I had more in mind about this but I’m very tired and have forgotten some of my points.
In chapter 1 of volume 1 of Capital, Marx says that when comes to commodities, their value comes from “the quantity of the value-creating substance, the labour, contained in the article.” This quantity consists in the “labour time socially necessary (…) required to produce an article under the normal conditions of production, and with the average degree of skill and intensity prevalent at the time.” This means that what “determines the magnitude of the value of any article is the amount of labour socially necessary, or the labour time socially necessary for its production.”
The concept of socially necessary labor time is very important in Marx’s work. It’s relatively easy to misunderstand and Marx’s own presentation is conducive to mistakes. I think the term itself is misleading. At the bottom of this post I include several quotes from specific passages of Marx related to all this, which I plan to eventually return to. I don’t want to initially get bogged down in specific quotes so I’m going to begin with just laying out the basic points I want to make.
One straightforward misreading of Marx that I’ve run in into is to think about socially necessary labor time as time spent doing tasks that are necessary for society from a prescriptive perspective based on a vision of a good society other than capitalism. (That is, something like “socially necessary labor time is time spent doing tasks like caring for the sick or producing food” as distinct from tasks like creating implements of torture, with the implication being that after capitalism is replaced by a better society were will have a more rational allocation of and way of doing the former and none of the latter.) I hold such a perspective but that’s not what Marx means by “socially necessary.”
A second misreading is trickier. Marx talks about socially necessary labor time as the average labor time required to make a commodity. The second misreading is to take this at face value, defining socially necessary labor time as the average of the actual labor times actually required to produce commodities. This is a very plausible misreadingr because, well, Marx more or less says that socially necessary labor time is the average labor time required to make a commodity, and he says this is what determines the value of a commodity. In my view Marx doesn’t really mean what he appears to mean, a close reading shows this, and Marx’s terms are unfortunate and misleading. In my view it would be better to drop ‘socially necessary labor time’ as a term and refer to something like ‘labor time for which remuneration is required.’ I think part of the problem here comes from a pitfall of Marx proceeding by a complicated immanent critique of capitalism, which often involves taking up assumptions of capitalist ideology in order to show their internal problems. This can make it look like Marx actually believes things he doesn’t, things that are actually part of the perspective he is criticizing.
Marx’s interest in socially necessary labor time derives in part from his wanting to show that the value of commodities does not derive from the actual labor time spent on them. That is, it’s possible for actual labor time spent on actual objects to differ from socially necessary labor time, and actual labor time is not the source of the value of commodities. In chapter 1 of Capital v1 Marx gives an example from the textile industry, saying that “The introduction of power-looms into England probably reduced by one-half the labour required to weave a given quantity of yarn into cloth. The hand-loom weavers, as a matter of fact, continued to require the same time as before; but for all that, the product of one hour of their labour represented after the change only half an hour’s social labour, and consequently fell to one-half its former value.” Here we see that actual labor time expended surpasses socially necessary labor time. The role of socially necessary labor time is why even though “the value of a commodity is determined by the quantity of labour spent on it” it is not the case that “the more idle and unskilful the labourer, the more valuable would his commodity be, because more time would be required in its production.” Because the value of a commodity is not pegged to actual labor time spent on that actual commodity. (http://www.marxists.org/archive/marx/works/1867-c1/ch01.htm)
In fact, not only is it possible for actual labor time to differ from socially necessary labor tine, Marx suggests in ch13 of v1 that this difference happens all the time. “In every industry, each individual labourer, be he Peter or Paul, differs from the average labourer. (…) If one workman required considerably more time for the production of a commodity than is socially necessary, the duration of the necessary labour-time would, in his case, sensibly deviate from the labour-time socially necessary on an average; and consequently his labour would not count as average labour, nor his labour-power as average labour-power. It would either be not saleable at all, or only at something below the average value of labour-power.” (http://www.marxists.org/archive/marx/works/1867-c1/ch13.htm) So an individual worker may well perform below average, taking longer than it normally takes. When this happens, people lose their jobs or can’t get jobs or have to put in unpaid time to make up the time.
So far so good. Here’s the problem. Marx talks about socially necessary labor time as an average, so it’s easy to think that socially necessary labor time is an average of actual labor time actually spent. That’s not quite true though. In chapter 3 of v1 Marx writes about changes in markets and in the wants and needs that provide people with some reasons for buying commodities in markets, again with reference to textiles.
