by Faun Rice
élever un nourisson; prendre soin de quelq’un, comme une mère s’occupe de ses
enfants” (1899: 624).
19. That would be mikarakara or mitandrina, or, if the meaning was just curing, mitsabo.
It’s important to emphasize in what follows that mitaiza has never been used as a
general term for any relation of caretaking, but only those of a particularly intimate
or, alternately, ritually significant kind.
270
ON KINGS
celebration of the ancestral dead thus ultimately turns into a war against death
itself, and just as an unlocked or unattended tomb can lead to ghosts escaping
to steal one’s babies, the most appropriate response to locking one, I was often
told, was to immediately go home and have sex, preferably with one’s wife or
husband, if not with anyone who’s willing and available, since doing so at that
moment was most likely to bring new babies into the world.
These are obviously extremely abbreviated descriptions, but they serve to
bring home two points I believe to be critical. The first is that the theme of nur-
ture, and, especially, women’s labor in the rearing of children, is a key way of im-
aging the creative power of ritual; the second, that such relations tend to become
the locus of reversals of authority structures, where ancestors turn into children,
or slaves turn into vehicles for kings. These reversals don’t necessarily overthrow
or challenge relations of authority (though they can), but they definitely become
a way of negotiating such relations. All this is important when one turns back
to royal ritual, because the moment one begins to look, one starts seeing all sorts
of obvious parallels—even though these are precisely the elements that might
otherwise be most likely to be overlooked if one did not have this larger ritual
context in mind.
Understanding the ritual logic is all the more crucial because highland king-
doms—whether the various warring principalities of the eighteenth century or
the great imperial state of the nineteenth—were essentially organized on ritual
terms. This is not to deny that they were not also vast forms of labor extraction;
rather, it is to suggest that within them no clear distinction between what we
would call “work” and what we would call “ritual” could be made (cf. Sahlins
1985: 113). Both were seen as necessary, and overlapping, aspects in the con-
stant human efforts required to create and maintain the material and social
universe.20
We might state the matter this way: What we call “societies” are always vast
coordinated systems of ritualized labor. Always, too, the elementary unit of any
such system is some kind of household; however it might be configured (and
as anthropologists know, there are an enormous number of possibilities in this
20. It is worth noting that while there is a Malagasy word for “work” ( asa), there is no
generic word for what we’d call “ritual”—the closest is, perhaps, fanasinana, which
is the term for the actions of making things masina, powerful or sacred. But not
all forms of ritual were considered fanasinana—mortuary ritual, for instance, was
not. What’s more, many types of ritual—notably, mediumistic curing—could often
themselves be referred to as “asa” or work.
THE PEOPLE AS NURSEMAIDS OF THE KING
271
regard), this household is the elementary unit of work, solidarity, domination,
and the creation and fashioning of human beings. In this sense women’s labor,
which tends to predominate within households, is also the most fundamental
form of work itself. These statements, I think, can be made of any known hu-
man society. Obviously, this is never all. Anything we feel merits the name of
a “society” will also have more encompassing structures through which all this
is coordinated: clans, temple complexes, joint-stock corporations, and so forth.
Usually, these look quite different from households. What makes monarchies
unique is that, much though there might be all sorts of things going on in the
middle of the hierarchy, the very top almost exactly resembles the very bottom.
Kingdoms not only begin with households, they also end with one. This apical
household is, of course, the royal household.
Now, certainly, royal households are almost invariably far more elaborate
than any ordinary households (which do not, for instance, ordinarily involve
harems attended by eunuch slaves, dancing dwarfs, and so forth), and often they
are organized in ways that can only be described as intentionally perverse, in the
sense of violating the basic moral principles that govern ordinary households
(led by men who murder their fathers, marry their sisters, and so forth), but they
are considered households nonetheless; in the final analysis, they are the same
sort of unit, a domestic unit creating, nurturing, and educating children—that
is, producing people—as the household units at the very base.
