5

PROVISIONAL AUTONOMOUS ZONE:

OR, THE GHOST-STATE IN MADAGASCAR
Shordy before I left for Madagascar I was ralking ro Henry Wright, an
archeologist who had worked there for more than a decade. "You have to be
careful," he said, "poking around the countryside." State authority was dis-
solving. In many parts of the island, he said, it had effectively ceased to exist.
Even in the region around rhe capital there were reports offokon'o/o11J1-vil-.
lagc assemblies-beginning to carry our executions.
This was one of the many concerns forgonen almost as soon as I actually
arrived in Madagascar. In the capital, there was quite obviously a function-
ing government; almost every educated person seemed to work for it. When
I moved to Arivonimamo, a cown about an hour to rhc West, things did not
S"m particularly differem. Cenainly, people talked about the government
all the rime; everybody acted as if there was one. There was an administra-
ti\"t SITUCture, offices where people typed up documents, registered things,
kept track ofbiHhs and deaths and the number of people's canle. One even
had ro get permission to carry out the most important rituals. The
111(01 ran schools, held national exams; there were gendarmes, a prison, an
airhdd with military jets.
It w.u only after I had been in Arivonimamo for some time-and even
more:, in retrospect, after I'd left-that I began to wonder whether what he
wid might actually have been true. Perhaps it was simply my own bias,
tht fact that I had always lived under an efficient and omnipresem govern-
ment, that made me read the cues the wrong way. Perhaps there really >A"asn"t
a state in Betafo at all; perhaps not even in Arivonimamo-or anyway, not
ont that behaves in any way like what I or other WestC'rnC'rs have comC' to
i.\Sume a state is supposed to behave.
fkfore I explain what I mean by this, though, perhaps it would help to
tht scene.
158
POSSIBILITIES
Arivonimamo and Berafo
I arrived in Madagascar on June 16, 1989. For rhe 6m six momlli, 1
lived in Anrananarivo, rhe capiral, scudying rhe language and doingarchivJ
research. The Narional Archives in Antananarivo are a remarkable
In their collection are thousands of documents from the nineteemh cenrur.-
kingdom of Madagascar, most from the highland province oflmerina, vhich
surrounded rhe capital. Almost all of ir was in Malagasy. I wem throug..1
hundm:ls of fOlders, carefully cop)'ing out everything concerned with tht:
dimicr ofEastC"rn I mama, rhe pare oflmerina in which I intended tov.m._
Eastern lmamo seemed, at rhe rime, to ha,e been a rather sleepy place,
hinterland far from the rumulruous political struggles of the capital, bu!
at rhe same rime insulated from the unstable fringes of Imerina,
territories full of raiding band irs, industrial projects, and periodic m-olu. h
was a place where nor much ever happened-and, rhus, rhe perfect fidd on
which to srudy the slowmoving processes of social and cultural change I \\l!
inrerested in.
Once I fdr I had a minimal command of Malagasy, I ser our fo:
Arivonimamo, the major town of rhe region. It was not at all difficult ro gn
there: Arivonimamo is only an hour from the capital by car. Before long I
had established myself in rown, and had begun making regular trips to d"
surrounding counrryside, gathering oral histories, keeping an eye our for 1
likely place to do more derailed research.
Arivonimamo is a town of some ren thousand people rhat clumn
around a stretch of the main highway leading wesr from the capital. In r!K-
1960s and 1970s, it had been the home of the national airport, which 521 in
a broad valley ro the south of town; bur though rhe airport brought
and employmem, it never seemed to become an imegral part of rhe town,
economy. It was largely a thing grafted on. The road from the airpon did
not pass through Arivonimamo itself; there wasn't even a place for mrekn
to spend the night there. In 1975, the airpon was replaced by another. ne.utt I
the capital. The old airport was given ro the military, which card}' ho.,.,"t'o-er. !
L

.. r ...
had once passed through here was rhe banered plywood shell of an emfl':!
restaurant, sranding where the airport road merges with the just on
the outskins or town.
The current town centers on a taxi srarion, a wide asphalt expall)e
flanked by tWo churches, Catholic and Protestant. At most hours, ir """15
crowded wirh vans and station wagons filling up with passengers; and hlp
and cr.nes a.nd heading otT the capital, or f urrher west down the
On the southern edge of rhe taxi stand is a wide spreading amont4114 trre.
very anciem sycamore that is considered the symbolic center of the town,rh=
I'ROVISIONAL AUTONOMOUS ZONE
!)9
mark rhar it wa.s once the place of kings. To irs nonh is a markc=tplace wirh
food stands and red-riled arcades, which every Friday fills co overAowing
with rural people and vendors under white umbrellas. The town itself clings
to 1he road (the only place rhere is elecrriciry); irs houses are mosrly two or
thm: stories, wirh graceful pillars supponing verandahs around rhe second
Aoor. and high-pirched roofs of lin or rile.
Ari\'Onimamo is rhe capital of an administrative district of the same:
name. h contains several governmem offices and three high schools: one
state school (CEG), one Catholic lycee, and one Protestant one. There is
a clinic and, on a high bluff somewhat to the west of rown, a small prison.
Together with a gendarmes' barracks nearer the old airporr, a post office=.
and a bank, these consrirute the government presence. There was once a
factory nearby bur it had been abandoned for years by the time I was there;
no one I knew was quite sure what, if anything, had ever been produced
rhere. The town's commercial economy fell almost completely outside the
formal (raxed, regulated) sector: there was a pharmacy and two large geneldl
stores, but that was about it. Otherwise. the population conformed to the
gtner.tl rule for Malagasy cowns: almost everybody grows food; everybody
sells something. Streets were fringed with dozens of little booths and stores,
all stocked with the same narrow range of products: soap. rum, candlc=s,
cooking oil, biscuirs, soda, bread. Anyone who had a car was a membc=r of the
tui collective; anyone who had a VCR was a theater operator; anyone who
had a sewing machine was a manufacturer of clothing.
The province of Imerina has always cenrcred on the gigantic irrigared
plains surrounding the nat"ional capital, Antananarivo, which have long had
a \"try dense population and been the cemer of powerful kingdoms. In the
nineteenth century, the Merina kingdom conquered most of Madagascar;
since the French conquest of 1895, Anrananarivo has remained 1hc center of
administration, and the surrounding territory the ancestral lands of
most of Madagascar's adminisuators and educated elite. The territory that
now makes up the district of Arivonimamo was always somewhat marginal.
It wl.s late to be incorporated into the kingdom, and it was never more than
integrated into the networks of cash and patronage centered on the
capi1al. So it remains. Now, as then, it is a political and economic margin, a
plact where not much ever happens.
To the north of Arivonimamo is a rolling country of endless red hills,
wme covered only with grass, mhers wooded with eucalyptus trees, stretches
of wpM-which look like dwarf oaks-and occasional stands of pine. The
hilb are cut by narrow twisting valleys, each carefully terraced for the cuhi-
of irrigated rice. Here and 1here rise gran ire moumains, supposed to
luvt been the sears of ancient kings.
In this back country, there are no paved roads. People walk-very
can afford bicycles. Goods are transported in ox-carts, along mud pathsthlt
are, even in winter, too runed for any but the toughest automobiles. With
the start of rhe summer rains, they become impassable. It is largely because
of the difficulties of communication that there is no large-scale commercial
agriculture, despite the proximity of the capital. Farmers do end up canm
a fair proportion of their crops co markets in cown, and much of rhis
up helping to feed the population of Antananarivo, bur it's all piecem(';ll,
individual cultivators selling to very small-scale merchants in an cndb
1
muhitude of tiny transactions, almost as if people were intentionally trying
to ensure char the meager profits to be had from buying and selling local
products ended up divided between as many hands as possible.