“To-day the product satisfies a social want. Tomorrow the article may, either altogether or partially, be superseded by some other appropriate product. Moreover, although our weaver’s labour may be a recognised branch of the social division of labour, yet that fact is by no means sufficient to guarantee the utility of his 20 yards of linen. (…) suppose that every piece of linen in the market contains no more labour-time than is socially necessary. In spite of this, all these pieces taken as a whole, may have had superfluous labour-time spent upon them. If the market cannot stomach the whole quantity at the normal price of 2 shillings a yard, this proves that too great a portion of the total labour of the community has been expended in the form of weaving. The effect is the same as if each individual weaver had expended more labour-time upon his particular product than is socially necessary. Here we may say, with the German proverb: caught together, hung together.”
That is: if no one wants to buy a product for what it really costs to make on average, then the labor time and resources spent on it are not socially necessary labor time. That is, the labor spent is socially superfluous, not socially necessary.
In my view, socially necessary labor time is a misleading phrase because even on a fairly good reading of Max the phrase sounds like “the time society needs in order to make something.” A better term might be something like “conventionally monetized labor time” meaning “labor time for which remuneration is typically required according to prevailing standards.”
My point is that “socially necessary labor time” is actually only the labor time that a capitalist is required to pay for, on average. If capitalists can get out of paying for certain periods of work time (say, workers forced to work through unpaid lunches, as discussed here by Disparaged CNA – http://disparagedcna.blogspot.com/ ) even though labor still happens, socially necessary labor time drops. If capitalists have to start paying for certain sorts of labor that they didn’t previously have to pay for, even though no new work is done, socially necessary labor time rises (say, for a long time worker worked through unpaid lunches then started getting paid for that time).
I heard an annoying radio program this morning while eating breakfast, about job hunting. Some consultant was talking about the need for some people to go get retrained so they can start new careers. The time and expense that people spend doing this is part of the total time and expense that goes into products, but much it doesn’t count as socially necessary labor time, because the time and expense are offloaded onto individual workers. If companies had to pay for these things, it would raise socially necessary labor time – again, without an increase in actual labor time spent.
Below are more bits from Marx to think about eventually. I’d like to write a piece comparing some of this with what Badiou calls a “count as one.”
Marx writes in volume III of capital that the value of commodities is determined “not by the labour-time necessary in the case of any individual producer for the production of a certain quantity of commodities, or of some individual commodity, but by the socially necessary labour-time; that is, by the labour-time, required for the production of the socially necessary total quantity of commodity varieties on the market under the existing average conditions of social production.” (http://www.marxists.org/archive/marx/works/1894-c3/ch38.htm)
(He writes a little later in that section that “The determination of value by the socially necessary labour-time asserts itself through the cheapening of commodities and the compulsion to produce commodities under the same favourable conditions.” http://www.marxists.org/archive/marx/works/1894-c3/ch38.htm I plan to come back to this, the assertion of the role of socially necessary labor time. In the original first chapter of Capital v1, in the first edition of the book, published in German in 1867, Marx writes that “the labour-time which is socially necessary for their production forcibly obtrudes itself as a regulating natural-law (…) The determination of the amount of value by the labour-time is consequently the mystery lurking under the apparent motions of the relative commodity-values.”
More succinctly, he writes in chapter 8 of v1 of Capital that commodities’ “value at any given time is measured by the labour socially necessary, i.e., by the labour necessary for their production under the then existing social conditions.”
Actual labor time can and often does differ from socially average labor time. Marx recognizes this in several places. For example he says in chapter 1 of Capital v1
“How, then, is the magnitude of this value to be measured? Plainly, by the quantity of the value-creating substance, the labour, contained in the article. The quantity of labour, however, is measured by its duration, and labour time in its turn finds its standard in weeks, days, and hours.