The difference, of course, is that the royal household, in the vast majority
of cases anyway, was only about the creation of people, and did not involve all
those other forms of production—of food, clothing, ironware, basketry, and so
on—that in ordinary households served as essential material elements for that
process of tending, growing, and nurturing human beings. In fact, the tendency
in those ordinary households is for what we would label “work,” “play,” “ritual,”
and “education” to be, not indistinguishable perhaps, or not usually, but in every
way entangled, overlapping, and mutually entailed. In contrast, royal households
could be seen as the first prefiguration of the modern consumer household,
which at least ideally is set in a sphere entirely opposed to the “production” of
material goods, which is just about the creating and shaping and maintaining
of people. Royal households largely divorce the making, shaping, and maintain-
ing of people from the making, shaping, and maintaining of things. They also
separate work, ritual, and play—at least, when royal figures do plow fields or lay
the foundations of buildings, it’s almost invariably as a form of ritual play; it’s
not considered actual labor. Royal households will tend to be full of servants and
272
ON KINGS
retainers, some of whose work is to continually teach members of the royal fam-
ily how to behave like proper royalty, just as, say, presidents and prime ministers
nowadays tend to be surrounded by hosts of aides and advisors whose function
is to teach and remind them how to be proper heads of state. But rarely do they
do a lot of what we’d consider productive work. What we’d call “material pro-
duction” tends to be outsourced onto other households.
This might seem an odd way to frame things. Surely, kingdoms are not run
by a royal household. They’re run by an individual called “the king.” Which is,
of course, true: no kingdom has ever been ruled by a household collectively (at
least, not officially), and in principle, the entire purpose of the royal household
is to produce that one single individual—the monarch—who is, properly speak-
ing, the entire focus of all ritual l
abor. But the apparatus for the creation of such
individuals is nonetheless crucial, since such individuals, for all their occasional
insistence to the contrary, are mortal, and can be replaced at any time.
I’m emphasizing all of this—at least some of which might seem self-
evident—because I think it’s important to problematize received categories in
order to understand what’s really going on in monarchies, which are, after all, an
extremely common form of government historically, even if they seem strange
and exotic to most people today. The terms of political economy, invented in
the North Atlantic world around the same time that modern republican forms
of government were being instituted in the late eighteenth century, really don’t
seem adequate to describe them. These terms propose a very simple schematic
version of what an “economy” is, one that has become so much a matter of com-
mon sense that we have to remind ourselves that for most people, the very idea
of a division between spheres of “production” (in workplaces) and “consump-
tion” (at home) would simply make no sense. Some of the attendant political
economy categories have been thoroughly critiqued by anthropologists. Others
have not.
Take, for instance, the commonplace notion that labor is basically about
“production”: that it’s typically directed at making things, transforming the
world by combining raw materials into finished products. This is simply as-
sumed in most theoretical literature. But it’s a very odd assumption. Even a
moment’s reflection should make clear that nowhere in the world is most activ-
ity we would ordinarily refer to as “work”21 directed at making anything. This is
21. Let us define “work” here as repetitive or formalized activity not performed for its
own sake, but primarily to change the state of something else.
THE PEOPLE AS NURSEMAIDS OF THE KING
273
true even if we restrict ourselves to work directed at material objects. Most such
labor isn’t aimed at producing things, but at cleaning or maintaining them or
moving them around. A ceramic coffee cup is “produced” just once; it’s washed
and stacked a thousand times. Even if it’s a disposable paper or styrofoam coffee
cup, far more time and energy is spent transporting, storing, and disposing of it
than in the relatively brief moment of its actual fabrication.22
This blindness has any number of pernicious effects, but for present purposes
I am just invoking it to make three points which I think are crucial to under-
stand the organization of labor in Merina kingdoms. The first is the obvious an-
thropological one: that we cannot presume Malagasy assumptions about what
work is and what is important about it necessarily mirror our own. And they
are indeed quite different. The second is that, as I’ve argued elsewhere (Graeber
2007c), assumptions about the nature of work tend to be organized around cer-
tain forms of what I will call “paradigmatic labor”—that is, certain varieties of
work effectively stand in, in the popular imagination, for a whole class of other
ones. In contemporary social science, and, to a large extent, popular discourse as
well, the two most important of these are factory work and childcare. The first
has become the paradigm for all paid work, the second, for unremunerated, do-
mestic labor. This is the imagery lurking in the background, for instance, when
Marxists speak of “productive” versus “reproductive” labor; this is what can allow
popular commentators to blithely declare that the decline in the number of fac-
tories in Britain or America means there is no more working class in such coun-
tries, even though there’s probably never been a single place in the history of the
world where the majority of working-class people were employed in factories.
The third point is that in monarchies like the Merina kingdom, the easiest
way to understand how work was imagined is by examining the forms of ritual-
ized labor surrounding kings.
