History of Economic Analysis part 64 pps

So far as this point is concerned, there seems to be no difference at all between Mill
and Marshall. Both recognized the importance that the desire to hold money, rather than
to spend it on goods and services, may acquire in certain situations—crises and
depressions in particular. And the only difference there is on this point between Mill and
Keynes is this: the former confined this excess demand for money to situations of this
kind, of which it is one of the consequences and which therefore cannot be explained by
it; whereas the latter considered the excess demand for money in depressions only as the
most spectacular form of a phenomenon that, in less spectacular forms, is well-nigh
ubiquitous or is well-nigh ubiquitous at least in certain phases of capitalist evolution, so
that it may become the cause of either cyclical downturns or ‘secular stagnation.’ Malthus
seems to have taken the latter view.
11

A much more important reason for Malthus’ dissent from Say and much more basic to
his principle of effectual or effective demand was, however, his opinion that saving, even
if promptly invested, may lead to deadlock if carried beyond a certain optimal point (op.
cit., ch. 7,
3). He did not go as far as Lauderdale,
12
who was the real anti-saver of the
age. He granted to the pro-savers even more than he should have done, namely, that
increase in capital cannot be effected in any other way than by saving. But he maintained
that, carried beyond an optimum point, saving would create an untenable situation: the
effectual demand for consumers’ goods from capitalists and landlords would not increase
enough to take care of the increased supply of products that results from an ever-
increasing conversion of revenue into capital; and the effectual demand for consumers’
goods from laborers, though it would increase indeed, cannot constitute a motive for
further accumulating and employment of capital. It is this which constitutes Malthus’
fundamental objection to Say’s law. The mistake involved will be analyzed below. But it
cannot be charged to Keynes. Though many passages in Malthus and also in Lauderdale
undoubtedly are suggestive of part of today’s (or yesterday’s) anti-saving argument, I

cannot help thinking that Lord Keynes should not have approved of Malthus’ every word
so sweepingly.
13
However, the idea of a schedule of aggregate demand for consumers’
goods taken as a whole, though without any awareness of the problems this concept
raises,
14
is in fact present in Malthus’ analytic set-up, and it may be therefore claimed
with justice that he anticipated Wicksell, who was the next first-flight economist to adopt
it.
Since the question of general gluts will come up again in the next chapter, I leave the
matter at this point. And since neither Say, nor Malthus, nor Mill was aware of the
problems of determinateness of equilibrium that the monetary factor may raise, we shall
leave this aspect for the next Part. But some readers may welcome a summary with
further reference to Keynes’s analysis, which accordingly I shall present now.
Keynes, of course, never meant to contradict the proposition that has been called Say’s
law above. This shows in his warning that his Aggregate Supply Function and Aggregate
Demand Function
15
must not be confused with
11
This is how I interpret, as did Lange (op. cit. p. 61), the passage in Malthus, Principles (footnote
on pp. 361–2 of the 1st ed.).
12
Malthus, in referring (op. cit. p. 352 n.) to Chapter 4 on Parsimony in Lauderdale’s Inquiry and
in expressing the opinion that Lauderdale went ‘as much too far in deprecating accumulation as
some other writers [Smith included] in recommending it,’ wrote a sentence that is worth quoting
both because of its wisdom and because it is so characteristic of the man: ‘This tendency to
extremes is exactly what I consider as the great source of error in political economy.’
History of economic analysis 592