Some people might think that if the value of a commodity is determined by the quantity of labour spent on it, the more idle and unskilful the labourer, the more valuable would his commodity be, because more time would be required in its production. The labour, however, that forms the substance of value, is homogeneous human labour, expenditure of one uniform labour power. The total labour power of society, which is embodied in the sum total of the values of all commodities produced by that society, counts here as one homogeneous mass of human labour power, composed though it be of innumerable individual units. Each of these units is the same as any other, so far as it has the character of the average labour power of society, and takes effect as such; that is, so far as it requires for producing a commodity, no more time than is needed on an average, no more than is socially necessary. The labour time socially necessary is that required to produce an article under the normal conditions of production, and with the average degree of skill and intensity prevalent at the time. The introduction of power-looms into England probably reduced by one-half the labour required to weave a given quantity of yarn into cloth. The hand-loom weavers, as a matter of fact, continued to require the same time as before; but for all that, the product of one hour of their labour represented after the change only half an hour’s social labour, and consequently fell to one-half its former value.
We see then that that which determines the magnitude of the value of any article is the amount of labour socially necessary, or the labour time socially necessary for its production. Each individual commodity, in this connexion, is to be considered as an average sample of its class. Commodities, therefore, in which equal quantities of labour are embodied, or which can be produced in the same time, have the same value.”
Compare with the original first chapter of Capital v1, in the first edition of the book, published in German in 1867, that “It might seem that, if the value of a commodity is determined by the quantum of labour expended during its production, the more lazy- and incompetent a man the more valuable his commodity is, because he needs all the more labour-time for its completion. But only the socially necessary labour-time is labour-time required for the constitution of some particular use-value, with the available socially-normal conditions of production and the social average-level of competence and intensity of labour. After the introduction of the steam-driven loom in England, for example, perhaps half as much labour as before was sufficient to change a given quantum of yarn into cloth. The English hand-weaver needed in order to accomplish this change the same labour-time as before, to be sure, but the product of his individual labour-hour now represented only one half a social labour-hour, and sank accordingly to half its earlier value.
So it is only the quantum of socially necessary labour, or that labour-time which is socially necessary for the constitution of a use-value which determines the quantity of the value. The single commodity counts here in general as average sample of its own kind. Commodities in which equally large labour-quanta are contained, or which can be produced within the same labour-time for that reason have the same quantity of value. The value of a commodity is related to the value of every other commodity, as the labour-time necessary for the production of the one is related to the labour-time necessary for the production of the other. All commodities, as values, are only particular masses of coagulated labour-time.” (http://www.marxists.org/archive/marx/works/1867-c1/commodity.htm)
Marx writes in chapter 22 of v1 of Capital, “National Differences of Wages,” that “In every country there is a certain average intensity of labour below which the labour for the production of a commodity requires more than the socially necessary time, and therefore does not reckon as labour of normal quality. (…) The average intensity of labour changes from country to country; here it is greater, there less.” (http://www.marxists.org/archive/marx/works/1867-c1/ch22.htm It’s worth noting here that Marx largely thought of national markets as relatively independent despite their connections via the world market. This degree of independence has significantly eroded since Marx’s day.)
Marx writes in chapter 13 of v1, on cooperation, that
“The labour realised in value, is labour of an average social quality; is consequently the expenditure of average labour-power. Any average magnitude, however, is merely the average of a number of separate magnitudes all of one kind, but differing as to quantity. In every industry, each individual labourer, be he Peter or Paul, differs from the average labourer. These individual differences, or “errors” as they are called in mathematics, compensate one another, and vanish, whenever a certain minimum number of workmen are employed together. The celebrated sophist and sycophant, Edmund Burke, goes so far as to make the following assertion, based on his practical observations as a farmer; viz., that “in so small a platoon” as that of five farm labourers, all individual differences in the labour vanish, and that consequently any given five adult farm labourers taken together, will in the same time do as much work as any other five.  