In Quesnay's account, the difference between fixed and circulating capital shows up as the difference between original advances and annual advances. He gets this right: it's a difference inside productive capital, capital that has gone into the actual process of production. Because he treats capital used in agriculture — the tenant farmer's capital — as the only really productive kind, these two kinds of advance only show up for the farmer's capital. That's also where the different turnover times come from: one part turns over every year, the other over more than a year — ten years. As the Physiocrats developed their ideas, they went on to apply this same distinction, almost in passing, to other kinds of capital too, to industrial capital in general. For society as a whole, the distinction between advances that return within a year and advances that return over several years stays so important that many economists, even after Smith, come back to it.

The difference between the two kinds of advance only exists once money that's been advanced has already been turned into the elements of productive capital. It is a distinction that lies entirely within productive capital — nowhere else. So it never occurs to Quesnay to count money itself as either an original advance or an annual advance. As advances for production — that is, as productive capital — both kinds stand opposite money, and opposite the commodities sitting on the market. And Quesnay is right about something further: the difference between these two elements of productive capital comes down to the different way each one enters the value of the finished product — and so to the different way its value travels along with the product's value in circulation, and so to the different way it gets replaced or reproduced: the value of one part is replaced whole every year, the value of the other only bit by bit, over longer stretches of time.

The one step forward Smith makes is to generalize the categories. In his hands they no longer apply to one special form of capital, the tenant farmer's capital, but to every form of productive capital. So it follows on its own that in place of the distinction — drawn from agriculture — between annual turnover and multi-year turnover, we get the general distinction between turnovers that take different lengths of time: one turnover of fixed capital always spans more than one turnover of circulating capital, whatever length of time those turnovers of circulating capital take — a year, more than a year, or less. This is how, in Smith, the annual advances turn into circulating capital and the original advances turn into fixed capital. But that generalizing of the categories is the whole of his advance. In working it out, he falls a long way behind Quesnay.

The raw, empirical way Smith opens the inquiry brings the confusion in with it from the start.

"There are two different ways in which a capital may be employed so as to yield a revenue or profit to its employer." (Wealth of Nations, Book II, chapter I, p.185, Aberdeen edition, 1848.)

The ways value can be invested to function as capital — to yield surplus-value for its owner — are just as varied, just as many, as the different spheres capital can be invested in. That's a question about the different branches of production capital can go into. Put this way, the question reaches further still: it also covers how value can function as capital for its owner even when it isn't invested as productive capital at all — say, as interest-bearing capital, or merchant's capital. So we are already a world away from what the inquiry is actually about: how splitting productive capital into its different elements — leaving aside which sphere it's invested in — affects its turnover.

Smith then goes straight on:

"First, it may be employed in raising, manufacturing, or purchasing goods, and selling them again with a profit."

All Smith tells us here is that capital can be used in agriculture, manufacture, and trade. So he is only talking about the different spheres capital can be invested in — including ones, like trade, where capital isn't part of the immediate process of production at all, and so isn't functioning as productive capital. With that he has already left the ground on which the Physiocrats set out the differences within productive capital and their effect on turnover. Worse, he immediately takes merchant's capital as his example, in a question that is only about differences within productive capital in the process of forming product and value — differences that in turn produce differences in its turnover and reproduction.

He goes on:

"The capital employed in this manner yields no revenue or profit to its employer while it either remains in his possession or continues in the same shape."

The capital employed in this manner! But Smith is talking about capital invested in agriculture, in industry — and he tells us later that capital invested this way splits into fixed and circulating! So being invested this way can't be what makes capital either fixed or circulating.

Or did he mean something else: that capital used to produce goods and sell them at a profit has, once turned into goods, to be sold — and that by being sold it passes, first, out of the seller's possession into the buyer's, and second, out of its natural form as a commodity into its money-form — and that it is therefore useless to its owner as long as it stays in his possession, or — for him — in the same form? But then the point comes down to this: the very same capital-value that had earlier been functioning in the form of productive capital, a form that belongs to the production process, is now functioning as commodity-capital and money-capital, in forms that belong to the circulation process — so it is no longer fixed or fluid capital at all, either one. And this holds just as much for the value added through raw and auxiliary materials — that is, through the fluid part of productive capital — as for the value added through the wearing-out of instruments of labour, that is, through fixed capital. This way, too, we don't get one step closer to the distinction between fixed and fluid capital.

Further:

"The goods of the merchant yield him no revenue or profit till he sells them for money, and the money yields him as little till it is again exchanged for goods. His capital is continually going from him in one shape, and returning to him in another, and it is only by means of such circulation, or successive exchanges, that it can yield him any profit. Such capitals, therefore, may very properly be called circulating capitals."

What Smith is calling circulating capital here is what I want to call capital of circulation: capital in the form it takes on in the circulation process, the change of form that comes with exchange — a change of substance and a change of hands — that is, commodity-capital and money-capital, as against the form it takes in the production process, that of productive capital. These aren't separate kinds that the industrial capitalist splits his capital into; they are different forms that the very same advanced capital-value takes on and sheds, one after another, over its life story. This is where Smith — and it's a great step backward from the Physiocrats — throws together two different things: the differences of form that arise within the circulation of capital-value, in its circuit through its successive forms, and the differences that arise while the capital-value is in the form of productive capital, from the different ways the different elements of productive capital take part in forming value and pass their value on to the product. We'll see the consequences of this basic confusion further on — the confusion between productive capital and capital in the sphere of circulation (commodity-capital and money-capital) on the one hand, and between fixed and fluid capital on the other. The capital-value advanced as fixed capital travels through the product in circulation just as much as the capital-value advanced as fluid capital does, and it turns into money-capital through the circulation of commodity-capital just as much as the other does too. The only difference is that its value circulates in instalments, and so also has to be replaced in instalments, over shorter or longer periods, and reproduced in its own physical form.