13
See General Theory, pp. 362–4 and especially the essay on Malthus in Keynes’s Essays in
Biography (1933, pp. 139 47) on the controversy between Malthus and Ricardo on our subject,
where generous enthusiasm carried Keynes beyond all bounds of reason. There he punctuated his
report with applause for Malthus and derogatory comments on Ricardo’s ‘blindness’ and became
blind himself to obvious weaknesses in the former’s and to all the strong points in the latter’s
argument. But the collection of extracts he presented is nevertheless interesting, especially because
it contains some pieces that have not as yet been published elsewhere.
14
Therefore, the principal comment to make upon Malthus’ dissent from Say is not that he may not
have done justice to possible elements of truth in Say’s practical conclusions, but that he did not
understand the theory at the back of them.
15
For the meaning of these terms, see General Theory of Employment, Interest and Money, p. 25.
The warning, with respect to the concept of aggregate supply price occurs on p. 24, n. 1. This does
not alter the fact that this terminology is misleading
supply and demand functions ‘in the ordinary sense.’ But he believed that Say’s law
asserts ‘that the aggregate demand price of output as a whole is equal to its aggregate
supply price for all volumes of output’ (op. cit. p. 26); that is, he interpreted Say’s law as
Lange did later on. Our own interpretation may be restated as follows if, in order to
facilitate comparison, we waive our objection to the concepts of aggregate demand price
and aggregate supply price: the law asserts that aggregate demand price of output as a
whole is capable of being equal to its aggregate supply price for all volumes of total
output; or, alternatively, that equilibrium within total output is possible for all volumes of
output whereas equilibrium is not possible for all outputs of shoes; or, still differently,
that there is no such thing as equilibrium or disequilibrium of total output irrespective of
the relations of its components to one another.
16
If correct, this interpretation seems to
remove Keynes’s objection. Actually this is not so, however. For the weaker proposition

which asserts only the possibility of equilibria at all levels of total output and no identity
of ‘demand for and supply of total output’ still yields the further proposition—which is,
however, not equivalent to it—that competition between firms always tends to lead to an
expansion of output up to the point of full utilization of resources or maximum output.
17

And this is the proposition to which Keynes really meant to object. Since, however, the
only reason he had for objecting was that people do not spend their whole income on
consumption and do not necessarily invest the rest
18
—thus barring, according to Keynes,
the way toward ‘full employment’—it would have been more natural not to object to this
proposition either, just as we do not object to the law of gravitation on the ground that the
earth does not fall into the sun, but to say simply that the operation of Say’s law, though
it states a tendency correctly, is impeded by certain facts which Keynes believed
important enough to be inserted into a theoretical model of his own.
19

So it gets down to this. A man of the name of J.B.Say had discovered a theorem of
considerable interest from a theoretical point of view that, though
16
From this it is no great step to a more usual formulation, which will be familiar to many readers,
viz., that total output is always in neutral equilibrium. In itself it is meaningless, since there is no
equilibrium of output as a whole. But I believe that some at least of the writers who expressed
themselves in that way meant just this. If so, they stated a true proposition although in a very
misleading way.
17
If readers refer to the second paragraph on p. 26 of the General Theory, they will find that
Keynes formulates this proposition much more strongly as the drift of his argument requires. But
General economics 593

there is no justification, from any standpoint, for going beyond the formulation of the text—unless
it be that Say was just as much given to overstatement.
18
This statement is crude. Possibly rigidity of wages constitutes another reason. But we cannot go
into this here. See below, Part V, ch. 5.
19
This would have made Keynesian theory a special case of a more general theory. But Keynes
preferred to start from a pattern that contained the obstructions to full employment, which he
believed he saw, and then to look upon what he called the classical theory as the theory of the
special or limiting case in which those obstructions assume the particular value of zero.
rooted in the tradition of Cantillon and Turgot, was novel in the sense that it had never
been stated in so many words. He hardly understood his discovery himself and not only
expressed it faultily but also misused it for the things that really mattered to him. Another
man of the name of Ricardo understood it because it tallied with considerations that had
occurred to him in his analysis of international trade, but he also put it to illegitimate use.
Most people misunderstood it, some of them liking, others disliking what it was they
made of it. And a discussion that reflects little credit on all parties concerned dragged on
to this day when people, armed with superior technique, still keep chewing the same old
cud, each of them opposing his own misunderstanding of the ‘law’ to the
misunderstanding of the other fellow, all of them contributing to make a bogey of it.
5. CAPITAL
Under this heading we shall carry on our discussion about the ‘classic’ analysis of the
structure of the productive process beyond the point reached in Chapter 5. But first we
must attend to some matters of terminology.
(a) Terminological Squabbles about Wealth and Income.
No better illustrations than these squabbles can be found for what has been said above on
the futility of the ‘method’ of hunting for the meaning of words, which nevertheless we
cannot afford to neglect entirely (1) because the manner in which writers conceptualize
may serve as a measure of their analytic maturity or experience; (2) because it is
interesting to see how they fitted recalcitrant facts into the conceptual arrangements they

adopted; and (3) because in many cases terminological discussion is only the garb of
more significant things and in particular reveals parts of a writer’s analytic set-up or
model.
1