As I have said, my first work was on oral history: I started visiting villag-
es usually accompanied by one or two Malagasy friends from Arivonimamo.
I ended up fixing on the village of Berafo in which to carry out.my imcnsi1t
fieldwork: a community that fascinated me, in parr, because it was divided
almost evenly between andriana (usually translated .. nobles") and the de-
scendants of their former slaves. Betafo lies along the southern Aank of a
long mountainous ridge called Ambohidraidimby, most of it only a thiny-to
forty-minute walk from the center of Arivonimamo. Ir is close enough thlt
one can live in rown and sdll cultivate one's fields in Betafo-as many peoplt
do-or have a house in both places and move freely back and fonh betwnn
them.
Most rural communities in Jmerina have some economic specializ.arion,
which occupies people especially in winter. In one village, the men will all
be butchers, in another the women all weave baskets, or make rope; spact1
in the marketplace in Arivonimamo arc mapped out as much by the ori
gin of the vendors as by the goods they have for sale. The people of Beufo
have been traditionally known as blacksmiths. Nowadays, roughly a third
of irs households still have a smithy out back. Of those who do nor, a nr:o-
largc number arc involved in supplying smiths with iron ingots, and sell-
ing the plows and shovels rhcy produce in markets and fairs in other paru
of lmerina. What had starred as a local effort had, by the rime I rhrn.
npanded dramatically, since in most of the region to the west of the upiul.
Betafo was mainly known for 'idling plows, despite the fact that no one in
Bcufo iudf actually produced plows-they were all manufactured in Olhcr
in rhc vicinity of Arivonimamo, with iron supplied by specularon
from Bc:tafo.
The imemihcarion of commerce is one response to the economic crurKh
thar has cauw:d a dramatic fall in standards ofliving throughout MadagaJCJI
5ince the 1970.1i. lr led ro a great increase in side occupations, !.0 rhar in any
PROVISIONAL AUTONOMOUS
IGI
one household, one woman might be spending much of her time running a
coffee stand in town, or weaving, another making fermented manioc to sdl
to vendors in the marker, one man driving an pan time and
ing several months a year selling pineapples in a different parr of lmerina,
while yet another might only drop by in the country occasionally, spending
most of his days refilling disposable lighters near the in town. All
this makes membership in a community like Betafo a bit hard to define. Not
that I was trying to gather much in the way of statistics. In fact, one of the
peculiar effects of my situation was that I had some fairly detailed bits of
formation about the demographics and property-holdings of the inhabitants
of8e1afo in the 1840s and 1920s, culled from the archives, I never managed
to get such statistical information for the rime I was actually there. This fact
is imponam. I think it reveals something quite profound, acmally, about
what son of place I was acwally in.
While I was living in Arivonimamo and working in Betafo, I spent a lot
of rime thinking about the political aspects of conducting research. Almost
all anthropologists do. In my case, it was especially hard nor to be a little
self-conscious in a milieu where urbanites seemed to find a special joy in tell-
ing me how terrified country folk were of Vazaha (people of European stock,
such as myself}-and country folk, in telling me how terrified children were.
For most Malagasy, the very word "Vazaha" evoked the threat of violence.
Fortunately for me, it also had as irs primary meaning "Frenchman," and
{as I endlessly had to explain) I did not even speak French. Speaking only in
Malagasy rook a bit of the edge off things. But even more crucial: conducting
research itself had associations. On the one hand, Imerina is a highly literate
society: no one had any problem understanding what I meam if I said I was
an American student carrying out research for his doctorate in amhropology.
Nor did anyone seem to doubt that this was a legitimate, even an admirable
thing to be doing. Bur techniques of knowledge were very closely identi6ed
with techniques of rule, and I quickly gO[ the impression 1hat there were
ctrtain sorts of inquiry people were much more comfortable with than oth-
as. Perhaps I was overly sensitive, but as soon as I got the feeling I was mO\r-
ing onto territory someone didn't want me delving imo, I desisted. I would
rather people talked to me about the things they wanted to talk about. A5 a
rtsulr, I know more about the distribution of property in Betafo in 1925-or
even 1880-than I do for the rime I was there. Property survep V.'C're the
\Oft of the thing governments would carry out, backed by the threat of force.
in order to aid in rhe forcible extraction of labor or taxes. This meant thar
there were extensive records in the archives; it also meant it was ex.actly wh.ar
people wamed to be sure I wasn't ultimately up 10. Even the act of synemati-
tally going from door to door surveying household size would have bttn ..
well, norhing would have been more guaranteed ro get people's backs
11
Lack of hard numbers seemed a minor price to pay. ?
The Very Existence of the State
Let me return, then, to the initial question of rhe stare.
Was there a government in Arivonimamo and the surrounding counuv.
sidd On one level, the answer was perfecdy obvious. Of course rhere

There were government personnel, government offices, and ar leasr in
1011
n
government-run schools, banks, and hospitals. Almost all economic rransac.
tions-even if rhey were generally off the books-were carried our uJing
government-issued Malagasy currency. The territory as a whole was claimrd
under the sovereign authority of a Malagasy stare that was recognized by .UJ
other stares in the world, and no one, in chis territory, was openly comesting
rhat state's sovereign authority. Certainly, there was nobody else claimin;
to represent a different stare or claiming ro represent a political ahernari1e
there were no insurrectionary communities, no guerilla movements, no po-
litical organizations pursuing dual power strategies.
From a different perspective, though, the situation looked quite
ent. Because the Malagasy stare, in this region at least-and this was a region
quite close to its center of power in the capital-was either unimeres1ed in,
or incapable of, carrying out many of what we consider ro be a Slate's
elementary, definitional functions.
The key issue in most Western definitions of rhe state is its power 10 co-
erce. States employ "force"-a euphemistic term for rhe threat of
to enforce rhe law. The classic definition here is Weber's: "A compulso)
political association with continuous organization will be called a if
and in so far as irs administrative staff successfully upholds a claim 10 the:
monopoly of rhe legitimate use of physical force in the enforcement of iu_
order" (1968 I: 54). Bur Weber's definirion was itself really jusr a matttr ol
repeating rhe convemional jural wisdom of his day. In fact, he seems ro
been drawing directly on rhe work of an earlier German legal theorist nama:!
Rudolph von lhering, who in 1877 had defined rhe state this way:
The Srau: is the only competent as well as the sole owner of social oxr-
cive force-the right to coerce forms the absoluu monopo(y of the Srarc.
Every association that wishes to realize its claims upon its members
means of mtchanical coercion is depcndcnr upon the cooperation ofrhe
State, and the State has in its power ro fix the ..:onditions under which it
will granr such aid (cited in Turner & Factor 1994: 103-104).
PROVISIONAL AUTONOMOUS ZONE
163
A definition like this is mainly a way m focus the mind; it is not of thar
much use for determining whether or not any panicular organization is a
state, since for that, everything depends on whether or not one feels a would
be state has been "successful" in claiming its monopoly. Nonetheless, these
definitions do capture rhe implicit common sense behind modern Western
insritucions of government-one in no way foreign to the Malagasy sme,
which was organized very much on this same model under the French colo
nial regime, and whose currenr form is based largely on colonial instimtions.
And most Malagasy, I think, would have agreed that the ability to apply
force in this way was, essentially, what made a state what it was. This made
it all the more striking that, in most of the Malagasy countryside, the state
had become almost completely unwilling to do so. Far from maintaining an
absolute monopoly of the right to coerce, or to authorize others to do so, the
state simply did not exercise what was ostensibly its primary function there
at all.
