In the last few years a new object of knowledge has appeared for sociology and economics: housework. Of course, it had not really been ignored up until then, but it was only tackled in a descriptive and quantitative way. Theoretically limited as they may be, we in France owe to the first time-budget studies (Stoetzel 1948, Girard 1958 and 1959) a recognition of the number of daily hours of housework performed by married women. We owe to the new feminists, however, the posing, for the first time in history, of the question of housework as a theoretical problem.1 Subsequently ‘scientists’ (and among them some of the most brilliant, such as J. K. Galbraith), although they did not produce a theory of housework, have at least not hesitated to recognize the importance of the problem. After a decent lapse of time, they have built into their accounts concepts forged for the most part outside the academic world.
The literature on housework grows each year – to the point where ‘schools’ are beginning to form. However, a consensus does still exist among the authors, mostly women, as to the major economic characteristics of housework. It is agreed, on the one hand, that housework is work (and that this is indeed the reason why it is being considered) and on the other that it is free/unpaid (which is why recognizing it as work was not automatic, but rather constituted a great step forward and a scientific discovery). But beyond these two points of agreement, the divergences begin. There is a debate, for example, about the ‘productivity’ of housework. Some put the stress on the aspect of ‘work’, in the sense of ‘tasks’ involved in housework; others on the ‘unpaid’ aspect, on the mode of production. Studies are sometimes focused on the use made of housework by – or the use for – capitalism, and sometimes on the meaning of this work for the direct producers: for women-wives.2 All these points are obviously connected, but in spite of the interest generated by the question, and the number of articles or books which have already been devoted to it, one point remains relatively hazy, albeit fundamental. And that is the very definition of housework, of the object of study.
The definition of housework remains hazy not because of an unresolved debate, but, on the contrary, because a tacit consensus exists as to its content. Housework is never formally defined, but the examples cited show authors adopting uncritically a ‘commonsense’ definition of housework (which is also the one used in time-budgets surveys). This takes housework to be the work done within the house by the wife: cooking, washing, ironing, sewing, shopping, cleaning and servicing (i.e. housewifery in the restricted sense of ‘doing the housework’), and care of children. But is this empirical definition adequate? Is it accurate? And is it compatible with the economic attributes of housework identified above?
The empirical content given to housework is inevitably affected by the theoretical interpretation given to the subject. It seems to me that a great many of the dead ends in which the debate on housework currently finds itself could be avoided if there was agreement on its principal characteristics. This implies that there should be a formal (and not an empirical) definition of the object of study, i.e. of the content of housework; and, conversely, that important features of housework will be found only when it is studied empirically from the point of view of its structural and economic characteristics.
Such a procedure might be taxed with being arbitrary, or the argument could be dismissed as circular, in that the economic (structural) characteristics attributed to housework at the start might be the ones one wanted to retain at the end as essential traits. The difficulty can best be avoided by taking as the point of departure only those characteristics of housework on which there is universal agreement: that is, its character as work, and its character as unpaid work.
This done, we can pose two questions. Does the commonsense content, the empirical definition of housework, cover all its economic characteristics? And, conversely, do these characteristics cover all its content? Putting this in other words:
1is it only housework (in the commonsense usage) which is unpaid)
2and is all household activity unpaid work?
I shall try to show, on the one hand, that the defining economic characteristics of housework cover a great deal more than the classical contents and apply also to ‘work’ which is supposedly not ‘of the house’; and on the other, that these economic characteristics do not apply to all household activity in the technical sense.
To this end, I shall make a detour via a subject which seems at first sight foreign to housework: the production which national accounting lists under the heading of ‘for self-consumption by households’. I shall restrict myself to describing production for self-consumption in agriculture (i.e. that of small farm households) because this is both quantitatively the most important in much of Europe and also the best researched.
Production for self-consumption on farms
The fact that agricultural households/enterprises produce for their own consumption is well known and could indeed constitute a definition of contemporary small farms as a social stratum (‘peasants’). Many authors consider that ‘self-sufficiency’ is the trait which most differentiates small farmers from the ‘rest of society’, whose members cannot directly consume what they produce. Further, this self-sufficiency makes this sector relatively independent of the exchange which necessarily unites the members of a society with a marked division of labour.