The chief divisions of ‘classic’ economics being production and distribution, the first
question seems to be what it is that is being produced and consumed. The answer was
Wealth.
2
But this only served to raise discussions on what this wealth is or, since it is
obviously identical with the goods produced and distributed (or, possibly, their value),
what ought to be included in these. These discussions display a surprising degree of
analytic immaturity. Authors wavered between wealth considered as a fund or stock and
wealth considered
1
Of this the reader can best convince himself by a perusal of Malthus’ Definitions in Political
Economy (1827), which may be called the standard work of the genus and, to repeat, merits much
History of economic analysis 594
more attention than it has received. Among other things it contains one of the best criticisms of
Ricardo’s theoretical set-up ever written (ch. 5). Also one cannot fail to admire the wisdom of the
Rules for the Definition of Terms (ch. 1).
2
Many authors, Senior in particular, in making Wealth the fundamental concept of economic
theory, emphatically disclaimed any idea of implying that Wealth was more important than
Happiness, Welfare, Virtue, and the like. As for Ricardo, it suffices to point out that the argument
for free trade that is so important a part of his work was entirely a welfare argument.
as a flow of goods;
3
they sometimes even failed to make it clear whether they meant a
social total or wealth per head; they gravely discussed the ‘problem’ of the relation

between wealth (‘riches’) and value or the ‘problem’ of the relation between social
(national) wealth and private wealth; in defining goods some were insensitive to
redundance or irrelevance of criteria; and even some of those who did not hold either a
social philosophy according to which labor alone produces the whole product or a labor
theory of value insisted on the element of human exertion as a definiens of wealth or
economic goods. To adduce instances of blemishes of this kind would serve no useful
purpose. It is sufficient to state that that discussion substantially centered on A.Smith’s
definition—material objects that are useful and transferable and cost labor to acquire or
produce—and that Senior partly improved and partly condensed this into ‘all things that
have exchange value.’ The improvement consisted in the replacement of the labor-cost
requirement by the requirement of ‘limitation of supply’: Senior at least realized clearly
the logical relation between the two, that is, the fact that limitation of supply is the
logically decisive criterion and that difficulty of attainment comes in only as one of the
factors that limit supply. But J.S.Mill did not see this clearly although he also defined
wealth by choosing ‘all useful and agreeable things’ for genus proximum and exchange
value for differentia specifica.
The way in which economists dealt with recalcitrant cases may be illustrated by the
case of human services not embodied in any physical commodity. No difficulty arose for
those who, like Lauderdale and J.B.Say, did not restrict the concept of economic goods to
material objects.
4
But those who did were faced by a spurious problem, that is, a problem
that owed its existence solely to their own conceptualization. We have already noticed,
first, an egregious instance of a verbal solution of a verbal difficulty (Ferrara’s handling
of a ‘material’ goods concept). Second, we may notice a device adopted by Senior. He
counted human beings and their ‘health, strength, and knowledge, and all the other
natural and acquired powers of body and mind’ as articles of wealth, a thing which was
done, then and later, by a great many economists.
5
And then he declared that, for

example, a lawyer does not sell services
3
The latter meaning prevailed as the popularity of the phrase, Distribution of Wealth, suffices to
show. It was the meaning adopted in the Wealth of Nations.
4
The question of nonmaterial wealth is, of course, wider than that, for such wealth also comprises
claims (which cancel out in a closed domain) and such things as patents and good will. The
question of what to do with them continued to attract undeserved attention throughout the century
and even beyond. Böhm-Bawerk’s first publication was on Rechte und Verhältnisse…(1881).
There is no need for us to go into this, however.
5
E.g. by Walras. There is a slight advantage in doing this: the three agents of production, land,
labor, and capital, then receive more symmetrical treatment. I take the opportunity to advert in
passing to the attempts that recur from time to time at evaluating statistically the economic value of
General economics 595
man. One of the best performances of this kind that belongs, however, to the next period is Ernst
Engel’s Der Wert des Menschen (1883).
but sells himself—the difference between him and a slave being that he does so by his
own will and for his own benefit and only for a definite time and purpose, whereas the
slave is sold by his owner and for good. The objections to this, however, should not be
that, legally, Senior’s construction is nonsense and that there is no such thing as a ‘sale’
for a limited time and purpose: for the construction might still be analytically convenient.
The true objection is that this conceptual arrangement offers no advantage and is
completely unnecessary. But it acquires a certain interest from the fact that Marx—and
later on Walras—adopted it also.
6