In the capital, there were police. Around Arivonimamo the closest thing
to a police force was a unit of gendarmes who had a barracks somewhat to
the west of town. Mainly, they patrolled the highway. Occasionally, I was
told, they would fight bandits further west; but they did not like ro travel
off the paved roads, over the rutted dirt tracks that led into the countryside
where almost everyone accually lived. In the countryside, gendarmes would
never show up unless someone had been murdered. Even then, it would usu
ally require something drastic-like a large number of wirnesses appearing
at their doorstep demanding they rake action, and, usually, having already
rounded up the culprit(s) themselves-before they would actually come and
take anyone away.
Even in town, they did not act much like police. In Arivonimamo I
heard a lot abour a bully named Henri, a large and powerfully built man,
perhaps insane (some said he was just pretending), who had terroriud its
inhabitants for years. Henri used to help himsclf[O merchandise at the local
shops, daring anyone to stop him; he was a particular danger to the town's
young women, who lived in constant fear of sexual assault. After much dis
cw.sion, the young men of the cown finally decided to join together and kill
him. This rook some time to arrange because, in fact, then: was an informal
tndition in that parr of the highlands that if one wishes to lynch somronC'.
one has [0 get their parent's permission first. Normally this is just an effecti\'e
~ ... y to reinforce parental authority, a kind of ultimate sanction-or, a Yr.IY of
allowing someone's mother or father to inform them that it's re.ally time one
should be getting our of town-but in this case, after many vain efforts to
apprise his son of the seriousness of the matter, Henri's father threw up hi!.
h.ands and allowed things to take their course. The next time he provoked
a fighr, a crowd immediately appeared armed wirh knives and agriculruQI
implemems. As ir turned ouc, they didn't quire succeed in killing him:
wounded, Henri managed m take refuge in rhe Catholic church and de.
manded sancmary, claiming persecution due m menral illness. There, no
one was willing m follow him. The Italian priest hid him in the backofavan
and smuggled him our roan insane asylum. He was soon discharged (he ba
1
rhe O[her patients), bur didn't dare show his face again in Arivonimamo fo
1
many years to come. The first rime I heard the smry 1 was mainly
in rhe derails of parental permission. Only later did it occur ro me rhar rhi
1
evenr rook place in a town with an actual police station. How could Henri
have managed to terrorize the town for years withom anything being dont
about him? "Why hadn't rhe gendarmes done anyrhing," I asked? "Ha\'tn'r
you seen Henri," people would reply. "He's enormous!"
"Bur the gendarmes had guns!"
"Yes, bur even so."
Events like this were in every way exceptional. The most significam
thing about violence around Arivonimamo was rhat there was very link
of it. Murders were shocking, isolated events; there were very few HenrU..
Nonetheless, rural assemblies had m develop all sorts of creative srraregies ro
overcome rhe reluctance of the forces of order to enforce rhe laws. ToYiltdJ
the end of my stay, there was afokon'olona meering in Berafo-a villageaJ
sembly-ro deal with an instance of violence. A man named Benja, notori-
ous for his fiery temper, had a quarrel with his sister over some murual busi
ness arrangement, and, the story went, had beaten her to within an inch of
her life. Acrually, stories varied considerably abour how badly she was rC'lll!
beaten, but the matter was considered a very serious affair requiring immtdi
ate atrention. After much deliberation, thefokon'olona ordered Benja ro
an undated letter confessing to having murdered his sister. and rhen, broughr
the confession down to be lodged ar the local gendarme station in (Own.
That way, if his sisrer was ever robe found the vic rim of foul play, he ...,oo1J:
already have confessed and could simply be delivered to the .authorities. The
message was rh.at his sisrcr's safety and welf.being were to be his pmo!U!
responsibility from then on. In rhis case, rhe state was being used as a
of ghost-image of authority, a principle but nor a rhreat, since if his sistcrw;l.l
found dead, the foko11'olona rhemselves would have to be the ones to .JJT(Sl
him and carry him down to the gendMmes' office; rhe would mm+.
make ir much more likdy rhat he would then have to spend some rime in i.Jii.
In other cases, the stare aurhorirics were bypassed entirely. The 1980s.. tOr
example, began ro see rhe revival of collecrive ordeals. In a case of thcfr-fix
instance, in Bemfo, afrcr someone had made off wirh rhe entire conrems ofJ
rice storage pir belonging ro a promincm elder-elders would gathe-r a wh..lit
PROVISIONAL AlfTONOMOUS ZONE
community together, and each would drink from a specially bowl
or ear a piece of a specially prepared liver, and call on their ancestors to
them down if they were guilty. The next person who a death was
thus presumed to be a victim of ancestral Two such or
deals had been held in Betafo in the decade before I there.
even rumors, further our in the countryside, of the of actual
poison ordeals. Everywhere, one began to hear about invisible powers enforc
ingjustice-buried charms, standing stones, of sacrifice newly
charged wirh the power to detect and punish evildocrs. Almost of
any wealth or political prominence started to begin hinting that they might
have access to dangerous magical powers: hail or lightning charms, vindic
rive ghosts, access to the protection of ancient kings. Anyone who
to amass-or maintain-a great deal of wealth had almost by ddinition to
be able to at least create the suspicion in others' minds that they might have
access to dangerous hidden powers of some sort or another. Bur it was a
delicate game: since anyone who boasted openly of such powers was assumed
almost by dehnirion nor to really have them, and anyone who employed dan
gerous magic against their fellow villagers was by definition a witch. I even
hard rumors of wealthy men deep in the coumrysidc who so infuriated their
neighbors by dark hints of magical powers that those neighbors eventually
sought counter-medicine, disguised themselves as bandits, and attacked and
ransacked their possessions.
The Slate as Guaranlor of Property Relations
Theories of social class almost always assume that a key role of the
nate-perhaps, iiS most important role-is ro underpin properry relations.
For a Marxist, cenainly, this is a state's primary reason for being. Conrracrual,
market relations can only exist because their basic ground, the basic rules of
1he game, are enshrined in law; those laws in turn are efftive only in so far
u everyone knows they will be backed up-in the last instance-by dubs
and guns and prisons. And, of course, if the ultimate guarantor of propeny
relations is state violence, then the same is true of social classes as well.
But, in the countryside around Arivonimamo, the state simpl)' did
not play this role. I cannot imagine a situation under which it would dis
patch armed men to uphold one person's right to another from their
Land-let alone to enforce a contract or investigate a robbery. This. too, \\""13
somtthing whose full signihcancc dawned on me only afterwards, because
everyone acted as if the government did play a crucial role in such mat
tc:n. The government kept track of who owned each piece of land: whencvn
somcom: died, the division of their fields and other property was mcticu
lously recorded ar the appropriate offices. Registering property, along wi
1
h
binhs and deaths, was one of the main things such offices did. There werr
all sons of laws concerning land, and no one openly contested them, just as
when talkiflg in rhe abstract, they always spoke as if they felt land regis!ra-
tion did give an accurate picture of who had ultimate rights to whar. In
practice, however, legal principles were usually only one, relatively minor,
consideration. If there was a dispute, legalities had to be weighed again
11
a welter of"rraditional" principles (which usually provided more than onr
possible solution to any given problem), the of former ownm, and
not least, by people's broader sense of justice-the feeling, for instance, !har
no accepted member of the community should be completely deprived of1he
means of making a living. Certainly no one would think of taking the ma!-
ter to court-except in a few rare cases where one of rhe disputants was an
outsider. Even then, the court served mainly as a neutral mediator; everyone
knew no police or any other armed official would enforce a court decision.