I believe this self-sufficiency, which is often presented as a major characteristic of the classical and atemporal peasant, to have been largely exaggerated. None the less, the quantity of agricultural produce which is directly consumed by its producers is substantial. According to Milhau and Montagne (1968): ‘French agriculture retains at least 25% of its gross produce for its own needs, and the proportion is much higher still in other countries.’
This high percentage should be treated with reservations, however, because of the ambiguity introduced by the term ‘gross product’. This makes it appear as if some intermediary production may be included; and even if we accept that the percentage consists solely of what the household appropriates from its final product, there is a further difficulty. The figure gives us an average rate, and in the actual situation of French agriculture, where there are enormous variations, an average, of whatever it may be, tells us next to nothing.
If we take the classical division of agriculture into three sectors:
1(so-called) subsistence agriculture;
2(so-called) artisanal agriculture (holdings which are traditional in production structure, but mechanized and of medium size); and
3(so-called) capitalist or industrial agriculture,
it is clear that the extent to which products are commercialized, and hence the extent to which self-consumption is possible, must vary enormously between sectors. In the ‘agribusiness’ of Beauce or Soissonnais, specialization in cereals is combined with the whole product being commercialized. This holds true generally for purely cereal holdings: they produce nothing except for the market. Households on such farms, like urban households, buy all the products which they consume in the market.
But it is not the scale of the holdings, be it in area or in total amount of business, which provides the appropriate criterion for dividing holdings which produce for their own consumption from those which do not do so. The extent of specialization is more to the point. Holdings specializing in fruit growing, market gardening, factory farming of pigs or chickens, or vineyards are all in the same position as the big cereal farmers.
At first sight, it seems that monoculture by its nature limits the possibilities for self-consumption. You cannot live exclusively on grapes, or pork or peaches or whatever. But this explanation is less ‘obvious’ than it appears, for in fact specialization of holdings itself results from an economic choice, from a decision to adapt to, or to orientate exclusively towards, the market.
Non-specialized holdings, on the other hand, those subsistence or artisan farms which make up the majority of France, are said to be ‘general’ farms. This term tends to confuse because it leads to – or derives from – two received ideas which do not correspond to reality. One is that these holdings have several types of market speculation: that they offer various types of produce on the market, whereas specialized holdings or monoculture offer only one. Second, that production for self-consumption, sometimes seen as the principal aim of traditional agricultural production, is also the consequence of ‘general’ production: it results from the production of diverse goods for the market.
In reality, general farm holdings usually have only one market speculation: stock rearing, either for milk or for meat. The growing of plants-which is what makes it mixed or ‘general’ farming – is an intermediary production. The plants do not reach the market. They are consumed in the process of producing market animals. Thus, so far as the market is concerned, to call such farms ‘general’ is misleading. There is however good and plentiful general farming on these holdings, but contrary to what Milhau and Montagne assert, this diverse production is not (or is no longer) motivated by a concern to reduce risks by offering several products on the market – to secure oneself against fluctuations in prices. It is for self-consumption.
It can thus be seen that it is not specialization as such which prohibits self-consumption, for general farm holdings are, from the point of view of the market, in the same situation as those which are specialized. Both offer only one final product, and the specialized holdings could, just as well as those in general farming, have diversified cultivation for their own consumption. In so far as they do not do so, it is because of their economic attitude. They calculate in terms of profitability, and this calculation leads them to choose commercialization at the expense of self-consumption. Similarly, the self-consumption typical of general farms does not result from a diversification given by what is produced — a diversity which exists prior to a choice to produce for self-consumption. On the contrary, a choice is made in favour of production for consumption at the expense of production for the market.
Production for self-consumption in agriculture is taken into consideration when national accounts are constructed. This production, although non-market in essence, is included in the gross national product and the gross national revenue. However, since, in the strict sense, it has no value – not having been the object of exchange – to be included it has to be attributed one. How is this value determined?