It was only toward the end of the period under survey—and then not so much in
England as on the continent of Europe—that economists started the discussion on what
‘should be’ called income, individual or national, that produced another not exactly

fascinating literature later on.
7
But we must not infer from this that the economists of that
period overlooked income aspects: elements of what we now call Income Analysis were,
on the contrary, much in evidence in their writings. The reason why the word Income
does not occur in them more often
8
is simply that they used other words. Wealth was one
of them. We have seen that the ‘classics’ were not very clear concerning the differences
between funds and flows, and between wealth and the services of
6
In Marx’s schema workmen do not sell labor (i.e. services) but their labor force or power
(Arbeitskraft). It might be urged that, in this case, this arrangement is not otiose but serves a
definite analytic purpose. We shall, in fact, see that it comes in handily for his exploitation theory.
But, quite independently of other objections to this theory, a little reflection will show that his
argument could also have been couched in terms of the labor services themselves. Moreover, the
reason why Marx liked his arrangement so much—he considered it as one of his main contributions
to economic theory—derives from a factual assumption that is patently false: he conceives that the
‘capitalist,’ having bought the laborer’s ‘force,’ then decides arbitrarily how many hours the
workman is to work. This is not so even where the labor contract does not specify hours: for these
as well as other conditions are always implied. An instructor in economics may be simply
‘appointed,’ but he knows well enough how many hours of teaching that means at the institution he
contracts with; and this applies to all kinds of einployments. The reader should make sure that he
understands why it is no objection to say that workmen, having no other sources of income but their
labor power, have no choice but to accept ‘any’ conditions—even in cases where as a matter of fact
this is or was true. With Senior, however, this construction served no such purpose and in fact none
except to remove an entirely imaginary difficulty. At the bottom of this difficulty was a disability
that Senior shared with most economists of his and even a later time: they found it surprisingly
difficult to grasp the distinction between wealth and service of wealth—so much so, in fact, that
there was some novelty about it even in 1906 when Irving Fisher, in Nature of Capital and Income,

insisted upon it.
7
Toward the end of the period under survey two factors helped to start that discussion: first, the
growing interest in income statistics (Robert D.Baxter’s National Income appeared in 1868), and,
second, the growing interest, on the Continent especially, in the problems of the income tax
(A.Held’s Die Einkommensteuer appeared in 1872).
8
This applies to the English ‘classics.’ Continental writers of the period did use it more. We have
already noticed in ch. 4 above the works of Storch and Sismondi.
History of economic analysis 596
wealth. Mostly, however, they actually meant flows of income goods (or even services)
when they spoke of wealth, so that at least in part we have already been commenting on
their concept of income when dealing with wealth. This applies in particular to A.Smith,
whose wealth is simply the ‘whole annual produce of a country’s land and labour,’
which, alternatively, he also called Gross Revenue (Book II, ch. 2, Modern Library ed., p.
271). Barring technicalities, this is substantially what we mean by Gross National
Product. This quantity minus ‘the expence of maintaining…capital’ is his Neat Revenue
or (again: substantially) our Department of Commerce National Income. Most
economists of the period discussed these definitions—some, like Say, accepting them
with minor amendments;
9
others, like Ricardo,
10
finding fault with them.
A.Smith then gave what he evidently thought was only another way of formulating the
same thing: ‘neat revenue’ or, as we should say, income was what people, individually
and collectively, ‘without encroaching upon their capital…can…spend upon their
subsistence, conveniences, and amusements’ (ibid., p. 271). This is the basis of what
became known in Germany as the Hermann-Schmoller income definition.
11