But, however that may be, it is clear, that the collective working-day of a large number of workmen simultaneously employed, divided by the number of these workmen, gives one day of average social labour. For example, let the working-day of each individual be 12 hours. Then the collective working-day of 12 men simultaneously employed, consists of 144 hours; and although the labour of each of the dozen men may deviate more or less from average social labour, each of them requiring a different time for the same operation, yet since the working-day of each is one-twelfth of the collective working-day of 144 hours, it possesses the qualities of an average social working-day. From the point of view, however, of the capitalist who employs these 12 men, the working-day is that of the whole dozen. Each individual man’s day is an aliquot part of the collective working-day, no matter whether the 12 men assist one another in their work, or whether the connexion between their operations consists merely in the fact, that the men are all working for the same capitalist. But if the 12 men are employed in six pairs, by as many different small masters, it will be quite a matter of chance, whether each of these masters produces the same value, and consequently whether he realises the general rate of surplus-value. Deviations would occur in individual cases. If one workman required considerably more time for the production of a commodity than is socially necessary, the duration of the necessary labour-time would, in his case, sensibly deviate from the labour-time socially necessary on an average; and consequently his labour would not count as average labour, nor his labour-power as average labour-power. It would either be not saleable at all, or only at something below the average value of labour-power. A fixed minimum of efficiency in all labour is therefore assumed, and we shall see, later on, that capitalist production provides the means of fixing this minimum. Nevertheless, this minimum deviates from the average, although on the other hand the capitalist has to pay the average value of labour-power. Of the six small masters, one would therefore squeeze out more than the average rate of surplus-value, another less. The inequalities would be compensated for the society at large, but not for the individual masters. Thus the laws of the production of value are only fully realised for the individual producer, when he produces as a capitalist, and employs a number of workmen together, whose labour, by its collective nature, is at once stamped as average social labour.” (http://www.marxists.org/archive/marx/works/1867-c1/ch13.htm)
So an individual work may well perform below average, taking longer than it normally takes. When this happens, people lose their jobs or can’t get jobs or have to put in unpaid time to make up the time.
Marx writes in a manuscript, the “Results of the Direct Production Process”
“the value of the means of production which enter into the process should not be greater than is necessary, hence that the commodities of which they consist should only contain, objectified, the labour time socially necessary for the purpose of production. This should be the case e.g. with the buildings, the machinery, etc., and it is the capitalist’s business to make sure of this when purchasing these means of production; to make sure that they have the appropriate average quality as use values needed for the formation of the product, whether as raw material or as machinery, etc., hence that they function averagely well, and do not oppose any unusual obstacles to labour, the living factor, e.g. through the quality of the raw material; and, also to be included here, the machinery, etc., employed should not pass on more than the average depreciation to the commodities, etc. All this is the capitalist’s affair.”
He adds later that “the worker must perform the socially normal quantity of purposeful labour within a given time, and the capitalist therefore forces him to ensure that his labour possesses at least the socially normal average degree of intensity. He will try to raise it as much as possible above this minimum, and extract from him over a given period as much labour as possible, for every [increase in the] intensity of labour over the average degree creates surplus value for him. He will also try to prolong the labour process as much as possible beyond the boundary of what has to be worked in order to replace the value of variable capital, of wages. If the intensity of the labour process is given, he will try to increase its duration as much as possible; if the duration is given, he will try to increase its intensity as much as possible. The capitalist compels the worker to give his labour the normal degree of intensity, and where possible a higher degree, and he compels him to prolong his labour process as much as possible beyond the period of time needed to replace his wages.”
“the quantitative calculation of the particular concrete labour as necessary average social labour (…) labour has to be considered in the dual form in which it is on the one hand represented as concrete labour (…) and on the other hand calculated as socially necessary labour (…) From the first point of view, everything depends on its particular use value, its specific character” while from the other point of view “its specific nature and the kind of thing it is are completely abstracted from when labour enters into calculations as a value-forming element (…) In this case it is undifferentiated, socially necessary, general labour, entirely indifferent towards any particular content, for which reason it receives an expression common to all commodities and only distinguishable in terms of quantity. This is its independent expression as money, the expression of the commodity as price. (…) From the one aspect the distinction between different kinds of concrete labour is expressed in the division of labour; from the other aspect in its monetary expression, which is undifferentiated.” (Marx repeatedly makes mention of his categories of use value and exchange value here in a way that I think is unclear and potentially misleading, and it may be a logical error on Marx’s part, this is all worth returning to but I’m not going to get into it now, see my old notes on sme of this; likewise worth returning to is the way Marx here lays out the relationships between the valorization process and the labor process.)
Chapter 21, on piece wages
“Let us suppose that, as the result of experience, a labourer who works with the average amount of intensity and skill, who, therefore, gives in fact only the time socially necessary to the production of an article, supplies in 12 hours 24 pieces, either distinct products or measurable parts of a continuous whole.”