The very fact of including a pig, produced and eaten by a household, in the GNR, follows from two presumptions:
1that the household, if it had not eaten the pig, would have been able to sell it; and
2that the household, if it had not produced the pig, whould have had to buy one.
These two propositions are equally legitimate. However, from an accounting point of view, they are not equivalent. According to which you choose, the pig has a different value. National accounting outside France chooses the first proposition and calculates the loss of earnings to agriculture. It invoices the pig at its selling price. French national accounting has favoured the latter proposition since 1963 and calculates the saving in spending for the household. It invoices the pig at its retail price. The latter is greater than the former.
According to the experts, the French solution presents some drawbacks because
it is the price of production which represents the cost of the consumption in question … commodities consumed on the farm do not carry the expenses of transport, the profit margins of intermediaries, and the taxes which encumber the retail price … evaluation at retail price … consequently includes … the value of services which have not in fact been provided
(Marczewski 1967).
Despite these drawbacks, the French national account changed tack because, although this method makes a supplementary production appear (that of transforming agricultural produce into the products of the agricultural and foodstuffs industries) which, according to Marczewski, hasn’t been effected, in other respects the augmentation of the gross internal product of the holdings (and thus the gross national product) is counterbalanced by the attribution of a debit of equal amount to the disposable income of households. Thus, according to our expert ‘the balance of accounts is not affected.’ In addition Marczewski admits that, if it is ‘certainly not true’ that agricultural households and holdings fulfil some commercialized functions (whose value is included in the retail price), it is ‘partly true that they fulfil for themselves the functions of transforming agricultural products’.
This admission is welcome, because if Marczewski had held to his first position, according to which agriculturalists do not carry out any transformation of agricultural products, we would have to suppose that they consume the same raw product as they sell, and thus that they devour pigs running on their four trotters.
Some of the transformations which are necessary to make agricultural products consumable are thus accounted, but only some. The products are valued at their retail price. The pig is (happily) killed, skinned and cut into joints, so French farmers cannot be taxed with eating live animals. But the national account stops there: at a point when the pig is still far from being consumable. The last necessary operations – preparation, cooking and serving the pig – are not taken into account. Once again agricultural households must be suspected of uncommon feeding habits – for example, a taste for eating raw meat directly off the slab – unless of course we want to suspect those who construct national accounts of being arbitrary.
But the arbitrariness of the national account is not surprising. For what are the transformations which continue to be invisible, what are the things which are not thought of as being done? Those which correspond to some of the operations covered by the term ‘housework’.
Housework as production
It is often argued that housework is unpaid (free) because it is not productive, and that it is not productive because it ‘does not enter into the system of values’, i.e. it does not pass through the market. Not only is this a poor explanation, it is above all a curious definition of productivity, because as we have just seen, some non-market products, consumed by their producers, are accounted and treated as productive. The absence of passage through the market (or not having been exchanged) is thus not the reason for the status of domestic work. The previous discussion has shown that housework is productive if we adopt the terms of national accounting, which are legitimate in that they define as productive all work which increases wealth. Accordingly, housework should be considered as productive along with other production accounted under the title of ‘for household self-consumption’. The process of production for one’s own consumption is a whole: either all the operations which lead to the final consumption are productive, or none of them are.
So we could ask why the French national account in particular makes such an arbitrary division within this process? If the prevalent principle in evaluating production for self-consumption made explicit reference to the lack of monetary gain – hence to the fact of goods not reaching the market – then the exclusion of the transformations to make food consumable – but of all the transformations to make it consumable – would be comprehensible (that is to say, if the principle stated that the value accounted was the sale price of the pig ‘on the hoof’, because only the potentiality of sale makes it a production). But such is not the case. On the one hand certain transformations are accounted, and, on the other, reference to potential marketability would in any case not justify the exclusion of housework.
We have seen that the only part of self-consumption which is accounted for is the value of goods in the state in which households usually purchase them on the retail market (i.e. before the final housework transformations). These final transformations, which are not counted in agricultural households, are not counted in any other households either. Only agricultural households raise pigs, but chops are cooked in all households. Consequently all households, and not just agricultural households, produce for their own consumption. Hence the sum of services and values which the national account ignores is substantial. To understand how and why this institution establishes this arbitrary cut, and to say why reference to marketability also does not justify excluding housework, we must make a new detour. At the end of the detour we will have the answers to these two questions and, moreover, will be able to reply to the first question posed: is housework the only unpaid work?