The modern
discussion on what it means to keep capital intact or to maintain capital—another
spurious problem—grew from that root.
On Productive and Unproductive Labor. We digress for a while in order to touch briefly
upon the famous controversy on productive and unproductive labor. The only reason why
this dusty museum piece interests us at all is that it affords an excellent example of the
manner in which the discussion of meaningful ideas may lose sight of their meanings and
slip off into futility. In the case before us, two meaningful distinctions may be discerned.
The one
9
I do not think that Marx was right in charging Say with the ridiculous mistake of overlooking
depreciation. All Say intended was to emphasize the basic importance of the ‘gross’ concept. See
Theorien über den Mehrwert.
10
Principles, ch. 26. This chapter read so strangely even to Ricardo himself that he felt it desirable
to insert qualifying footnotes. But on referring to this chapter, the reader will find that on the last
page of it (including footnote) Ricardo successfully corrected an error committed by A.Smith and
another error committed by Say. The first four paragraphs of the chapter seem to restrict the net
income of a country to profits and rent, and to treat wages like depreciation charges. The reason
given for this misleading arrangement is that only profits and rent constitute the national surplus
from which taxes and savings can come. But profits are, according to Ricardo himself, not or not
wholly a disposable surplus, and wages, according to his own admission, do contain, in general,
some disposable surplus—another example of Ricardo’s exasperating way of first insisting on a
proposition with tremendous energy and then blowing it up himself. But the argument points to a
unitary concept of profits plus rent—wholly foreign to Ricardo’s usual reasoning—from which
Marx may have learned something. The argument also points toward a conception of income that
may have some uses and certainly has a great hold on the popular mind, namely, the conception of
income as a surplus over necessities.
11
Hermann, see above, ch. 4. Gustav Schmoller, ‘Die Lehre von Einkommen…,’ Zeitschrift für die

gesamte Staatswissenschaft, 1863.


General economics 597
springs from the fact that a private-enterprise system generates incomes that provide for
consumption in two ways: directly for the consumption of those who ‘earn’ them, and
indirectly for the consumption of those who are ‘supported’ by them, for example,
children and the retired aged. It stands to reason that the relation between the two, in our
example determined (in part) by the age distribution of the population, is no matter of
indifference but on the contrary one of the most important characteristics of a society’s
economic life. Controversies about the question whether or not there are also types of
employment that, for some purposes or for all, should be treated as ‘supported’ out of the
incomes earned in the business process, for example, whether public officers should be so
treated on the ground that their incomes derive from the taxation of other incomes, may
be entirely meaningful.
12
The other meaningful distinction springs from the fact that
services of labor (or of natural agents) that are directly bought and consumed by
households, such as the services of servants, teachers, and physicians, occupy a position
in the economic process that is different from the position of services of labor that are
bought and ‘consumed’ by firms and have, economically speaking, still to go through a
business process. That this is not a distinction without a difference—although, of course,
these services, in the form of products, also reach the consumers’ sphere eventually—is
readily seen from the fact sufficiently expressed by the common slogan that these
services are paid from some firm’s capital, whereas the former are paid from some
household’s income or revenue.
13
So soon as the servant has received his wages or their
equivalent in goods, there is no further problem. When the factory worker has received
his wages, further problems arise of selling the product he has helped to produce, of lags,

risks, discounts, and so on, all of which are pertinent to the determination of those wages
themselves. Thus the distinction is indeed relevant to the structure of the economic
process and imposes itself upon the analyst at many turns of his way (e.g. in matters of
the wage-fund doctrine, see below, subsec. 6f).
It will be seen that these two distinctions are entirely independent of one another: each
has meaning without reference to the other. But both—and a lot of confusion in
addition—were bequeathed to the writers of that period by A.Smith. On the first page of
his Introduction, he placed great emphasis upon ‘the proportion between the number of
those who are employed in useful labour, and that of those who are not so employed.’ For
lack of space, I must leave it to the reader to satisfy himself that, with an admixture of
extraneous matter, this passage really adumbrates the meaning of our first distinction.
Only it does so hazily and, by employing the vague term ‘useful,’ gives the clue to all the
confusion that disfigured the subsequent controversy on productive and unproductive
labor, though this phrase does not figure in Book I of the Wealth of Nations. This phrase
does emerge in Chapter 3 of Book II, where A.Smith, having experienced physiocratic
influence, developed his theory of Accumulation. He had no use, of course, for the
physiocratic proposition that only labor employed in agriculture is productive any
12
Compare the controversies that have arisen in our own time in connection with the statistics of
national income on the question whether or not public administration should be considered as an
industry like any other so that there would be no analytically relevant difference between the salary
of a government official and the wages of, say, a workman in a motor-car factory.
13
As will be pointed out presently, this must not be confused with the case of people who live on
derived income in the sense intended under the first distinction, e.g. the case of the retired aged.
History of economic analysis 598
more than he had use for the ‘mercantilist’ proposition that only labor employed in export
industries is. But pouring away the physiocrat wine, he retained the bottles and filled
them with wine of his own: he defined labor as productive that ‘adds to the value of the
subject upon which it is bestowed’ (op. cit. p. 314) and exemplified this by the case of