In Arivonimamo, in fact, there was one man wirh a gendarme's uniform
who would occasionally rent himself our to money-lenders or merchants 10
imimidate people imo paying debts or surrendering collateral. An acquain-
tance of mine from Betafo was terrified one day when he showed up in rhc
company of a notorious loan-shark-even after his neighbors explained w
him that the man could hardly be a real policeman, because, even if you
could find an officer willing to trudge out into the counrry on such a ui,i1l
matter, lending money at interest was against the law for private indi,iduail
and a real gendarme would have had just as much cause to arrest his credi1or
as he. This struck me as a parricularly telling case, because it underlinrd jui!
how litrle the forces of order cared about economic affairs. Normally, thm
nothing more guaranteed to infuriate police than the knowledge that some-
one is going around impersonating an officer. Doing so strikes at the \"C:f!.
essence of their aurhority. If this particular impostor got away with
he apparently did-it appeared ro be because he confined his acti\itirs w
a domain in which the gendarmes had no imercst. Afrer all, rhe genJarroo
never did anything to protect shopkeepers from Henri, eirher-and that w15
in town; the counterfeit officer seems to have confined his activities alm0:11
exclusively to rhe countryside.
There arc various ways one might chose co asses.s this situation. Ooc
would be ro conclude thar people of rural Imerina, or in Madagascar in gm
eral, had a different conception of the state than Marxists and Weberi.msJtt
used to. Maybe the prorcction of property is simply nm one of the functions
anyone expects a government to fulfill. To the cxrenr people secmaJ ro
otherwise, they might just be paying lip service to alien principles
by the French colonial But, in fact, tile precolonial Merina sutC'\\1.1
PROVJSIONAL AlJTONOMOUS ZONE
16i
veritably obsessed with protecting property. King Andrianampoinimerin.a,
its founder, emphasized this role constantly in his speeches (Larson 2000:
192). Law codes, beginning with his own, always made the regulation of
inheritance, rules about buying and renting, and rhe like, one of thtir most
imponam areas of concern. Even the registration of lands predates the co-
lonial period; records began to be kept in 1878, sevenreen years befort the
French invasion.
On the other hand, existing evidence gives us no reason 10 believe thai
people then paid much more attention to this elaborate legal struCiure rhan
they do today-although neither is there any record of anyone openly chal-
lenging it. Legal systems have always been accepted in principle, and ap-
pealed. to only very selectively in practice. Mostly, people go about their busi-
ness much as they had done before. It is this phenomenon, I chink, which
gives the best him as to what's really going on.
Let me make a broad generalization. Confronted wich somtone bem
on imposing unwanted authority, a typical Malavsy response will be to
agree hearrily with whatever demands that person makes, and chen, as soon
as chey are gone, to try to go on living one's life as if the incidem had nC\er
ha.aqfine<LOoe_ c:v_en sav _chis was the archetvgipllv ,Malae:aJl' ,wav ,
of dealing with authoriry: one's fi.rst line of defense is simply to deny thou
rhe event in question (a governmem offi.cial coming to count canle and an-
nounce the required tax payments, or negotiate the: requisitioning oflabort:n
to replant trees or build a road) ever occurred. Admittedly, it is hardly a srr.u-
egy limited to Madagascar. Something along these lines is often considered
a typically "peasant" strategy: it is an obvious course to take: when one is in
no w.ay economically dependent on those trying to tell onr what to do. But
thrre are many other routes to take, all sorts of possible combinations of
confromation, negotiation, subversion, acquic:scence. In where
there is often a strong distaste for open confrontation in daily life in gt:neral.
the preferred approach has always been to do whatever it t:lkcs to make the
annoying outsider happy until he goes away: then, insist that he had nt"Vtr
bec:n there to begin with, or if that doesn't work. to simply ignore wh.atC\cr
one: has .agreed with and see what the consequences might be. It even t.akC1
on a cosmological dimension. Malagasy myths on the origins of dr.ath claim
rhatlife itself was won from God in a dealt hat humans nl"Ver rc-alh intendn.l
to kttp (hence, it is said, God kills us). Here is one, drawn early in rhe cc:n-
tury from the Betsimisaraka of thC' C'ast coast. There :nc c:ndleu
most obviowly tonguc-in-chc-c:k, with the often bearing an
rescmbl.anc.e to the son of passing colonial official who would
Uw:.w up in villages, with armed retainers, demanding rhe p.z,-mmt of tun:
168 POSSIBILITIES
Once upon a time, a Vazimba [aboriginal] couple were the only two
of the earth. They were sad because they had no children,
50
one day they found some clay and gave it human form. They made two
figures, one a liulc boy, the other a liule girl. The woman blew in 1hc:ir
noses to ani moue rhem bur she wasn't able to give them life. Then, one
day, she happened to meet a god who was traveling on the canh. The
woman asked him ro give life to the two statues and promised him, ifhe
succeeded, two cows and a sum of money. So he did so.
When the children grew up, the parems married them to one an-
other. Then the god returned to claim his payment.
"We have no money," rhe parcnrs said, "because we're old, bUI in
twelve years our children will pay you."
"Because you have= tricked me," replied the god, "I will kill
And he did.
After twelve years the god returned ro again ask the children for hiJ
payment.
"You've killed our parents," S3id the couple, "so the money we've
gathered up to pay you h3s all been spent. We have to ask you for ten
more years to acquit our debt."
Ten yrars later, the god returned 3nd the couple had three children
but no money.
"I will kill you," said the god, "you and your descendams, whc1her
you be old or young."
Since that day, humans have been mortal, 3nd when one quiu life.
Malagasy people say, "they are taken by the god that made 1hem:
(Rene! 1910 lll: 17-18; my translation from the French).
The mythological poim is, to say the least, suggestive. One might wdi
argue that this whole ani tude is ultimately one with the logic of s.1crifict.
which at least in Madagascar is often explicidy phrased as a way offobbing
off the Divine Powers wirh a portion of what is rightfully theirs, so as to wir.
the rest for living people. The life of the animal, it is often said, goes to GoJ;
hence (implicitly), we get to keep our own. Consider, then, the curious f.Kl
that all over Madagascar, sacrificial riruals-or their functional equi,-alrms..
such as thcfomadihana (reburial) riruals oflmerina-always seem ro rrquirt
govcrnmcnr permits. The facr rh:u this permir has been received. 1h.u !he-
paperwork has been properly done, is often made much of during rhr en-
c:mony itself. Here is a fragment of a Betsimisaraka speech, spoken O\'Cf the
body of a sacrificial ox:
PROVISIONAL ALnUNOMOUS ZONE
For this ox is not the kind of ox that lazes in its pen or shiu anywhnc
on entering the village. hs body is here with us, but its life is with you,
the government. You, the government, arc like a great bean lying on in
back: he who turns it over sees its huge jaws; so we, comrades, cannot
turn that beasr over! It is this official permit that is the knife that dartl
to cut its hide, the ax that dares to break its bones, which comtl from
you who hold political authority (Aiy 1994: 59-60).
169
Not only is che state figured simultaneously as a potential force of vio-
lence and its victim; the ace of acquiring a permit becomes equated with
the act of sacrifice irself. The main point lam trying to make hert is about
autonomy. Filling out forms, registering land, even paying taxes, might be
considered the equivalenrs of sacrifice: linle ritualized actions of propitiation
by which one wins the autonomy to cominue with one's life.