That Smith means nothing here by circulating capital except capital of circulation — that is, capital-value in the forms it takes on in the circulation process, commodity-capital and money-capital — is shown by the especially clumsy example he picks. He takes as his example a kind of capital that doesn't belong to the process of production at all, that only lives in the sphere of circulation, that consists only of capital of circulation: merchant's capital.

How absurd it is to start with an example where capital doesn't figure as productive capital at all, he says himself right afterwards:

"The capital of a merchant is altogether a circulating capital."

But the distinction between circulating and fixed capital is supposed — as we're told later — to arise from essential differences within productive capital itself. So on one hand Smith has the Physiocratic distinction in mind, and on the other the differences of form that capital-value passes through in its circuit. And the two get run together every which way.

But how a profit is supposed to arise from the change of form between money and commodity, from the mere conversion of value out of one of these forms into the other, simply cannot be seen. And the explanation becomes completely impossible here, because he starts with merchant's capital, which moves only within the sphere of circulation. We'll come back to this; for now, let's hear what he says about fixed capital.

Capital can also go into improving land, or into buying useful machines and tools, or into other such things that bring in an income or profit without changing owners or circulating any further. Capital of that kind, Smith says, can rightly be called fixed capital. Different trades need very different proportions of fixed and circulating capital. Every master craftsman or manufacturer must have some part of his capital tied up in the tools of his trade — a small part in some trades, a large part in others. But by far the greater part of the capital of craftsmen like tailors, shoemakers, and weavers is circulating: it goes out as wages to their workers or as payment for their materials, and comes back with a profit through the price of the finished work.

Setting aside the naive idea about where profit comes from, the weakness and confusion show up right away. Take a machine-maker, for instance: for him the machine is a product that circulates as commodity-capital — in Smith's own words, something that:

is parted with, changes hands, and goes on circulating.

So on his own definition, the machine would not be fixed capital at all — it would be circulating capital. Once again, this confusion comes from mixing up two different things. One is the distinction between fixed and fluid capital, which comes from the different ways the different elements of productive capital circulate. The other is a difference of FORM that the very same capital passes through: as productive capital while it is at work inside the process of production, and as circulation-capital — commodity-capital or money-capital — once it is out in the sphere of circulation. So depending on which stage of capital's life a thing happens to occupy, the very same things can count, on Smith's own reasoning, as fixed capital (when they are means of labour, elements of productive capital) and also as "circulating" capital, commodity-capital (when they are the product just cast out of production into circulation).

But Smith suddenly switches his whole basis for the distinction, contradicting what he had used to open the entire inquiry only a few lines before. This happens above all in the sentence:

"There are two different ways in which a capital may be employed so as to yield a revenue or profit to its employer" — namely, as circulating capital or as fixed capital. On this reading, then, these are two different ways of using two different capitals, independent of one another — the way capital might go into industry, say, or into agriculture instead. But now it says:

"Different trades need very different proportions of fixed and circulating capital employed in them."

Now fixed and circulating capital are no longer two separate, independent investments — they are different portions of one and the same productive capital, which in different lines of business make up different shares of that capital's total value. So this is a distinction that comes from how productive capital itself naturally splits up, and it only holds good with reference to that capital. But this again contradicts something else: elsewhere merchant's capital gets set up as simply circulating capital, opposed to fixed capital. Smith himself says:

"A merchant's capital is entirely circulating capital."

It really is capital that functions only within the sphere of circulation, and as such it stands opposed to productive capital in general — to capital that is bound up in the process of production. But for that very reason, it cannot also stand opposed, as the fluid (circulating) portion of productive capital, to the fixed portion of productive capital.

In the examples Smith gives, he calls fixed capital the tools of the trade, and circulating capital the part of capital laid out on wages and on raw materials — auxiliary materials included — which comes back with a profit through the price of the finished work.

So the starting point is simply the different pieces that make up the labour process: labour-power (labour) and raw materials on one side, means of labour — tools — on the other. These count as parts of capital because a sum of value meant to function as capital has been laid out on them. In that sense they are the material elements, the actual forms taken by productive capital, capital at work in the process of production. So why is one part called fixed? Because, as Smith puts it, "some part of the capital must be fixed in the instruments of trade."

But the other part is just as much fixed — in wages and in raw materials. Yet machines and "instruments of trade... such like things... yield a revenue or profit without changing masters, or circulating any further. Such capitals, therefore, may very properly be called fixed capitals."

Take mining, for example. Here no raw material gets used at all, since the object being worked on — say, copper — is a natural product that labour has to appropriate in the first place. The copper still to be appropriated, the product of the process, which later circulates as a commodity, or commodity-capital, forms no element of productive capital: no part of its value has been laid out there. On the other side, the other elements of the process — labour-power and auxiliary materials like coal, water, and so on — likewise do not physically enter the product. The coal is used up entirely, and only its value passes into the product, exactly as a portion of a machine's value passes into the product. And the worker remains just as separate from the product, the copper, as the machine does. Only the value he produces through his labour becomes part of the value of the copper. So in this example, not a single element of productive capital changes hands, and none of them circulates any further, because none of them physically enters the product. Where, then, is the circulating capital here? On Smith's own definition, the whole capital used in a copper mine would consist of nothing but fixed capital.