factory workers who, as he adds by way of explanation (ibid. p. 316), live on ‘that part of
the annual produce of the land and labour which replaces capital’ (with a profit); and he
defined labor as unproductive that does not add (exchange) value to anything and
exemplified this by the labor of the menial servant and that ‘of some of the most
respectable orders in the society’ such as the sovereign ‘with all the officers both of
justice and war who serve under him’ and ‘are maintained by part of the annual produce
of the industry of other people.’ Two things are clear: he had got hold of our second
distinction; and he had confused it with the first.
The first man to see this quite clearly was Marx, who adopted our second distinction,
giving A.Smith ample credit for having uncovered so important an element of the
structure of capitalist society, and pointing out that this piece of insight was, in the work
of A.Smith, wrapped in considerations that Marx thought superficial and in any case
quite unconnected with it.
14
Of course, nobody missed the point entirely—most writers
made, tacitly or explicitly, use of it when analyzing the demand for labor. But when they
discussed the distinction as such, they lost sight of it and always thought of the first
distinction. Nor is this all. We have seen that this distinction also may be made
meaningful. But giving themselves up to the associations evoked by the terms ‘useful’
and ‘productive,’ economists concentrated on such ‘issues’ as which activities were
worthy of these honorific epithets. Teachers and civil servants did not like to be called
‘unproductive,’ feeling—sometimes rightly and sometimes wrongly—that this phrase
was intended to carry a derogatory meaning.
15
And so a meaningless discussion became a
standard item of nine-
14
Marx elaborated this at length in his discussion of Smithian doctrine in the Theorien über den
Mehrwert. From his standpoint the decisive distinction was between labor that does ‘produce
surplus value’ and labor that does not. But the distinction between labor that is paid from business

capital and labor that is paid from ‘revenue’ is preferable: a servant might work more hours than
are embodied in the ‘value’ of his labor and hence might be ‘exploited’ just as may a factory
worker. The former’s employer may also derive a surplus. The point is, to continue in Marxist
language, that this surplus need not be ‘realized’ in any market.
15
Some such feelings assert themselves again each time modern economists discuss, for the
purposes of national-income statistics, the conceptual treatment to be accorded to government
salaries.





General economics 599
teenth-century textbooks in spite of the increasing awareness of its futility, which
eventually killed it. An account of all the ramifications and of all the misspent ingenuity
that sometimes went into it would fill a volume. But it could serve one purpose only,
namely, to display the word-mindedness of economists and their inability to tell a real
problem from a spurious one.
16
[J.A.S. intended to have this digression On Productive
and Unproductive Labor printed in small type so that the average reader could skip it
easily.]
(b) The Structure of Physical Capital.
17

On its most abstract level, the analysis of economic choice, which is really all that is
involved in what we are accustomed to learn in the particular form of a theory of value,
can be carried out in terms of unspecified things called ‘goods’ that have no other
properties than those of being desired and of being scarce. It stands to reason, however,

that in order to make headway beyond the most arid generalizations we must pick, from
our vision of reality, further restrictions upon economic choice such as are implied in our
‘know-how’ or, less colloquially, in the limitations of a given technological horizon,
which will permit some, and exclude other, transformations of our initial stock of goods.
In any case, we must postulate given wants, a given technological horizon, given
environmental factors such as land and personnel of given kinds and qualities, and a
given stock of produced goods with which to start. But this is not enough. This initial
stock of goods is neither homogeneous nor an amorphous heap. Its various parts
complement each other in a way that we readily understand as soon as we hear of
buildings, equipment, raw materials, and consumers’ goods. Some of these parts must be
available before we can operate others; and various sequences or lags between economic
actions impose themselves and further restrict our choices; and they do this in ways that
differ greatly according to the composition of the stock we have to work on.
18
We express
this by saying that the stock of
16
This discussion induced a related one on the concepts of productive and unproductive
consumption that may be illustrated by a statement of Senior’s (op. cit. p. 57): ‘if a judge…required
by his station to support an establishment costing £2,000 a year, should spend £4,000, half of his
consumption would be productive and the other half unproductive.’ Productive consumption, then,
was ‘that use of a product which occasions another product.’ The idea that commodities and
services do not leave the economic process for good as soon as they enter the sphere of the
households that consume them, but that they ‘produce’ there the productive services of the
members of these households, turns up again and again. In our day it has been adopted by Leontief,
in whose system households are treated as an industry that consumes productively like any other.
17
Some readers will find this subsection difficult reading. It attempts to explain an unconventional
view of the role of physical ‘capital’ within the logic of the economic process that might be pressed
into this phrase: from the standpoint of analysis, capital means a set of restrictions. This will