This theme of autonomy crops up in any number of other siUdics of
colonial and postcolonial Madagascar-notably, those of GerJid Althabc
(1969, 2000), about these same Bcrsimisaraka, and Gillian Fedcy-Hamik
(1982, 1984, 1991) on the Sakalava of the northwest coast. But in these au-
thors it takes on a sort of added twist, since both suggest that, in Madagascar,
rhe most common way to :1chievc autonomy is by creating a b.lse image of
domination. The logic seems to be this: a community of equals can only bt
created by common subordination to some overarching force. Typically. it is
conceived as arbitrary and potentially in much the same w;ay as thr
traditional Malagasy God. But it can also be equally far from everyday hu-
man concerns. One of the most dramatic responses to colonial rule:, among
both peoples, was the massive diffusion of spirit possession; inn-cry commu-
niry, women bc:gan to be possessed by the souls of ancient kings, "hose: will
was (at least in theory) ro have all the authority it would had
they been alive. By relegating ultimate social authority to cmr:&nccd womrn
1pcaking with the voices of dead kings, the power to constitute' communities
is displaced to a zone where French officials and policr would have no way to
openly confront it. In either case, there was the same kind of mme: one man-
ages to create a space for free action, in which to live one's life out of thr grip
of power, only by creating the image of absolute domination-but onrwhKh
U ultimately only that, an image, a phantasm. completely manipulabk by
those it ostensibly subjects.
To put the marter crudely, one might s.ay that thr people I knew -ne
engaged in a kind of scam. Their image of govc:rnmem had, at last I-ince thr
colonial period. been one of something essentially alien, prcdaro.-,.. concivt.
The: principal emotion it inspired was fear. Under the French. rhe pnn-
mem apparatus was primarily an engine for extracting mon<'y and forced
170
POSSIBILITIES
labor from irs subjects; ir provided rela[ively linle in way of social benchrs fo.;
rhe rural population (certainly, from rhe poinr of view of the rural popula-
tion ir didn'c). In so far as ir did concern itself wirh irs sub jeers' daily netd.J, ir
was wirh rhe conscious intention of creating new ones, of transforming rhti;
desires so as m create a more deeply roared dependence. Nor did mantn
change much after independence in 1960, since the first Malagasy regill'.t
made very few changes in irs policy or mode of operation. For rhe vast rru.-
joricy of the population, rhe common-sense attirude was that the srare W"'&J
something to be propitiated, then avoided, in so far as it was in any "''!
possible to do so.
Ic was only after the revolution of 1972 that things really began ro
change.
An anti-colonial revoh in its origins, the 1972 evems introduced a suc-
cession of state-capitalist, military-based regimes-from 1975 untill991,
dominated by the figure of President Didier Ratsiraka. Rarsiraka found hU
political inspiration in Kim II Sung of North Korea. In theory, his rc:gimt:
was dedicated to a very centralized version of socialist developmem and
mobilization. From the beginning, though, he was uninterested in what fit
considered a stagnant, traditional peasant sector with lit de revolmionary po-
temial. In agriculmre as in industry, his government concentrated its ellom
on a series of colossal development schemes, often heroic in scale, involving
massive investment, funded by foreign loans. Loans were easy enough to gn
in the 1970s. By 1981, the government was insolvent. Ever since, M a l a ~
economic history has mainly been the story of negotiations with the IMF.
There is no room here to enter imo details on the effects ofiMF-ordrrr.:l
austerity plans. Suffice it to say their immediate result was a catastrophic r ~ u
in living standards, across the board. Hardest hit were the civil service .1nd
other government employees (who made up the bulk of the middle cl.ml
bur-aside from a narrow elite surrounding the President himself. who !tok
libcrally-pauperi7.ation has been well-nigh universal. Madagascar is now
one of the poorest coumries on earth.
For Ratsiraka's "peasant sector"-rural areas nor producing key com-
modities-this whole period was marked by rhe gradual withdrawAl of tilt
state. The most onerous taxes from the French period-the head rn. cJnk
tax, house tax-intended to force farmers to sell their products and 1hw
10 goad them into the cash economy, were abolished immediatd)' after tlx
revolution. Ratsiraka's regime first ignored rural administration; after 1981.
it increasingly became the object of triage. The srate, its resources ewr more
limited as budgets were endlessly slashed, was reduced to administering .1nJ
providing minimal social services to those rowns and territories its rukr.
found economically important: mainly, those which generated sdmt kinJ
PROVISIONAL AVTONOMOUS ZONE
171
of or.change. Places like Arivonimamo, where almost all production
and distribution was carried our outside rhe formal sector wac of
no imert:St to them. Indeed, it is hard to imagine anything that could hap
pen of rhe area becoming the base for armed guerrillu (hardly
a possibiliry)-that would seriously threaten the interests of rhc mrn who
tta.lly ran the country.
2
Resources for rural areas dried up. By the time I was in Arivonimamo,
the: only sector of administration that was receiving any signi6canr funding
\\'2.5 the: education system. Even here the sums were paltry: the: main govern-
ment role wa.s to post the reachers (who were sometimes paid, at lcm in part,
by parents' associations), provide curricula, and administer the tests. The lat-
rc:r, particularly the baccalaureate examination, were of particul.ar concc:rn to
the: center because they were the gateway into the formal, state: sector: those
who passed their baccalaureaTe were obliged to undergo several wttb of
military training and then carry out a year's .. National Service," though-as
I've: pointed out-this mainly consisted of lounging around in meaningless
make-work jobs. But National Service was, I rhink, important. It was
of marking passage into a domain where effective authority really did aist,
Yo-here orders had to be obeyed. For those not ensconced in the educational
system, the government provided nothing, but it also had almost no immtd.i-
atc: power over their lives. j
Still, even in the countryside, government offices to exist. The:
typewriters were often crumbling, functionaries were ofte-n reduced ro bur-
ing the-ir own paper, since they could no longer requisition any, but people
dutifully continued to 611 out forms, requesting permission uprooting
trt'CS or elthuming the dead, reponing binhs and deaths. and regism-ing the
number of their cattle. They must have re-alized that, had they rcfwa.i, noth-
ing ""uld have happened. So: why did they play along1
One mighc, I suppose, call it inertia, sheer force of habit: people- were:
nill running the same scam, propitiating the state without having noticed
iu huge: jaws were toothless. Cenainly, memories of colonial violence: mrc
11ill vivid. I was told many times of the early days of ma!l C"'c-cutions. or of
bow tcrri6ed rural people. used to be when had ro eru" a government
Clfficc:, of the endless pressure of taxation. Burl think the: rc:al answtr is more:
IUbtle.
Mmtoric::s of violence were: mainly important they defined
what people imagined a state to be about. I found little- notion that t),( state
!for all iu socialist pretensions) cxistcd to provide services; at least, no one:
much complained about the lack of them. Pe-ople srcmcd to accc:pt 1h.at a
gortmment essentially an arbitrary. predatory, coercive pawt"f. But the
one- thone of of6cial ideology everyone- did to take sniousJy w:u the
idea of Malagasy uniry. In rhe highlands, at least, people saw rhemsd\esl
1
"Malagasy"; they hardly ever referred to themselves as "Merina."

uniry was a constant theme in rhetoric; it was the real meaning, I thin(
of rhe Malagasy Rags that inevitably accompanied any major ri[Ual (whOSe
official meaning was ro mark that the forms had been filled out, the e>.err.
approved). lr seems to me that ir was the very empriness of the state which
made it acceprable as a unifying force. When it was powerful, the
1
n
Imerina was essentially seen as something French-this remained true eo.rn
in the early years of independence. The 1972 revolution was lirsr and fore-
most an efforc ro achieve genuine independence, ro make rhe state truh
Malagasy. For the highland popularion, I would say, rhis effort was
successful-if only because, at the same time, rhe stare was stripped of
mosr all effective power. In other words, the governmenr became something
along rhe same lines as the ancient kings discussed by Althabe and Feeley-
Harnik: absolute, arbitrary powers that consritute those they subjugate 3.5 2
community by vircue of their common subjugation, while at the same time,
exrremely convenienr powers to be ruled by, because, in any immediate
tical sense, they do nor exist.