Take a different industry instead, one that uses raw materials which actually form the substance of the product, plus auxiliary materials — like heating coal — that enter the product only by their value, not bodily. With the product — yarn, say — the raw material it is made from, cotton, also changes hands and passes out of the process of production into the process of consumption. But so long as the cotton is functioning as an element of productive capital, its owner does not sell it — he works it up, has it made into yarn. He does not let it out of his hands. Or, to use Smith's crudely false and trivial phrase, he makes no profit "by parting with it, by its changing masters, or by circulating it." He no more lets his materials circulate than he lets his machines circulate. They are fixed in the process of production, exactly as the spinning machines and the factory buildings are. Indeed, a part of productive capital must be just as constantly fixed in the form of coal, cotton, and so on, as in the form of means of labour. The only difference is that the cotton, coal, and so on needed for, say, a week's production of yarn get completely used up producing that week's output, and so must constantly be replaced by fresh supplies. So these elements of productive capital, though they stay identical in kind, constantly consist of fresh individual supplies of the same kind — whereas the very same individual spinning machine, the very same individual factory building, goes on contributing, without being replaced by a new individual of its kind, to a whole series of weekly productions. As elements of productive capital, all its components are constantly fixed in the process of production, since it cannot proceed without them. And all the elements of productive capital, fixed and fluid alike, stand equally, as productive capital, opposed to circulation-capital — that is, to commodity-capital and money-capital.

The same holds for labour-power. A part of productive capital must be constantly fixed in it too, and it is the very same individual workers, just like the very same individual machines, that a given capitalist goes on using over a long stretch of time. The difference between them and the machines does not lie in the machine being bought once and for all — which isn't even true in every case, since it may be paid off in instalments. The difference lies in this: the labour the worker expends passes entirely into the value of the product, whereas the value of the machine passes in only bit by bit.

Smith mixes up different definitions when he says of circulating capital, as against fixed capital:

"Capital employed in this manner yields no revenue or profit to its employer while it either remains in his possession or continues in the same shape."

He puts on the same level two things that are quite different: the merely FORMAL change a commodity — the product, the commodity-capital — goes through in the sphere of circulation, which mediates the change of hands of commodities, and the BODILY change that the different elements of productive capital go through during the process of production. Turning commodity into money and money into commodity, buying and selling, he simply lumps together here with turning elements of production into a product. His example for circulating capital is merchant's capital: capital turned from commodity into money, and from money into commodity again — the change of form C-M-C that belongs to the circulation of commodities. But for functioning industrial capital, this change of form within circulation matters because the commodities the money gets turned back into are elements of production, means of labour and labour-power. That is what keeps the capital's function going continuously, keeps the process of production going as a continuous process, a process of reproduction. This whole change of form takes place in circulation; it is what mediates the actual passing of commodities from one hand to another. The changes that productive capital goes through within its own process of production, by contrast, belong to the labour process: they are necessary in order to turn the elements of production into the intended product. Smith fastens on the fact that part of the means of production — the actual means of labour — serves within the labour process (which he wrongly describes as "yielding a profit to their master") by not changing its natural form, only wearing away gradually; while another part, the materials, does change, and fulfils its role as means of production precisely through that change. But this difference in how the elements of productive capital behave physically within the labour process only forms the STARTING POINT for the distinction between fixed and non-fixed capital — it is not that distinction itself, as is shown by the fact that it holds equally for every mode of production, capitalist or not. What corresponds to this difference in physical behaviour is how value is passed on to the product, and correspondingly how that value gets replaced through the sale of the product — and it is only this that makes up the real distinction. So capital is not fixed because it is fixed in the means of labour, but because part of the value laid out in the means of labour stays fixed there, while another part circulates as it passes into the value of the product.

Smith: "If capital is invested to bring in future profit, it must bring that profit either by staying with its owner, or by leaving him. In the one case it is fixed capital, in the other it is circulating capital."

The first thing that stands out here is a crude, everyday notion of profit, taken straight from how an ordinary capitalist sees things — and it flatly contradicts Smith's own better, deeper insight elsewhere. In the price of the product, the cost of the materials is replaced, and the cost of labour-power is replaced, and so is the portion of value that the tools and machines handed over to the product through wear and tear. None of this replacing is ever, in any case, where profit comes from. Whether a value advanced for production comes back whole or in pieces, all at once or gradually, through the sale of the product — that can only change the manner and timing of the replacing. It can never turn what both kinds of replacement have in common — simply getting the value back — into the creation of surplus-value. Underneath this lies the ordinary idea that, because surplus-value is only realized once the product is sold, once it circulates, it must therefore spring from the sale, from circulation itself. In truth, this talk of profit "arising differently" is just a false way of saying that the different elements of capital serve differently — that they work differently as productive elements within the labour process. And in the end the difference isn't drawn from the labour process, from the process by which value grows, from the function of productive capital itself — it's only meant to hold subjectively, for the individual capitalist, for whom one part of his capital happens to be useful in one way and another part in another way.

Quesnay, by contrast, had drawn these distinctions from the process of reproduction itself and from what that process needs. For it to run without interruption, the value of the year's running advances has to be replaced in full, every year, out of the value of the year's product. The value of the fixed investment, though, is only replaced bit by bit, so that it takes a stretch of years — ten, say — before it's replaced in full and must be fully renewed, that is, swapped out for new specimens of the same kind. Smith falls well below Quesnay here.