The difference between occupational work and household work on farms
On agricultural smallholdings, women take part in all production, as much in that destined for the market as in that destined for the accounted aspects of self-consumption. This is true for men too. But women also do some additional unaccounted work which men do not do, whose products are consumed by the household. This work is designated ‘housework’.
Time-budget studies of the wives of small farmers distinguish their ‘housework’ from their ‘occupational work’. Most of the authors of such work say how difficult it is to distinguish the two, and they underline the concrete overlap of tasks. Becouarn (n.d.) and Bastide and Girard (1959) deem that the wives of small farmers do an average of four hours of farm work a day; Allauzen (1967) raises this figure to five hours.
But what criteria distinguish housework from occupational work? Apparently the distinction between ‘holding’ and ‘house’. This itself recovers, or rather applies to the agricultural holding, an opposition between business and household; which itself hides another opposition, between production and consumption, without which the first has no meaning. But this cannot be used to single out housework alone, precisely because on farms, part of what is produced by the ‘business’ is directly consumed by the household. The distinction between exchange-value and use-value (between production for the market and production for consumption) which may be being referred to implicitly, is also not relevant either. If it were used, it would not separate ‘housework’ from ‘occupational work’, but rather ‘housework-and-some-occupational work’ from ‘the-rest-of-the-occupational-work’, because ‘occupational work’ covers all the holdings’ production, and a substantial part of this is self-consumed (i.e. constitutes use-values).
How, then, is the holding itself defined? It is neither a place (because many occupational activities take place inside the house) nor a business (because it does not produce exclusively for the market). In which case, how is occupational work to be defined, since ‘occupational work’ refers back to (is the work associated with) the ‘holding’ and no definition of holding exists?
In the absence of definitions, we have to look and see what the empirical objects are with which these terms coincide. We then see that, in actuality, the term ‘occupational’ is applied to activities which have as their object accounted production; and as we saw in the last section, all the household and holding production is accounted – except for the part called ‘housework’. Housework is itself implicitly defined in opposition to, or rather as what is left when, occupational work is subtracted. It is what is not occupational work.
We thus arrive at a reasoning of remarkable circularity, and certainly at a theoretical dead end. There is no formal, economic, definition of occupational work or the holding. (Since these two terms refer back to one another, if one lacks a definition so must the other.) Likewise occupational work, having no economic definition, has no features which distinguish it formally from ‘housework’, which at the same stroke has no definition either. These two terms certainly relate to one another, but, in the absence of definition, not as two concepts but as two empirical objects. They are opposed, but in some way other than in the economic logic from which they are deemed to arise.
We therefore have to look once again to the empirical world to find our way out of the impasse; or rather, to see in what domain these two types of ‘work’ are defined. We can then establish that rural researchers call ‘housework’ certain tasks which are in fact in no way distinguished on the economic level from the other tasks carried out by the wives of small farmers. The only thing which distinguishes these tasks from the others carried out by the same women, is what does not distinguish them from tasks carried out by women in non-agricultural households. What distinguishes housework on farms is that it is not specifically agricultural. It is done elsewhere as well. In sum, researchers call ‘housework’ work which is not specific to agricultural production for self-consumption, i.e. which is common to all production for self-consumption.
‘Occupational work’ on farms in turn is thus nothing – it cannot be anything – other than what is left once ‘housework’ has been subtracted. The definition of ‘occupational work’ depends on housework, defined as a package of tasks; and the only sociological characteristic of this package is that it is work for self-consumption which is done by non-agricultural households as well as by agricultural households.
To sum up, the definition of accounted self-consumption on farms seems to be: all agricultural production for self-consumption less the production for self-consumption common to all households, rural and urban. This definition might seem to show that urban criteria are applied improperly to agricultural households: that production is counted when it is done outside the household, while all that is done inside the household is counted as non-productive, even though on farms it is productive.