become quite clear before long, and I believe that the reader will derive some benefit from taking
the trouble involved in mastering this subsection.
18
The stock of wealth of all kinds that exists at an instant of time has been called Capital by Irving
Fisher (Nature of Capital and Income, 1906, p. 52), who has successfully shown that, properly
thought out, the capital definitions of most, if not all, of
History of economic analysis 600
goods existing at any instant of time is a structured quantity or a quantity that displays
structural relations within itself that shape, in part, the subsequent course of the
economic process. Naturally we wish, for the purposes of pure theory, to reduce these
structural characteristics to as few and as general ones as possible, steering as best we can
between the Scylla of unmanageable lifelikeness and the Charybdis of sterile simplicity.
Ever since the time of Cantillon and Quesnay, when scientific model-building began,
economists have, of course, been aware of all this. In the preceding chapter we have
already had a glimpse of the manner in which the writers of the ‘classic’ period took—
haltingly—the first two steps in the analysis of the structural properties of the economic
process: the one was to recognize capital as a ‘requisite’ of production and the other was
to adopt the physiocrat (Cantillon-Quesnay) idea of ‘advances.’ We have now to fill in
the more important of the remaining elements of this analysis, which constitutes what is
commonly known as theory of capital.
The reader need not be afraid that we shall have to wade through another morass of
verbal controversy. The theory of capital does indeed enjoy a reputation for this kind of
thing that is rivaled by few other fields. People kept on asking the meaningless question:
What is Capital? And some have tried to answer it by speculations about the original
meanings of the words caput, capitale,
and the like. Senior even held that
‘the term Capital has been so variously defined that it may be doubtful whether it have
any generally received meaning’ (Outline, p. 59). In a sense, this is true.
19
But it is true

only: first, owing to relatively minor faults of conceptualization committed by the
individual authors, and these we can disregard if their analytic intention is clear enough;
second, owing to the wish, the father of so many futile controversies, to have a unitary or
all-purpose concept of capital, and this wish I do not share; third, owing to the not less
unwarranted wish of many authors to approximate the ‘capital’ that is useful in their
analysis to either the asset or the liability side of a business concern’s balance sheet;
fourth, owing to occasional waverings between physical capital concepts, on the one
hand, and monetary concepts, on the other, which will be noticed in the next footnote.
For the rest, the matter is much simpler than it looks because there really is only one
dominant analytic purpose to describe that practically all the leading economists tried to
serve.
Since it was a requisite of production, capital consisted of goods.
20
More-
the writers of the period under survey come precisely to this. We shall not adopt Fisher’s concept
here but instead we shall simply use the excellent Smithian term Stock. This will make it easier to
distinguish it from various other meanings of the term Capital without adding each time: ‘in our
sense.’
19
Readers interested in the history of economists’ use of the term are referred to Irving Fisher (op.
cit. ch. 4, 2) or to the chapter on the concept of capital in the second volume of Böhm-Bawerk’s
great work.
20
But even those authors who expressed themselves most strongly in favor of a physical capital
concept sometimes drifted into the preserves of a monetary one. Ricardo and J.S.Mill, among
others, occasionally penned sentences which carry meaning only if they refer to monetary capital.
This has been critically noticed, for the first time I think, by Tchernychevsky, in L’économie
politique jugée par la science (1874), an elaborate analysis of Mill’s Principles. It is still more
obvious in the Manual of Mill’s follower, Fawcett. With Mill, capital is ‘expended’ on raw
General economics 601