So what's left of Smith's definition of fixed capital comes down to just this: it's the tools that keep their shape through the production process and go on serving in production, until they wear out, as against the products they help to make. This forgets that every element of productive capital — tools, materials, labour-power alike — constantly stands, in its own physical form, opposite the product, and opposite that product once it circulates as a commodity. And the only real difference between the part made up of materials and labour-power and the part made up of tools is this: labour-power is bought fresh each time, not bought once for the whole span it will serve, the way tools are; materials aren't the same identical stuff either — fresh specimens of the same kind keep entering the labour process. At the same time this creates the false impression that the value of fixed capital never circulates at all — even though Smith himself had already shown, earlier, that the wear and tear of fixed capital becomes part of the price of the product.

When circulating capital is set against fixed capital, it goes unremarked that it only stands in that opposition as the part of productive capital that has to be replaced in full out of the value of the product, and so has to go all the way through the product's changes of form — whereas fixed capital does not. Instead it gets lumped together with the shapes capital takes on when it passes out of production and into circulation — commodity-capital and money-capital. But both of those forms, commodity-capital and money-capital, carry the value of the fixed part of productive capital just as much as the fluid part. Both are capital-of-circulation, as against productive capital — but that is not the same as circulating (fluid) capital as against fixed capital.

Finally: through this whole lopsided story — fixed capital makes its profit by staying inside production, circulating capital by leaving it and being circulated — something gets hidden. Variable capital and the fluid part of constant capital move through turnover in exactly the same form. That shared form hides the real difference between them in the process by which value grows and surplus-value gets made — which only darkens further the whole secret of capitalist production. And through the shared label "circulating capital," this essential difference gets wiped out. Later economics carried this even further, treating not the opposition of variable and constant capital, but that of fixed and circulating capital, as the essential thing, the only distinction that mattered.

Having first described fixed and circulating capital as two particular ways of investing capital, each of which, taken on its own, brings in a profit, Smith says:

"No fixed capital can yield any revenue but by means of a circulating capital. The most useful machines and instruments of trade will produce nothing without the circulating capital which affords the materials they are employed upon, and the maintenance of the workmen who employ them."

Here it becomes clear what those earlier phrases — "yield a revenue," "make a profit," and so on — actually mean: simply that both parts of capital serve to help form the product.

Smith now gives the following example:

"That part of the capital of the farmer which is employed in the implements of agriculture is a fixed, that which is employed in the wages and maintenance of his labouring servants is a circulating capital."

(Here, then, the difference between fixed and circulating capital correctly refers only to the different way different parts of productive capital circulate — to their different turnover.)

"He makes a profit of the one by keeping it in his own possession, and of the other by parting with it. The price or value of his labouring cattle is a fixed capital" — here again Smith gets it right that the difference turns on value, not on the physical thing — "in the same manner as that of the instruments of husbandry; their maintenance is a circulating capital, in the same way as that of the labouring servants. The farmer makes his profit by keeping the labouring cattle, and by parting with their maintenance."

(The farmer keeps the feed for his cattle — he doesn't sell it. He consumes it as feed, just as he consumes the cattle itself as a working tool. The only real difference is this: the feed that goes into keeping the cattle working gets wholly used up, and has to be constantly replaced with fresh feed out of the farm's produce or out of its sale. The cattle themselves are only replaced one at a time, as each animal in turn becomes unfit to work.)

"Both the price and the maintenance of the cattle which are bought in and fattened, not for labour but for sale, are a circulating capital. The farmer makes his profit by parting with them."

(Marx's correction:) Every producer of commodities, including the capitalist one, sells his product, the result of his production process — but that doesn't mean the product itself forms either the fixed or the fluid part of his productive capital. By the time it's sold, it exists in a different form altogether: it has already been thrown off from the production process and now has to function as commodity-capital. Cattle bought for fattening function in the production process as a raw material, not as a working tool the way labouring cattle do. They enter the product as its very substance, and their whole value passes into it, the same way the value of an auxiliary material — the feed — does. That is why fattening cattle are a fluid part of productive capital — not because the finished product, the fattened animal, happens to have the same physical shape as the raw material, the not-yet-fattened animal. That's just an accident. At the same time, this very example could have shown Smith that it isn't the physical shape of a productive element that makes the value lodged in it fixed or fluid — it's the function that element performs within the production process.

"The whole value of the seed too is a fixed capital. Though it goes backwards and forwards between the ground and the granary, it never changes masters, and therefore it does not properly circulate. The farmer makes his profit not by its sale, but by its increase."

Here the sheer thoughtlessness of Smith's distinction comes into the open. On his account the seed would be fixed capital if no change of masters happened — that is, if it were replaced directly out of the year's product, simply deducted from it. But it would be circulating capital instead if the whole product were sold and someone else's seed-corn bought back with part of its value. In the one case a change of masters happens; in the other it doesn't. Smith is confusing fluid capital with commodity-capital all over again. The product is indeed the material carrier of commodity-capital — but of course only the part of it that actually goes into circulation, not the part that goes straight back into the production process it came out of as product.

Whether the seed is deducted directly as part of the product, or the whole product is sold and part of its value used to buy someone else's seed instead — in both cases only replacement happens, and this replacement makes no profit either way. In the one case the seed enters circulation as a commodity together with the rest of the product; in the other it only figures in the bookkeeping as a value-component of the capital advanced. But in both cases it stays a fluid part of productive capital. It is wholly used up in finishing the product, and it has to be wholly replaced out of that product for reproduction to be possible at all.