“Let us now consider a little more closely the characteristic peculiarities of piece wages.
The quality of the labour is here controlled by the work itself, which must be of average perfection if the piece-price is to be paid in full. piece wages become, from this point of view, the most fruitful source of reductions of wages and capitalistic cheating.
They furnish to the capitalist an exact measure for the intensity of labour. Only the working-time which is embodied in a quantum of commodities determined beforehand, and experimentally fixed, counts as socially necessary working-time, and is paid as such. In the larger workshops of the London tailors, therefore, a certain piece of work, a waistcoat, e.g., is called an hour, or half an hour, the hour at 6d. By practice it is known how much is the average product of one hour. With new fashions, repairs, &c., a contest arises between master and labourer as to whether a particular piece of work is one hour, and so on, until here also experience decides. Similarly in the London furniture workshops, &c. If the labourer does not possess the average capacity, if he cannot in consequence supply a certain minimum of work per day, he is dismissed. 
Since the quality and intensity of the work are here controlled by the form of wage itself, superintendence of labour becomes in great part superfluous. piece wages therefore lay the foundation of the modern “domestic labour,” described above, as well as of a hierarchically organized system of exploitation and oppression. The latter has two fundamental forms. On the one hand, piece wages facilitate the interposition of parasites between the capitalist and the wage-labourer, the “sub-letting of labour.” The gain of these middlemen comes entirely from the difference between the labour-price which the capitalist pays, and the part of that price which they actually allow to reach the labourer.  In England this system is characteristically called the “sweating system.” On the other hand, piece-wage allows the capitalist to make a contract for so much per piece with the head labourer — in manufactures with the chief of some group, in mines with the extractor of the coal, in the factory with the actual machine-worker — at a price for which the head labourer himself undertakes the enlisting and payment of his assistant work people. The exploitation of the labourer by capital is here effected through the exploitation of the labourer by the labourer. 
Given piece-wage, it is naturally the personal interest of the labourer to strain his labour-power as intensely as possible; this enables the capitalist to raise more easily the normal degree of intensity of labour.  It is moreover now the personal interest of the labourer to lengthen the working-day, since with it his daily or weekly wages rise.  This gradually brings on a reaction like that already described in time-wages, without reckoning that the prolongation of the working-day, even if the piece wage remains constant, includes of necessity a fall in the price of the labour.
In time-wages, with few exceptions, the same wage holds for the same kind of work, whilst in piece wages, though the price of the working time is measured by a certain quantity of product, the day’s or week’s wage will vary with the individual differences of the labourers, of whom one supplies in a given time the minimum of product only, another the average, a third more than the average. With regard to actual receipts there is, therefore, great variety according to the different skill, strength, energy, staying-power, &c., of the individual labourers.  Of course this does not alter the general relations between capital and wage-labour. First, the individual differences balance one another in the workshop as a whole, which thus supplies in a given working-time the average product, and the total wages paid will be the average wages of that particular branch of industry. Second, the proportion between wages and surplus-value remains unaltered, since the mass of surplus labour supplied by each particular labourer corresponds with the wage received by him. But the wider scope that piece-wage gives to individuality tends to develop on the one hand that individuality, and with it the sense of liberty, independence, and self-control of the labourers, and on the other, their competition one with another. Piece-work has, therefore, a tendency, while raising individual wages above the average, to lower this average itself.”
In “the Value Form,” an appendix published with the first edition of v1 of Capital Marx wrote that
“The fact that the products of labour – such useful things as coat, linen, wheat, iron, etc. – are values, definite magnitudes of value and in general commodities, are properties which naturally pertain to them only in our practical interrelations (in unsrem Verkehr) and not by nature like, for example, the property of being heavy or being warming or nourishing. But within our practical interrelations, these things relate to one another as commodities. They are values, they are measurable as magnitudes of value, and their common property of being values puts them into a value-relation to one another. Now the fact that, for example, ‘20 yards of linen = 1 coat’ or ‘20 yards of linen are worth 1 coat’ only expresses the fact that:
1. the different types of labour necessary for the production of these things count equally (gleichgelten) as human labour;
2. the fact that the quantity of labour expended in their production is measured according to definite social laws;
3. that tailors and weavers enter into a definite social relation of production.