Provisional Aulonomous Zone
In contemporary anarchist circles it has become common to talk of thr
"TAZ," or "temporary autonomous zones" (Bey 1991 ). The idea is that, whilr
there may no longer be any place on earrh cnrircly uncolonized br
and Capital, power is not completely monolithic: rhere are always rcmpour_.
cracks and fissures, ephemeral spaces in which communitici
can and do cominually emerge like eruptions, covert uprisings. Free spms
Ricker into existence and rhen pass away. If nothing else, they provid('(on
Stant testimony to the fact rhar alrernativcs are still conceivable, that hum&n
possibiliries are never fixed.
In rural lmerina, it mighr be better to talk about a "provisional auronc--
mous zone," rather than a "temporary" one: in part, ro emphasize that it
not stand quite so ddlanrly outside power as the image of a TAZ impliM.
but also, because there is no reason ro necessarily assume i[S indeptndrtM
is all that temporary. BerafO, even to a large cxrcnt Arivoninumo.
outside the direct control of the state apparatus: even if the proplc who]i,-c
there passed back and forth between them and zones, such as tht" cJpiti
which arc very much under the domination of rhe state. Their autonomy"'ll
tentative, uncertain. Ir might be largely swept awJ.)' rhe moment .1 new
sion of guns and money restores the appar:Jrus; but then again. it might n..'l
Some might consider the current situation Mrselt: I consitkr JiJ
remarkable accomplishment. After all, austerity plans have been imposed on
nations all over the world; few governments have reacted by abandoning the
bulk of the population to govern themselves; nor would many populations
have been so well prepared to do so.
Why were they able co do so? I would guess there are various reasons.
One is the maintenance of active traditions of self-governance, and what
would, ifir were observed in, say, European or Larin American social move-
mrnrs, undoubtedly be called a culture of direct democracy. The art of com-
ing to decisions by consensus was something everyone simply learned as part
of growing up. It was so much a part of everyday common sense that it was
difficult, at first, for an outsider to even notice it. For instance, there was a
general principle that no course of acdon that might have negative conse-
quences on others should legitimately be carried our without chose others'
prior consent; the resultant meetings were called "fokon'olona" meetings-
meaning, basically, "everybody"-bur despite rhe consistent misunderstand-
ing of colonial ethnography, "the" fokon'olona was nor a formal institution,
but a Aexible principle of deliberation by groups char could vary from five
to a thousand, depending on the dimensions of rhe problem they were col-
ltctivdy trying to solve. Within those meetings, however, anyone, male or
female, old or young, formally had equal right to speak: the only criteria
was to be old enough to be able to formulate an intelligent opinion.
4
What's
more, anyone engaged in an ongoing project had the power to engage in
what would in contemporary consensus process be referred to as a "block":
onr could simply declare "I am no longer in agreement" (t.ry manaiky aho)
'>''ith the general direction of rhings, and it would cause a general crisis until
one'sconcerns had been publicly addressed. Suffice it to say, then, that even
during the colonial period, when all political gatherings were technically il-
legal. ordinary people had maintained instirutional suuctures and political
habits that allowed them to govern their own affairs with minimal appeal
to ourside force. They had also managed to develop forms of resistance suf-
hciently subtle that, when the state was emptied of its substance, they were
.1ble to allow it to effectively collapse with minimal loss of face.
I don't mean to romancicize the situation. What autonomy rural com-
munities h.ave has been won at the cost of grinding poverty; it is hard to
enjoy one's freedom if one is in a constant scramble to have enough meat.
lnsrirurions of rule-mosr obviously schools and Christian churches-still
functioned, and in the same hierarchical way as ever, even if they did now
Lugdy l.acked the power to back up their efforts with the threat of physi-
c.aJ force. There were certainly profound social inequalities within many of
tksr rur.al communities, not to mention in rown: both differences of wealth
;perfups minor by world standards, bur nonetheless reaD. and even more.
174
POSSIBILITIES
divisions between what were called "whire" and "black" people, dcscendan
11
of nobles or commoners in the anciem kingdom, and rheir former slavC"!.]o
order ro what places like Berafo were like, rhen, one must fin:
rhar It was a srood our side srare power; then, that r.
d1d nor stand enmely ours1de lt. For all the effons to maimain ,
autonomy, reality of coercion has by now reshaped the terms by
people deal With each or her; in cercain ways, it has become embedded in th
very scrucrure of experience.
In Imerina, just about everyone considers themselves a Christian (about
two thirds of the population is Protestant, one third Catholic). Many
ularly anend church. The governmem may no longer have the meam
10
compel children to anend school, bm accendance is still close to uni\'try[
at least on the primary level. At the same rime, however, there is a
ambivalence about both these institutions, particularly rhe schools. As I..J.
ready remarked when speaking of the politics of research, the educational
system in Imerina has always been seen as a tool of power, and always, too,
idemified with Vazaha. The present educational system rook form
French colonial regime. lr is important to bear in mind rhar this was not._
regime that could ever make rhe most remotest claim to being the cxprosion
or popular will. It was a regime imposed by conquest, maintained onl)
constant threat of force.
lr is worth considering for a moment what maintaining a credible thrnr
or force actually requires. Ir is not merely a maner of having an
number of men willing to use violence; not even a maner of arming
training them. Mostly. iris a matter of coordination. The crucial thing isw
be able to ensure that a sufficiem number of such violent men will alw.i.\'"
be able to show up, whenever and wherever there is an open challenge
one's authority-and that everyone knows rhat they will indeed do so. Be:
this, in turn, requires a great deal. It requires an nrensive cadre or tr.iir..
functionaries capable of processing information, not to mcnrion an infn
mucrurc or roads, telephones, typewriters. barracks, repair shops. pcrrokul:l
depots-and rhe srafr {0 maintain rhern. Once buih, such an infr..tsrrur.:turt
can and doubtless will serve other purposes as well. Roads builr to
soldiers will also end up carrying chickens to marker and people ro \'isit rhcr
ailing relatives. But, if it wasn'r lOr the soldiers, rhe would nevrr
bn rhrre, and at lr;&st in people set."med pert(-ctly wdl J.W'lrt
of that.
Mosr ur rhe people who work in a state burt'aucracy-prc=ny much .1m
state on a day to day level, mtJCh mort (lYl
cerncd wt.th processing. informacion than with people's skulls. &t
thr samr ts uur of soldtcrs ;a.nJ police. R . .uher rh.
111
stt this 1;, ....
1
JS pnXlt'th..:
[IROVISIONAL AUTONOMOUS ZONE 171
violence plays a minor role in the operation of a state, it might bmtr to
ask oneself how much these technologies of information are themsdvo pan
of the apparatus of violence, essential elements in ensuring that small Nnd
ful of people willing and able to break skulls will alw.ays be abk to shD'N up .u
the right place at the right time. Surveillance, after all, is a techniqur of w.r,
and Foucault's Panopticon was a prison, with armed guards.
Viewed from Madagascar, the essentially violent of tht state is
much harder to deny. This w:as not only because ofirs colonial history. h was
also because most Malagasy-at least the ones I knew-were accustomed ro
different standards of perception. The best way to put it is that, unlike most
Americans, they did not see anything parcicularly shameful about ftar. This
was one of the things it took me longesc to get used to Sing grown
men, for instance, gazing into the street and casually remarking ""sa.ry can,
''I'm scared of those oxen." For someone brought up as I had hem ir was my
disconcerting. I may not come from a panicularly macho background. by
American standards, but I had been brought up to assume conftssioru ol
ftar, at least fear of being physically harmed by or hers, at least 3 link
. bit Cmbarrassing. Most Malagasy sttmed to find the subject pleas:am and
amusing; they took a veritable delight in relling me how afr.aid some pcopk
wtre of Vauha, sometimes, even, how much they themselves wnt>. Tha1
governments work largely through inspiring k.a.r in their subjects was simpl}
obvious to them. It seems to me char, in so far as Western science
a .tendency to downplay the imponance of coercion, ir is partly because of 01
hadden embarrassmem: we find it shameful ro admit the dcgm: to which OUI
own daily lives arc framed by the fear of physical force.,
Schools, anyway, are ultimarely a p:m of this appuarw of violence.