As Volume I put it: raw materials and auxiliary materials lose the independent shape they had as use-values when they entered the labour process. Not so with the actual instruments of labour. A tool, a machine, a factory building, a vessel, and so on, serve in the labour process only for as long as they keep their original shape, entering the labour process again tomorrow in exactly the same form as yesterday. And just as they keep their independent shape against the product throughout their working life, so they keep it after that life is over. The carcasses of machines, workshops, and factory buildings still exist on their own, separate from the products they helped to make.

There is a real difference in how means of production get used up in forming the product: some keep their own shape against the product, others change it or lose it entirely. This difference belongs to the labour process as such — it holds just as much for labour processes aimed at simple self-sufficiency, like that of a family producing only for itself, with no exchange and no commodity production at all. Smith falsifies this real difference in two ways. First, he drags in a notion of profit that has no business here — as though some elements bring the owner a profit by keeping their shape, others by losing it. Second, he lumps together the changes some elements of production undergo within the labour process with a completely different change of form — the one that belongs to the exchange of products, to the circulation of commodities, buying and selling — which also carries with it a change of ownership in the commodities that circulate.

Turnover assumes that reproduction is carried out through circulation — that is, through selling the product, turning it into money, and turning that money back into the elements of production. But where part of his own product serves the capitalist producer directly again as a means of production, he in effect appears as seller of it to himself, and that is how it figures in his books. This part of reproduction is then not carried out through circulation, but directly. The part of the product that re-enters production this way still replaces fluid capital, not fixed capital — as long as, first, its whole value passes into the product, and second, it is itself fully replaced, in kind, by a new specimen out of the new product.

Smith now tells us what fixed and circulating capital are made of. He lists off the actual things that count as fixed capital and the things that count as circulating capital, as though this belonged to the things themselves by nature, rather than coming from the particular job a thing does within capitalist production. Yet in this same chapter he notes that a house kept for someone to live in, though it "may yield a revenue to its proprietor, and thereby serve in the function of a capital to him," cannot yield any revenue to the public that way, nor serve as capital to the public — and the income of the whole people can never be increased by it in the slightest.

So Smith is saying plainly here that being capital is not a property a thing has in itself, under all conditions. It is a function — something a thing wears at some times and not at others, depending on circumstances. And what holds for capital in general holds for its subdivisions too.

The very same things belong to fluid capital or to fixed capital depending on what job they do in the labour process. Take cattle: as a draft animal — a tool for work — an ox constitutes a material mode of existence of fixed capital. As livestock being fattened up — raw material — the same ox is part of the farmer's circulating capital. The same thing can also switch between being part of productive capital at all and belonging to the fund people consume from directly. A building used as a workshop is a fixed part of productive capital. The same kind of building used as a dwelling is no form of capital at all, precisely because it's a dwelling. In many cases the very same tools can serve now as means of production, now as means of consumption.

This was one of the mistakes that followed from Smith's way of thinking: treating fixed and circulating as characters belonging to the things themselves. The analysis of the labour process already showed how the roles of means of labour, material worked on, and product change depending on what part the same thing is playing in the process at a given moment. The distinction between fixed and non-fixed capital is itself built on top of those particular roles — the roles these elements play in the labour process, and therefore also in the process that forms value.

There is a second problem. In listing the things that make up fixed and circulating capital, it becomes completely clear that Smith mixes up two different distinctions. One is the distinction between fixed and fluid parts — which only makes sense for productive capital, capital in its productive form. The other is the distinction between productive capital on one hand and, on the other, the forms capital takes during its circulation: commodity-capital and money-capital. He writes, in the same place:

Circulating capital, he says, consists of the food, materials, and finished goods of every kind sitting in the hands of the merchants who deal in them, plus the money needed to circulate and distribute all of it.

Look closely, though, and here — differently from before — circulating capital gets equated again with commodity-capital and money-capital. But those are two forms of capital that don't belong to the production process at all. They aren't fluid capital as opposed to fixed capital; they are circulation-capital as opposed to productive capital. It is only alongside these that the parts of productive capital advanced as materials — raw material or semi-finished goods, actually built into the production process — turn up again. He writes:

The third and last of the three portions the whole stock of society naturally divides into is circulating capital, whose mark is that it only brings in revenue by circulating — by changing hands. This too has four parts: first, money...

(But money is never productive capital — never capital functioning in the production process. It is always just one of the forms capital takes while moving through circulation.) Second, the stock of food sitting with the butcher, the grazier, the farmer, and so on, which they hope to sell at a profit. Fourth and last, finished goods still sitting in the hands of the merchant or manufacturer. And third, materials — whether raw or worked up to some degree — for clothes, furniture, and building, not yet turned into any of those three finished shapes, still in the hands of growers, manufacturers, silk and cloth dealers, timber merchants, carpenters and joiners, brickmakers, and so on.

Items 2 and 4 are nothing but products that have already been pushed out of the production process and now have to be sold — in short, they now function as commodities, as commodity-capital. In this form they occupy a place in the process where they are no longer an element of anyone's productive capital, whatever they're eventually meant for — whether they end up feeding individual consumption or feeding production. The goods under 2 are foodstuffs; the goods under 4 are all the other finished products, which themselves break down into finished tools on one hand and finished consumer goods (other than the food already counted under 2) on the other.

That Smith brings the merchant into this shows the confusion. Once the producer has sold the goods to the merchant, they are no longer any part of the producer's capital at all. Looked at socially, the goods are still commodity-capital, just now sitting in someone else's hands rather than the producer's — but precisely because they are commodity-capital, they are neither fixed capital nor fluid capital.