It is a definite social relation of the producers in which they equate (gleichsetzen) their different types of labour as human labour. It is not less a definite social relation of producers, in which they measure the magnitude of their labours by the duration of expenditure of human labour-power. But within our practical interrelations these social characters of their own labours appear to them as social properties pertaining to them by nature, as objective determinations (gegenständliche Bestimmungen) of the products of labour themselves, the equality of human labours as a value-property of the products of labour, the measure of the labour by the socially necessary labour-time as the magnitude of value of the products of labour, and finally the social relations of the producers through their labours appear as a value-relation or social relation of these things, the products of labour.” (http://www.marxists.org/archive/marx/works/1867-c1/appendix.htm)
So, socially necessary labor time here
Marx sometimes talks about socially necessary
chapter 3 of Capital v1
“The leap taken by value from the body of the commodity, into the body of the gold, is, as I have elsewhere called it, the salto mortale of the commodity. If it falls short, then, although the commodity itself is not harmed, its owner decidedly is. The social division of labour causes his labour to be as one-sided as his wants are many-sided. This is precisely the reason why the product of his labour serves him solely as exchange-value. But it cannot acquire the properties of a socially recognised universal equivalent, except by being converted into money. That money, however, is in some one else’s pocket. In order to entice the money out of that pocket, our friend’s commodity must, above all things, be a use-value to the owner of the money. For this, it is necessary that the labour expended upon it, be of a kind that is socially useful, of a kind that constitutes a branch of the social division of labour. But division of labour is a system of production which has grown up spontaneously and continues to grow behind the backs of the producers. The commodity to be exchanged may possibly be the product of some new kind of labour, that pretends to satisfy newly arisen requirements, or even to give rise itself to new requirements. A particular operation, though yesterday, perhaps, forming one out of the many operations conducted by one producer in creating a given commodity, may to-day separate itself from this connexion, may establish itself as an independent branch of labour and send its incomplete product to market as an independent commodity. The circumstances may or may not be ripe for such a separation. To-day the product satisfies a social want. Tomorrow the article may, either altogether or partially, be superseded by some other appropriate product. Moreover, although our weaver’s labour may be a recognised branch of the social division of labour, yet that fact is by no means sufficient to guarantee the utility of his 20 yards of linen. If the community’s want of linen, and such a want has a limit like every other want, should already be saturated by the products of rival weavers. our friend’s product is superfluous, redundant, and consequently useless. Although people do not look a gift-horse in the mouth, our friend does not frequent the market for the purpose of making presents. But suppose his product turn out a real use-value, and thereby attracts money? The question arises, how much will it attract? No doubt the answer is already anticipated in the price of the article, in the exponent of the magnitude of its value. We leave out of consideration here any accidental miscalculation of value by our friend, a mistake that is soon rectified in the market. We suppose him to have spent on his product only that amount of labour-time that is on an average socially necessary. The price then, is merely the moneyname of the quantity of social labour realised in his commodity. But without the leave, and behind the back, of our weaver, the old-fashioned mode of weaving undergoes a change. The labour-time that yesterday was without doubt socially necessary to the production of a yard of linen, ceases to be so to-day, a fact which the owner of the money is only too eager to prove from the prices quoted by our friend’s competitors. Unluckily for him, weavers are not few and far between. Lastly, suppose that every piece of linen in the market contains no more labour-time than is socially necessary. In spite of this, all these pieces taken as a whole, may have had superfluous labour-time spent upon them. If the market cannot stomach the whole quantity at the normal price of 2 shillings a yard, this proves that too great a portion of the total labour of the community has been expended in the form of weaving. The effect is the same as if each individual weaver had expended more labour-time upon his particular product than is socially necessary. Here we may say, with the German proverb: caught together, hung together.”
“Relative Surplus Value (…) As we saw in our analysis of the commodity, the productivity of labour does not increase the value of the product or the commodity in which the labour manifests itself. If we presuppose that the labour time contained in the commodities is, under the given conditions, necessary labour time, socially necessary labour time — and this is always the presupposition we start from once the value of a commodity is reduced to the labour time contained in it — what takes place is rather the following: The value of the product of labour is in an inverse ratio to the productivity of labour. This is in fact an identical proposition. It means nothing more than this: If labour becomes more productive, it can represent a greater quantity of the same use values in the same period, it can embody itself in a greater amount of use values of the same kind. Accordingly, an aliquot part of these use values, e.g. a yard of linen, contains less labour time than previously, has therefore less exchange value and indeed the exchange value of the yard of linen has fallen in the same proportion as the productivity of the labour of weaving has grown. Inversely, if more labour time than previously were required to produce a yard of linen (let us say, because more labour time was required to produce a pound of flax), the yard of linen would now contain more labour time, hence would have a higher exchange value. Its exchange value would have increased in the same proportion as the labour required to produce it had become less productive.”