In Malagasy, one does nor speak of educ..rion as convt'ying facu and in
formation so much as skills: the word used,fohaiutW, meant .. skill. know
how, practical knowledge." The kind of fohaiuM one acquirn 011 w.:hool
howt"Ver w.u seen as an essentially foreign one, a fohalurw Vu.ah.l. oppostd.
u such, to Malagasy forms of know-how. The rrchniquts lauttht in w:hool
wen seen a.s, esscmially, techniques of rule. In pan rhi.s i1 bttaut.t 1ht Khool
system was itself pan of the infrastructure of violcnn: it w015 dnittncd. pri
m.arily to train funcrionaries: xcondarily. technit:ians. The uyk of ttx-hinJ
wu amhoritarian, with a heavy emphalis on rotc mcmonurion .and
the skills th.at were taught were taught wirh 1he npnuuon thty .-m- Ill
be cmploytd in officn, workshops. or danroom1 orttaniud around c.ena1n
forms of social relation-what mi(lbt be rd"crrrd 1o as rtl.atloru of comnund
lnc assumption wu aiWOif' th.at peopk would br otdcn. 01h
m thc'fC to obq. In mtK-r words, not only was rhu. dnaprd k
produce rM comprtenco 10 maimain an infwtructul't ol.......,t
176
POSSIBJLrJlES
it was premised on social relations com pie rely unlike rhose current i
aspects of daily life, ones rhar could only be maimained by a
of physical harm. e.:
The rowards and book learning, then,
on a perfec_rly_ sens1ble apprectanon of rhe situation. Everyone comida:.;:
knowledge_ m nself a valuable, even a pleasant, rhing. Everyone recogni:ztt
thar the sktlls one learned in school opened spheres of experience rhat
nor mherwise be available, co types of informacion and nerworks of
municarion that spanned the globe. Bur rhese skills were also rechniqut1 J
1
repression. By training people in cenain mer hods of organization and!)):
others (how to keep lisrs and inventories, how ro conduct a mee1ing. .. ). ti-t
sysrem ensured chat no maner whar their purposes, any large-scale nrn'O:l
they put together capable of coordinating anyrhing-wherher it be an ..
rorical preservation society, or a revolutionary parry-will almosl incviubfr
end up operating somewhat like a coercive bureaucracy. Cenainly, one con.
and many did, cry to rework these devices ro operate in a more consenswl.
democratic manner. Jr can be done, bur ir is extrcmdy dif6cuh;
tendency, the drifr, is for any system creared by people trained in these com-
petencies, no maner how revolutionary their inrenrions, ro end up
ar least a litde like rhe French colonial regime. Hardly surprising then th.u
most people wrote rhese techniques off as inherently foreign. and uicd l..i
much as rhey could ro isolate them from "Malagasy" conrcxts.
But, ar the same rime, there was another, perhaps more subtle effect ri
the existence of these hierarchical insrirurions. They allowed people to mal:e
clear disrincrions between everything rhat was ''gasy"-Malagasy-lr...i
everything rhar was considered "Vazaha," alien, authoritarian,
French. They guaranteed that everyone had ar least some experience of tb::
latter, rhar zone where the stare was "rhe only competent as well as 1he !Ak
owner of social coercive force": even ifir was simply a matter of being foo:li
to stand in uncomforrablc lines as a child, jump ar orders in gym class, .tnJ
dutifully copy and memorize boring and apparently pointless lessons. Th
Ciperience of smtclike discipline became a way of constantly remindin!!orr.:-
self what was, in conuasr, considered "Malagasy"-rhe habits of constns:!>
decision-making, for example, rhe reluctance ro give orders to ftllow .a;duh."
the general suspicion of anyrhing thar smacked of confronra1ion or C"o"tr.
charismatic leadership (compare Bloch 1971). h is fairlv clear that
these 1rairs had nm always been considered Malag.a.sy.
though I susptct thai Malagasy had, from rhe very beginning of their stttk-
the island, always tended ro define themselves againsr foreigtl(M l'i
some kmd or another. c. In this way, puadoxically enough. the
natun: of local autonomy actually becomes, in a sense, sdf-susraining. \fc
PROVISIONAL AtrrONOMOUS ZONE
IT7
all live in a larger world of gross of weahh and power. Malagasy
rice farmers and blacksmiths and seamnresses and vidro opcntoD were all
well aware of rhar. Bur precisely through such constant reminders, pcoplr
managed, co a large degree, to insulate themselves as wdl.
A Final Q.uestlon
I doubt rhat the himerland of Arivonimamo is an isol:ued case. AJ
Henry Wrighr had pointed our to me, similar rhings were happening all
over Madagascar: in Facr, probably they had bn for much longer and in
more profound ways in many other partS of the island, since Arivonimamo
was, after all, with its milirary airport and gendarmes and prison, an hour
away from the capital, one of the last places one would the sr.ate au
thority to disappear. In Madagascar itself, state authoricy appe:m to havt
ebbed and flowed, sometimes assening irself, sometimes retreating, in the
intervening years; bur in much of the counny-panicularly areas like
Arivonimamo, do not comain vanilla plantations, bauxite mines, or n.uun:
preserves-the situation has remained essentially unchanged. One wooden
if then: mighr not be hundreds, even thousands, of similar communitia in
other pans of the world-communities that have from or drifta.l
away from the effective conuol of national governments and become to all
intents and purposes self-governing, but whose members arc sci II performing
the external form and tokens of obeisance in order to disgui.se th.:u f.lc1.
his a question we might well ponder when reading the conremponry lit
er.&ture on "Failed states" arid particular, the crisis of uareauthority in Afria.
As James Ferguson has rccendy nored (2006), in many pam of Africa, about
the only significant meaning of ,.Slate sovereignty" left is international rec
ognition of a government's legal right to represent iu citizens in intcrna
tiona) arenas, and particularly, to guarantee contracts concerning access to
within irs territory, for those from other states. Few cvtn prrtend
to maintain a monopoly of violence in the manner Jesc::riba.l by Rudolph
von lhering or Max Weber. The withdrawal of resources, the ab.andonmctu
of any sense that the government can or would C'\'C'n wish to pfO\'iJe rqw.lly
for the basic needs of all its citizen!>, ha.s had deva.st.:uinp; dTC'Cu on he-alth.
educarion, and livelihood. But at the same time. cvm IMF-impos.rd auneri1y
plans have known ro have their curious unintended side-dTecn.
It is, in fact, something of an irony that it is only when "an.an.hy: in
the sense of the breakdown of stare power, resuhs in chaos, vioknce .. md
destruction-as in the case of s.ay, Somalia in 1he 1990s. or many oi
southern and cenrral Africa today-that non-Africaru an- lilcd,. ro hrar
abour ir. Whar I observed in MadaK"'5Car suggcus that for evny lUC.