Let me introduce another distinction here, and cite a term I introduced in an
earlier work (Graeber 2007c): “emblematic labor.” If paradigmatic labor is what
you imagine to be the model for work in general, or a whole broad category of
work, emblematic labor is work seen as typical of a certain group of people, a
kind of work that defines what sort of people they are seen ultimately to be. A
“fisherman” might spend only a relatively small proportion of his time actually
fishing, he might even get a larger proportion of his income from something
22. In this sense, even my previous comments about households and production aren’t
precisely right.
274
ON KINGS
else, but a “fisherman” is nonetheless what he is basically seen to be. In many
societies, this emblematic labor is the kind of work that sort of person does in
ceremonial contexts. Caste systems are an obvious case in point: drummers or
washermen or barbers in an Indian village do not spend most of their working
days making music, washing, or cutting hair. Yet those activities define their role
in the larger society, largely by determining what role they play in important
cosmological rituals.
A. M. Hocart (1950) in fact insisted that caste systems of this sort originate
specifically in royal rituals, where, as in Fiji, different groups in a kingdom were
defined by the kind of labor they performed specifically for the king. Thus in
Fijian kingdoms, there were certain groups identified as fishermen—not be-
cause they spent most of their time fishing, or even because they spent more
of their time fishing than anyone else (everybody fished), but because it was
their responsibility to provide fish to the court or royal rituals. Each Fijian clan
was characterized by its own form of emblematic labor, and this was seen as
establishing what kind of people they really were. Eighteenth- and nineteenth-
century Merina arrangements have been referred to as “caste-like” for similar
reasons (e.g., Bloch 1977): different ancestries, and different orders of nobility,
were defined by the work they performed for the royal household (particularly,
the products they brought as first-fruits or santatra), and the role they played in
building royal houses and royal tombs.
I think examining emblematic labor in the Merina kingdom in this fashion
is the best way to understand the structure of the kingdom, and the role of
the royal household. But in order to do so we first must say something about
paradigmatic labor. In highland Madagascar, the paradigm for work in general
is not production or even, precisely, childcare. When people think of “work,”
they think first and foremost of the bearing of burdens: moving, dragging, and,
especially, carrying things, which includes everything from carrying babies to
dragging trunks of wood to moving earth with shovels. The semantic range and
web of associations is quite different than we are used to. But once we under-
stand this, a lot of other things begin to make much better sense.
Speaking, carrying, and making
The essentials of the matter do not seem to have changed much sin
ce the nine-
teenth century. Then as now, work was seen as centered on the household, and
was primarily the business of women. This is not to say men did not work at
THE PEOPLE AS NURSEMAIDS OF THE KING
275
all—often they worked very hard—but women spent more of their time in ac-
tivities viewed as working ( asa), and women’s general dispositions were seen as
pragmatic, industrious, and generally work-oriented in a way that men’s simply
were not.23 If I were to guess what kind of paradigmatic image was called to
mind by the idea of work—parallel to the clock-punching industrial worker
of our own imagination—it would most likely be a young woman, infant slung
over her back, laboring in a rice field, or carrying water or produce on her head.
Domestic labor here represents the perfect fusion of child-rearing and physi-
cal work because women tend to attend to children and carry out other duties
simultaneously.
The paradigmatic form of work, then, was—and is—not seen as a making
or building anything, or even maintaining anything, but, rather, lifting things
up and moving them around. The importance of such matters can be seen in
that fact that, traditionally, who carries what for whom in what circumstances
is carefully regulated, at least in formal or ritual moments—indeed, formal or
ritual moments are above all those in which the niceties of etiquette in such
matters are strictly enforced.24 Who carries what for whom is probably the most
important way of indicating rank. Even well-educated, not particularly tradi-
tional women, I found, would on trips occasionally make (half-hearted) offers
to carry my backpack, noting that, properly, if a man and woman are of roughly
the same age, the man shouldn’t be the one shouldering the burden. But mainly
these rules applied to seniority. As elsewhere in Madagascar, the ranking of
children by age is especially important. Indeed, insofar as one can speak of an
“atom of hierarchy’” in highland society, it is embodied in the principle—re-
peated in proverbs—that elder brothers or sisters should speak for their juniors,
and juniors, carry burdens for their seniors.25 Often this is treated as a reciprocal
23. Women in general endlessly denounced males as lazy ( kamo lahy), either individually