Capital v1, ch12, The Concept of Relative Surplus Value
“The value of labour-power, i.e., the labour-time requisite to produce labour-power, determines the labour-time necessary for the reproduction of that value. If one working-hour be embodied in sixpence, and the value of a day’s labour-power be five shillings, the labourer must work 10 hours a day, in order to replace the value paid by capital for his labour-power, or to produce an equivalent for the value of his daily necessary means of subsistence. Given the value of these means of subsistence, the value of his labour-power is given;  and given the value of his labour-power, the duration of his necessary labour-time is given. The duration of the surplus-labour, however, is arrived at, by subtracting the necessary labour-time from the total working day. Ten hours subtracted from twelve, leave two, and it is not easy to see, how, under the given conditions, the surplus-labour can possibly be prolonged beyond two hours. No doubt, the capitalist can, instead of five shillings, pay the labourer four shillings and sixpence or even less. For the reproduction of this value of four shillings and sixpence, nine hours’ labour-time would suffice; and consequently three hours of surplus-labour, instead of two, would accrue to the capitalist, and the surplus-value would rise from one shilling to eighteen-pence. This result, however, would be obtained only by lowering the wages of the labourer below the value of his labour-power. With the four shillings and sixpence which he produces in nine hours, he commands one-tenth less of the necessaries of life than before, and consequently the proper reproduction of his labour-power is crippled. The surplus-labour would in this case be prolonged only by an overstepping of its normal limits; its domain would be extended only by a usurpation of part of the domain of necessary labour-time. Despite the important part which this method plays in actual practice, we are excluded from considering it in this place, by our assumption, that all commodities, including labour-power, are bought and sold at their full value. Granted this, it follows that the labour-time necessary for the production of labour-power, or for the reproduction of its value, cannot be lessened by a fall in the labourer’s wages below the value of his labour-power, but only by a fall in this value itself. Given the length of the working day, the prolongation of the surplus-labour must of necessity originate in the curtailment of the necessary labour-time; the latter cannot arise from the former. In the example we have taken, it is necessary that the value of labour-power should actually fall by one-tenth, in order that the necessary labour-time may be diminished by one-tenth, i.e., from ten hours to nine, and in order that the surplus labour may consequently be prolonged from two hours to three.
In order to effect a fall in the value of labour-power, the increase in the productiveness of labour must seize upon those branches of industry whose products determine the value of labour-power, and consequently either belong to the class of customary means of subsistence, or are capable of supplying the place of those means. But the value of a commodity is determined, not only by the quantity of labour which the labourer directly bestows upon that commodity, but also by the labour contained in the means of production. For instance, the value of a pair of boots depends not only on the cobbler’s labour, but also on the value of the leather, wax, thread, &c. Hence, a fall in the value of labour-power is also brought about by an increase in the productiveness of labour, and by a corresponding cheapening of commodities in those industries which supply the instruments of labour and the raw material, that form the material elements of the constant capital required for producing the necessaries of life. But an increase in the productiveness of labour in those branches of industry which supply neither the necessaries of life, nor the means of production for such necessaries, leaves the value of labour-power undisturbed.
The cheapened commodity, of course, causes only a pro tanto fall in the value of labour-power, a fall proportional to the extent of that commodity’s employment in the reproduction of labour-power. Shirts, for instance, are a necessary means of subsistence, but are only one out of many. The totality of the necessaries of life consists, however, of various commodities, each the product of a distinct industry; and the value of each of those commodities enters as a component part into the value of labour-power. This latter value decreases with the decrease of the labour-time necessary for its reproduction; the total decrease being the sum of all the different curtailments of labour-time effected in those various and distinct industries.” (http://www.marxists.org/archive/marx/works/1867-c1/ch12.htm)