But the problem goes far beyond making categories of industrial accounting adequate to agriculture. For in fact, even though the urban household is held to be non-productive, in reality it is productive. It could be argued that not only occupational work in agriculture, but also occupational work as such, is empirically defined as what remains once housework is eliminated. The only justification in fact for declaring the work done in non-agricultural households non-productive would be if all work for self-consumption were declared non-productive, including therefore all the production for self-consumption in agriculture which is officially accounted.
Housework as unpaid work
The reason why housework is not considered to be productive and why it is not accounted is because it is done, within the confines of the home, for free: because it is not paid, or exchanged in the general fashion. And this is not because of the nature of the services which make it up, because one can find any and all of them on the market; nor is it because of the nature of the people who do it, because the same woman who cooks a chop unpaid in her home is paid when she does it in another household (see The main enemy, Chapter 4). It is because of the particular nature of the contract which ties the female worker – the wife – to the household of her ‘master’.
But, you may say, the non-remuneration of work is not specific to housework. It is true of all the work people do for their own consumption. No one gets paid for work they do for themselves and their families.
Some authors, however, including Dalla Costa (1972), have argued that housework – which they extend to cover all the work women do to reproduce themselves and their families (which I would call production for self-consumption) – is unpaid, in the sense of it being productive and necessary work which should properly be paid. Dalla Costa, and the group which formed around her work, have concluded that all housework, including the housework done by single women for themselves as much as the work of wives, should be paid for by the state. They have demanded wages for housework.
As I see it, they make an intellectual slip, due to the same lack of theoretical rigour as is shown by other authors who define housework as a series of tasks. They call work, and hence unpaid work, all the services you do for yourself (including the work single men do for themselves). They do not define the social conditions of production which would qualify an activity’s being treated as unpaid work. I believe only services furnished for others should be so designated.
To take again the example of self-consumption in agriculture, let us suppose that a small farmer (who could be a man or a woman) makes his or her own bread. This activity is considered productive even by the national account. If the farmer does not eat the bread, he or she can sell it; if they had not baked it, they would have had to buy it. Should this accounted activity also be paid for? (Is it ‘unpaid’?) If farmers bake bread, should they be paid for the baking? Of course not. It is held, and rightly, that they remunerate themselves: that if they did not bake their own bread, the baking should certainly be paid for, but by them, to the baker. If they do bake, they collect the price of baking in being thrifty. This is why it is legitimate to add this value to the national product. A loaf has effectively been baked. That the producer has consumed it forthwith (or almost) in no way detracts from the fact that the bread has been well and truly added to the overall total of individual and national wealth. That the bread does not pass through the market means the values of transport and commercial services are not added – but it changes nothing about the production of the bread itself. It is legitimate to add the value of home-baked bread to the income of small farmers because they profit from this added wealth. If they had sold it instead of eating it, they would equally have found themselves richer by the value of a loaf.
To consider that, not having been paid (in money) for work, an individual has performed free work, is a gross error. He or she has been remunerated, whether in saving the baker’s service or in consuming an extra loaf. In either case, they have been remunerated. It may be well or badly, in terms of monetary calculation (when the saving made is compared to the time spent, for example), but this matters little, for the individual has chosen his or her own remuneration.
It is thus not contradictory to say, both, that work is productive, and that it should not be remunerated by an external source, in so far as productive work adds an element to the riches of its producers, and thereby carried its own remuneration. To pay for such work would be absurd. It would be to pay for it twice: to add a second remuneration to the first. Similar reasoning could, and should, be applied to all the services that you do for yourself. They could, and should, be accounted, but since you then consume them, they are remunerated. We could even skip the transition ‘since you then consume them’, since by definition you consume a service you produce for yourself. Indeed it is often consumed by you at the same time and in the very act of its production (e.g. combing your hair).