POSSIBILITIES
there might well be dozens, even hundreds that oursiders simply d
know about, precisely because local people managed to make rhc u
0

111
peacefully. Like Malagasy villagers, they avoided confrontation
.
state had feel publicly humiliated or to lose
ar the same.tlme, It as d1fficult possible for them ro gmern, .
easy as poss1ble ro s1mply play along With the Neither is rhi
51
t
1
,,
egy, the existence of newly autonomous communities, likely to be
to Afnca. There are many pans of rhe world-in sourheasr Asia, Oua,-d
1
most nmably, bur even, say. parts of Larin America-where rhe presenar;
the stare has always been a somewhat sporadic phenomenon. hs visitdJtot
perhaps, always borne less resemblance to the forms of consant
and surveillance we are familiar with in both totalitarian srates or mdu\:
trial democracies, and more rhe occasional, if often dis:mrous, apparan::
of a vindictive Malagasy god. So, often, with the worldsystem as a
Such gods can rarely be eliminated entirely, any more than the monsoorucf
earthquakes that they are often seen to resemble. Bur rheir visitations
rendered equally occasional.
Of course, rhe insrirurional srrucrure did remain: rhere were S<hook
banks, hospitals. They ensured that the "stare form", as Mario Tronri 10:
insrance calls ir, was always prescm: everyone had some idea wh3t it
to live inside insrirutions that were premised on coercion, even iff or thr mo<t
parr these were ghostly shadows of real stare institutions, since rhr actual \iG-
Ience had been stripped away. Or perhaps one should be more precisr hm
The violence was sri II there. Ir had simply retreated. There were cenainlyni!'
police in rhe city, or anywhere where there was, say, a bauxire mine. orollxr
resource rhar generated significant foreign exchange. Even more, rhe glolu!
allocation of resources-what medicines and equipment acrually appraml
in rhe local hospital, for example-was maintained by rhe sysrema1ic
of violence ro enforce propeny arrangements. In a place like Arivonimamo
however, one could only deal with its distant effecrs, and s1range. hollo11
insrirutions that largely served to remind local people of precisely how 1he-
were nor supposed ro ordinarily behave.
Endnotes
One mighr cumras1 rhc siiU:uion here wir h whar ob1ains in, say. much oi
che siiUalion is quire rhe opposite, sinLe police. cfl'ccri\dy. our or-"r
mccrcsced m enforcing proprny righiS, and c:u
1
he
10
ignore men: cnt'
of murdrr-unleu, cha1 rhe viccim is a member of chc- dilr
PROVISIONAL AliTONOMOUS ZONE
179
2 The gendarmes' occasional zeal in pursuing bandits probably did have wmething
to do with a perception that they were the only organizW, armed group rhat lud
the capacity to form the nucleus of a rebellion-unlikely though that might havr
been. There had been times, mainly in the nineteenth century, when bandiuactu
ally had turned into rebels. But I suspect the concern was rooted in deeper under-
standings about what a state was all about: under the Merina Kingdom. bandiu
(referred to in official documents simply as fllhavakJ, encm{) were, along
with witches, the archetypal anti-stare, rhat which legitimate royal aUihority de-
fined itself against. The connection with witches also helps explain the orhcrwUc
puzzling fact that, much though they were unconcerned with Henri's deprcda
tions, Arivonimamo's gendarmes did leap into action to arn:st and a
re-cnagc girl suspected of being behind an outbrnk or AmbtJ/awLJM, or pce:Wion
by evil ghosu, which affected a whole dormful of studenu at the stare high school
in 1979.
Medical services for instance were in thtoryprovidcd fm, but had bcm dTectrvcly
privatized by corruption, which, in turn, became universal once gavernment sab
rics declined to next to nothing.
As Jacques Dcz (1975: 54-57) notes in a genen.lly cxccllcnr summary: though in
rhe end, he reproduces colonial assumptions by concluding that
wu by the late-cightcc:nth century king Andrianampoinimcrina. On
the underlying ethos of consensus decision-making sec Andriaman1ato (1957).
5 In Europe or North America, this is more true of men than women: in Mad..:aga1ear
il w:tS, if anything, the other way around.
6 Conrempor.uy archeologists now believe th:u significant human in
Madagascar wu surprisingly l:ue: perh2ps from the eighth century CE, and ar fint
seem to h111vt consisted ofheterogenrous popul111tions probably of very different ori
gins, Austronesian, African, and perhaps others. Durins this early period then:- W21
even a smalllslillmic city, Mahilaka, almost cerrainly mpgcd
in lively trade with East Africa 111nd the Arabian ptnrnsul.a.. E.ariy Malillgasy rhu
had experience of Slilltes and world religions from the very beginning; and the: mo-
ment of when contemporary Mabg:uy cuhun- appnn to havt born,
SC"Cms to have occurred around the 1ime of the height or perhaps even
ofMillhilillkill. After this, howC"Ver, it proved surprisins)y pcni1tcnt rhrnugMultbc
i1land and o[ resisting Crequem Islamic ancmpu to convert and incorpo-
r.ue the island's popul111tion. I strongly swpcct th.u insofar as Malapl,
emerged as a coherent entity. it w:u in consciow. contr:ut to everything thai ....-u
considered "SiiAmo" -Swahili. lslamic-jwt u it is maintained in con
1ns1 to cverything;thilll i1 todilly.
IHU
POSSIBILITIES
Bibliography
Althabc, Gl!rard
1969 OpprtSJion rt Lib/ration tid11s 1'/maginnirr: Irs oommufiiJUtb ,b {.,,.
on'mt.1U dr Mad4gascar. Paris: M:upero.

2000 Anthropologir politiqur d'unr drcoloniZBtion. Paris: l'H:armanan.
Aly, Jacques
1984 '"Lc Discours Rilud chez lcs Bc(Simisaraka de Ia Clhc ES[ de M:adagucar." Prtltr'..:r
Afri<ainr 132: 54-61.
Andriamanjato, Richard
1957 Lr Tliny n lr Tody d4nr 14 prmlr Malgachr. Paris: Presence Africanc.
Bey, Hakim
1991 T.A.Z. T/Jr Tunpornry Autonomow Zonr, Onto/og,(a/ Antmhy, PomC
New York: Autonomcdia.
Bloch, Maurice
1971 '"Decision-making in Councils Among the Mcrina" In Councils in 11nion (AwJe;
Richards and Adam Kuper, cds.). Cambridge Papers in Social
Cambridge: Cambridge University Press.
On, Jacques
1975 '"Prrmitrt structure d'cncadrcmcnr rural: lc Fokonolona."
6: 31-69.
Fcdcy-Harnik, Gillian
1982 "The King's Men in Madagascar: Slavery, Ciriunship and MoJUrd:y:
Africa 52: 31-50.
1984 .. The Political Economy of De:nh: Communicadon and Change in MaUgUT
Colonial History." Amt'rioan Etlmalogist 8: 231-254.
1991 A Gum Estau: Rmoring l11dt'pmdmu in Mndagamu. Washington:
Ferguson, James
2006 GltJbal Shadows: Africa;, tht' N,a/ibmtl WarM Ordt'r. Duke: IJnivmifl
PrcJS.
Urson, Pier
2000 History and Mt'mory in tilt Agt af Em/4wmmt: Buaming Mtrina i11 H,f!Uv
MadagiUcar, 1770-1822 (Social History of Africa). Ponsmourh: Hcmc:mann_
Lnel, Charles
1910 Conm dt MadagiUcllr (3 volumes). IJ:&ris: Ernest Leroux.
Turner, Stephen, and Regi Factor
1994 M.u \\'ltbtr: Tht as Soria/ 71Jinlur. i.onJon: Roulledge.
Weber, Mia
1968 An Outlillt of hutrpmil't" .\lh-iDlog. 2 \'Ob. (Guenther
ant! Wuru:h. et!s.; Ephraim FilhoiT. New York: BNminuN