Whenever production isn't just for the producer's own direct use, the product has to circulate as a commodity — it has to be sold, not to make a profit but simply so the producer can survive. Under capitalist production there's an extra reason: selling the commodity is also how the surplus-value built into it gets realized. Either way, the product leaves the production process as a commodity, and so is neither a fixed nor a fluid element of that process.

Here Smith actually undercuts his own point. Finished products — whatever their physical shape, whatever their use-value or usefulness — are all, at this stage, commodity-capital: capital in a form that belongs to the circulation process. While they're in that form, they are no part of their owner's productive capital, if he has any. That doesn't stop them, the moment they're sold, from becoming — in the buyer's hands — elements of productive capital, fluid or fixed. So the very same things that at one moment appear on the market as commodity-capital, standing opposite productive capital, can, once they're taken off the market, go on to function — or fail to function — as fluid or fixed parts of someone's productive capital.

Take yarn, the cotton spinner's product. For the spinner, yarn is the commodity-form of his capital — commodity-capital. It can't go back into his own productive capital, not as material and not as a tool. But in the hands of the weaver who buys it, that same yarn is folded into the weaver's productive capital as one of its fluid parts. For the spinner, meanwhile, the yarn carries the value of part of both his fixed and his fluid capital (leaving surplus-value aside). The same goes for a machine. As the product of the machine-maker, it's the commodity-form of his capital, commodity-capital for him — and as long as it stays in that form, it is neither fluid nor fixed capital. Sold to a manufacturer who puts it to work, it becomes a fixed part of that manufacturer's productive capital. Even when, by its own physical form, a product can go straight back as a means of production into the very branch it came from — coal feeding back into coal-mining, say — the part of the coal meant for sale is still neither fluid nor fixed capital, but commodity-capital.

On the other hand, a product may, by its very nature, be completely unable to serve as any element of productive capital — not as material, not as a tool. Some food, say. Even so, it is still commodity-capital for whoever produced it, and it still carries the value of both fixed and fluid capital — of one or the other depending on whether the capital laid out to make it has to be replaced wholly or only in part, whether it has transferred all or only part of its value onto the product.

In Smith's item 3, raw material — raw stuff, semi-finished goods, auxiliary material — appears in two contradictory ways. On one hand it's treated not as something already built into anyone's productive capital, but simply as one particular kind of use-value among the goods making up society's product as a whole, alongside the food and finished goods listed under 2 and 4. On the other hand these same materials are also listed as already incorporated into productive capital, elements of it sitting in the producer's hand. The confusion shows up because they're pictured partly as sitting in the hands of producers — growers, manufacturers — and partly as sitting in the hands of merchants — silk and cloth dealers, timber merchants — where they are mere commodity-capital, not elements of productive capital at all.

So in this whole list of what makes up circulating capital, Smith has actually forgotten the distinction between fixed and fluid capital altogether — a distinction that only makes sense for productive capital in the first place. What he's really doing instead is setting commodity-capital and money-capital, the two forms belonging to the circulation process, against productive capital. But he does even that without realizing it.

Finally, it's striking that in listing the elements of circulating capital, Smith forgets labour-power entirely. This happens for two reasons.

We have just seen that, apart from money-capital, circulating capital is only another name for commodity-capital. But labour-power, as it circulates on the market, is not capital at all — not any form of commodity-capital. It isn't capital in any sense. The worker is no capitalist, though he does bring a commodity to market: his own skin. Only once labour-power is sold and taken up into the production process — that is, once it has stopped circulating as a commodity — does it become an element of productive capital: variable capital, the source of surplus-value, a fluid element of productive capital as far as the turnover of the capital-value laid out on it goes. Because Smith here confuses fluid capital with commodity-capital, he has no way to fit labour-power under his heading of circulating capital. So variable capital shows up instead in the shape of the commodities the worker buys with his wage — his means of subsistence. On this reading, the capital-value laid out in wages is supposed to belong to circulating capital. But what actually gets taken up into the production process is labour-power, the worker himself — not the food and other things by which the worker keeps himself alive. True, we have already seen that, viewed at the level of society as a whole, the worker's own reproduction through his individual consumption also belongs to the reproduction process of social capital. But that doesn't hold for the single, self-contained production process under consideration here. The "acquired and useful abilities" that Smith lists under fixed capital in fact belong, on the contrary, to fluid capital — once they are the abilities of a wage-labourer who has sold his labour along with those abilities.

Smith makes a serious mistake when he splits all social wealth into three boxes: the fund people consume directly, fixed capital, and circulating capital. Split correctly, wealth falls into only two things: the consumption fund, which is no part of the capital actually at work (even though bits of it can continually act as capital), and capital itself. So one part of wealth counts as capital, the other part doesn't — it is simply the consumption fund. And on Smith's scheme it looks as if every piece of capital simply has to be either fixed or fluid, the way every mammal simply has to be either male or female. But we have already seen that the fixed/fluid split only applies to the elements of productive capital. Alongside those there is a great deal of capital — commodity-capital and money-capital — sitting in a form that can be neither fixed nor fluid.

Except for the part of output that individual capitalist producers consume themselves, directly, as means of production, without buying or selling it, the whole mass of social production, on a capitalist basis, circulates on the market as commodity-capital. So it is clear that both the fixed and the fluid elements of productive capital, and everything that goes into the consumption fund, are drawn out of this commodity-capital. In other words: on the basis of capitalist production, means of production and means of consumption alike first turn up as commodity-capital, even though they are destined later to serve as means of consumption or means of production. In just the same way, labour-power itself turns up on the market as a commodity — though not as commodity-capital.