Work is certainly being done, and appropriated, but by you. It is work which is remunerated even though it is not paid. It is work precisely because it ‘profits’ someone, but because it is rendered to yourself, it carries its proper remuneration in itself: it is not unpaid, it is not ‘free’ work. It follows therefore that the only work which should be called ‘unpaid work’ is work which is both unpaid and unremunerated. This can only be work done for someone else. In sum, work can be:
1accounted and remunerated (e.g. the work done by a peasant farmer who bakes bread for him or herself);
2accounted and not remunerated (e.g. the work done by the wife of a farmer who bakes bread for her husband);
3not accounted and not remunerated (‘housework’).
The fact that particular work is noted by those who construct national accounts does not imply that it is remunerated, nor does it induce it to be remunerated. On the other hand, the fact that certain work is not remunerated certainly does induce (lead) to its not being accounted. The national account’s differential treatment of work which is equally non-remunerated (e.g. the butchering and cooking of pigs), if it makes no difference to the objectives of the national account (i.e. to balancing the accounts), certainly does introduce a division into the work for self-consumption done within households.
The national account takes the household as a unit. It does not want to know who has done what, nor under what conditions (of remuneration or unpaidness) within it. It is the ‘household’ which ‘produces’ for itself and which consumes. By definition things are handed over to the outside by the household and vice versa. Nothing goes on inside this cell. For the national account there are no individuals, nothing is exchanged or extorted from anyone, within households, and consequently there are no modalities (of exchange or non exchange) to be studied, since nothing takes place. From the point of view of individuals, however, their relations of production are fundamental. Housework therefore cannot begin to exist as a subject of study until we pass the barrier of the household as a unit.
Conclusion: the definition of housework
On the basis of the arguments above, we can begin to specify what we mean by saying that housework is ‘free’ or unpaid. We mean that it is work which is not paid, and which is not remunerated either, since it is done for others. From this point onwards, housework can no longer be defined as either this or that or the other task, or even the totality of the tasks of which it is composed. It must be defined as a certain work relationship, a particular relationship of production. It is all the work done unpaid for others within the confines of the household or the family. From this point of view, there is no difference at all between ‘housework’ and all the other work of the wives of small farmers, nor between their work and the work of wives of artisans or businessmen or wage-workers. Nor is there any difference between the work of wives and the work of other unpaid family workers (male or female) – called in France ‘family helps’.
This doubtless explains why the wives of small farmers can only draw a distinction between their ‘household’ activities and their ‘occupational’ activities with great difficulty, and at the behest of sociologists. One of the reasons why the distinction is difficult to make is certainly that the different tasks overlap concretely, and women pass incessantly from one to another within the same hour. But this ‘reason’ is itself determined by the underlying and principal reason why they do not draw a distinction, which is that the work-of-the-holding and the work-of-the-house are carried out within the same relations of production. Although they may consist of different tasks (each whole, ‘occupation work’ and ‘housework’, being itself disparate and constituting a lot of separate tasks), they constitute but one and the same job: the job of a wife.
While it is legitimate to look at certain tasks women perform in order, for example, to compare or re-evaluate the techniques used now and in the past, as soon as we start to interpret the economic nature of housework, we must adopt an economic definition. This in turn requires that we take into account all the work of wives, that we treat all the work done within the same relations of production as a whole.
To talk of the structural characteristics of ‘house work’ while continuing to define this as a collection of tasks, is a contradiction in terms. If the purpose is to explain (or at least to interpret) the unpaid nature of ‘housework’, then all unpaid work (both ‘work’ and ‘unpaid’ having been defined) must be taken into account. Only then can the boundaries of the object of study be drawn. At present the limits have been set in advance, and empirically. In other words, the empirical object itself has yet to be theoretically determined.
This is why, since the characteristic relations of production of housework are not specific to it, or are not restricted to it, but also characterize other types of tasks and work (e.g. wives’ ‘occupational’ work on farms), I propose that we henceforth substitute the concept of domestic3 work for that of housework – for the object of study is certainly free work carried out in the domus in the wide and sociological sense.
The defective definition of housework, or rather the contradiction between its common practical definition and the study of its relations of production, has, I believe, greatly limited the latter. The demand for wages for all household tasks is a proof of this by absurdity. It is not, however, the most serious manifestation of this limitation. But that is another matter.
* This article appeared first in A. Michel (ed.), Les femmes dans la société marchande, PUF, Paris 1978.