the one hand, and with Wittgenstein on the other. I will deal here only marginally with this issue,1 focusing rather on the second aspect, concerning the distinction between causal models of scientific explanation and schemes of ‘analytical determination’, i.e., non-causal representations (or descriptions). An example of the former, according to Sen, is the ‘causal determination’ of the marginalist theory of prices. An example of the latter is Sraffa’s scheme, in which inputs and outputs are fixed, ‘as in a snapshot of production operations in the economy’.2 Sen’s conclusion is that, by adopting a method corresponding to the second approach, from the start Sraffa would give up the attempt to construct a theory of prices and distribution alternative to marginalist theory. So would his contribution be of no use in constructive terms? Sen rejects such a drastic conclusion, arguing (or conceding) that this method encouraged the classical political economists and Marx to expand the area of interest for economists, proposing, albeit in a ‘merely’ descriptive domain, important issues for social and political communication.