This produces a new confusion in Smith. He writes:

"Of these four parts (four, that is, because Smith takes capital in its circulation forms — commodity-capital and money-capital — and then subdivides the commodity-capital by its material contents), three — provisions, materials, and finished goods — are regularly taken out of it, either every year or over some longer or shorter stretch, and placed either in fixed capital or in the stock kept for immediate consumption. Every fixed capital originally comes from a circulating capital, and has to be constantly kept up by one. All useful machines and tools originally come from a circulating capital, which supplies the materials they are made of and pays for the upkeep of the workers who make them. They also need capital of the same kind to keep them constantly in repair."

With the same exception as before — the part of the product that producers consume themselves again, directly, as means of production — the general rule for capitalist production is this: all products come onto the market as commodities, and so circulate for the capitalist as the commodity-form of his capital, as commodity-capital. It makes no difference whether, by their physical form and use-value, these products must or can then function as elements of productive capital — as means of production, fixed or fluid — or whether they can only serve individual, not productive, consumption. All products are thrown onto the market as commodities; so all means of production and all means of consumption, everything used in productive or individual consumption, must be bought back out of the market as commodities. This truism is of course correct, and it holds equally for the fixed and the fluid elements of productive capital, for means of labour just as much as for materials of labour, in every form. (This already sets aside the elements of productive capital that exist by nature and are not products at all.) A machine, like cotton, is indeed bought on the market. But it does not follow from this — and it only seems to follow because Smith confuses capital belonging to the sphere of circulation with fluid, that is not-fixed, capital — that every piece of fixed capital originally comes from a fluid one. What is more, Smith contradicts himself here. By his own scheme, machines as commodities belong to the fourth item of circulating capital. So saying they "come from" circulating capital only means they functioned as commodity-capital before they functioned as machines — materially, they simply come from themselves, just as the cotton that is a fluid element of a spinner's capital comes from the cotton sold on the market. And when Smith goes on to derive fixed capital from fluid capital on the grounds that labour and raw material are needed to make machines, he forgets two things. First, means of labour — that is, fixed capital — are also needed to make machines. Second, fixed capital, machinery and the like, is also needed to make raw materials, since productive capital always includes means of labour, but does not always include material to work on. He himself says, right after this:

"Lands, mines, and fisheries all require both fixed and circulating capital to work them;"

So he admits that raw material, too, needs not only fluid but also fixed capital to produce it. "and" — here comes a fresh error — "their produce replaces, with a profit, not only those capitals, but all the other capitals in society as well."

This is completely wrong. What land, mines, and fisheries produce does supply the raw materials and auxiliary materials for every other branch of industry. But its value does not replace the value of all the other capitals in society; it only replaces its own capital-value, plus surplus-value. Here Smith's memory of the Physiocrats is showing through again.

Looked at socially, it is true that the part of commodity-capital made up of products that can only serve as means of labour must, sooner or later — assuming they were not simply produced for nothing and can find a buyer — actually go on to function as means of labour. That is: on the basis of capitalist production, once they have stopped being commodities, they must become real elements — as before they were only prospective ones — of the fixed part of social productive capital.

Here a distinction comes in that arises from the physical form of the product itself.

A spinning machine, say, has no use-value unless it is put to work spinning — that is, unless it functions as an element of production, or, from the capitalist's standpoint, as a fixed component of a productive capital. But a spinning machine can be moved. It can be exported from the country where it was made and sold abroad, directly or indirectly, whether for raw materials or for champagne. In the country where it was made, it will then only have functioned as commodity-capital — never, not even after being sold, as fixed capital.

Now take products that are fixed in place because they are built into the ground, and so can only be used where they stand — factory buildings, say, or railways, bridges, tunnels, docks, and land improvements. These cannot be physically exported, lock, stock, and barrel. They cannot be moved. Either they are useless, or, once sold, they must go on functioning as fixed capital in the country where they were built. For the capitalist who builds factories or improves land purely to sell them, these things are a form of his commodity-capital — on Smith's terms, a form of circulating capital. But looked at socially, unless they are to be useless, these things must eventually function as fixed capital, within a production process that is itself tied to their particular location. This certainly does not mean that anything immovable simply counts as fixed capital: as dwelling-houses, for instance, such things can belong to the consumption fund and not to social capital at all, even though they are still part of social wealth, of which capital is only one part. The producer of such things, to put it in Smith's own terms, makes a profit by selling them — so, circulating capital! The person who actually uses them, who buys them for good, can only use them by putting them to work in a production process — so, fixed capital!

Titles of ownership — in a railway, say — can change hands every day, and their owners can even sell them abroad and make a profit that way, so that the titles are exportable even though the railway itself is not. But none of this changes the fact that, in the country where they are located, these things must either lie idle or function as a fixed part of a productive capital. In the same way, one factory owner can make a profit by selling his factory to another — but that does not stop the factory going on functioning as fixed capital as before.

So: means of labour that are tied to a particular place, inseparable from the land, may function as commodity-capital for the person who produces them — his fixed capital consists instead of the means of labour he needs to build such buildings, railways, and so on — and yet still have to function, necessarily and predictably, as fixed capital within the country itself. But it does not follow, the other way round, that fixed capital must consist of immovable things. A ship or a locomotive only does its work by moving; yet for the person who uses them — though not for the person who produces them — they function as fixed capital. On the other side, things that really are fixed within the production process, that live and die inside it and never leave it again once they have entered it, count as fluid elements of productive capital. Take the coal burned to drive a machine in the production process, or the gas burned to light a factory building. These are fluid, not because they physically leave the production process along with the product and circulate as commodities — they do not — but because their whole value passes into the value of the commodity they help produce, and so has to be replaced in full out of the sale of that commodity.

In the passage from Smith just quoted, one more phrase is worth noting:

"A circulating capital which furnishes... the maintenance of the workmen who make them" — referring to machines and the like.

In the Physiocrats, the part of capital advanced as wages is correctly counted among the annual advances, as against the original advances. But for them, what figures as part of the productive capital the tenant-farmer employs is not labour-power itself, but the means of subsistence handed to the farm workers — "the maintenance of the workmen," as Smith puts it. This follows directly from their particular doctrine. For them, the portion of value that labour adds to the product — just like the portion added by raw material, tools, and the other material elements of constant capital — equals only the value of the means of subsistence paid to the workers, the amount needed to keep them going as labour-powers. Their doctrine itself makes it impossible for them to discover the difference between constant and variable capital. If it is labour that produces surplus-value, beyond reproducing its own price, then it produces surplus-value in industry just as much as in farming. But since their system holds that surplus-value arises only in the one branch, agriculture, that surplus cannot come from labour itself — it must come instead from nature's special contribution in that branch. And that is the only reason farm labour counts, for them, as productive labour, as against every other kind of labour.

Smith defines the workers' means of subsistence as circulating capital, as against fixed capital, for one reason to begin with: he confuses fluid capital — as against fixed capital — with the forms of capital that belong to the sphere of circulation, that is, with capital-of-circulation, a confusion that later writers have inherited from him uncritically ever since. So he confuses commodity-capital with the fluid component of productive capital. And once that confusion is made, it seems obvious that, since the social product takes the form of a commodity, the workers' means of subsistence — like those of non-workers, like materials, like the means of labour themselves — must all be supplied out of commodity-capital.

But a second thing is going on: a physiocratic idea slips into Smith here too, even though it contradicts the deeper — genuinely scientific — part of his own argument.

Capital that has been advanced is turned, in general, into productive capital — that is, it takes the shape of elements of production, which are themselves the product of earlier labour. Labour-power is one of these elements. Only in this shape can capital function within the production process at all. Now suppose that, instead of labour-power itself — which is what the variable part of capital has been converted into — one puts the worker's means of subsistence in its place. Then it is clear that, as far as forming value goes, these means of subsistence are no different from the other elements of productive capital: from the raw materials, or from the fodder given to draught animals, with which Smith, following the Physiocrats, does indeed lump them together, as quoted earlier. Means of subsistence cannot valorize their own value, or add any surplus-value to it. Their value, like that of the other elements of productive capital, can only turn up again in the value of the product — they cannot add more value to it than they themselves have. They differ from fixed capital made up of means of labour in the same way raw material or half-finished goods do: they are, at least for the capitalist who pays for them, wholly used up in the product they go into forming, so their value has to be replaced in full at once, whereas with fixed capital that happens only gradually, bit by bit. So the part of productive capital advanced in labour-power — or, on this substitution, in the worker's means of subsistence — now differs from the other material elements of productive capital only in its physical make-up, not in how it behaves in the labour process and the valorization process. It differs only in that it falls, together with part of the objective product-forming materials (what Smith calls "materials" in general), into the category of circulating capital, as against the other part of the objective product-forming materials, which falls into the category of fixed capital.

That the part of capital laid out in wages belongs to the fluid part of productive capital — sharing its fluidity, as against the fixed component of productive capital, with part of the objective product-forming materials, the raw materials and so on — has absolutely nothing to do with the part this variable capital plays, as against constant capital, in the valorization process. It is only about how this part of the advanced capital-value has to be replaced, renewed, reproduced, by means of circulation, out of the value of the product. The buying and re-buying of labour-power belongs to the circulation process. But it is only inside the production process that the value laid out in labour-power turns — not for the worker, but for the capitalist — from a fixed, constant magnitude into a variable one; and only through this does the advanced value become capital-value at all, become capital, become value that valorizes itself. Now, because Smith defines the fluid component of productive capital not as the value laid out in labour-power, but as the value laid out in the worker's means of subsistence, grasping the difference between variable and constant capital — and so grasping the capitalist production process at all — is made impossible. What should be the defining fact about this part of capital, that it is variable capital as against the constant capital laid out in the objective product-forming materials, gets buried under a different fact: that the part of capital laid out in labour-power belongs, as far as turnover goes, to the fluid part of productive capital. And the burying is completed once the worker's means of subsistence, rather than labour-power itself, are counted as the element of productive capital. Whether the value of labour-power is advanced in money or directly in means of subsistence makes no difference in itself — though of course, on the basis of capitalist production, the second case can only ever be an exception.

By fixing Smith's definition of circulating capital as the decisive fact about the capital-value laid out in labour-power — a physiocratic idea, kept without the Physiocrats' own premises — Smith succeeded, for everyone who came after him, in making it impossible to recognize the part of capital laid out in labour-power as variable. The deeper, correct account he gave elsewhere did not win out; this lapse of his did. Later writers went further still. They made it not merely the decisive fact about the capital laid out in labour-power that it counts as circulating rather than fixed capital; they made it an essential feature of circulating capital as such that it is laid out in means of subsistence for the workers. From this naturally followed the doctrine of a wages-fund made up of necessary means of subsistence, treated as a fixed quantity — one that, on the one hand, physically limits how much of the social product the workers can share in, and, on the other hand, must be spent in full on buying labour-power.