Great Zimbabwe

Houses of stone

Drafted from Pikirayi, African Archaeology Series (2001)formation

On a granite plateau in southern Zimbabwe there is a wall. It is about two hundred and fifty metres round, up to eleven metres high and five metres thick at the base, built of shaped granite blocks laid in courses without mortar, with a chevron pattern running along the top of one stretch. Inside it there is a solid conical tower, ten metres high, that is not a granary and is not a tomb and has no known function.

Around and above it are further enclosures on the hill, and a valley of smaller ones. The whole complex covers something like seven hundred hectares and is the largest stone structure in Africa south of the pyramids.

Nobody who built it wrote anything down.

Everything here comes from the ground

There is no chronicle, no inscription, no king list and no dated event. The occupation is known from radiocarbon and from imported ceramics; the economy from gold workings and from the direction the imports came from; the phasing from how well the walls are coursed. No ruler of this state is known — not a name, not a title, not a length of reign — because there is nobody who wrote one down.

That is worth pausing on, because it is a different kind of silence from the usual one. Most states that seem undocumented are documented by their neighbours, or by their successors, or by an archive that partly burned. This one had no literate neighbour within three thousand kilometres and no successor who wrote. What is left is the stone and the rubbish.

The gold, and the pots

The one thing that can be said with confidence about how this state paid for itself is said by other people's imports.

Excavation has produced Chinese celadon and stoneware, Persian and Syrian glazed wares, glass beads by the tens of thousands, and a coin struck for a thirteenth-century sultan of Kilwa.

Kilwa is three thousand kilometres up the coast, on an island off what is now Tanzania, and its whole position rested on the monsoon: a ship out of Arabia could reach Kilwa and return inside one year's winds and could not reach Sofala and return, so the last sixteen hundred kilometres to the gold belonged to Kilwan boats. The gold in that sentence came off this plateau.

A copper coin in a museum case is the physical proof of a relationship between two places that never exchanged an ambassador, a treaty or a word of writing — only metal, cloth and pots.

Why it stopped

Probably not one thing, and the fifth chapter is about it: land exhausted after three centuries of concentration, a gold trade whose centre moved north towards the Zambezi, and possibly climate.

The site was not destroyed. It was left, over some decades in the fifteenth century, and people went on living in the area and building smaller versions of the same architecture elsewhere — Khami, Danangombe, and something over a hundred and fifty other zimbabwes across the plateau.

Nobody conquered it. There was nobody to.

How the gold was got

Drafted from Theal, Records of South-Eastern Africa, volume VI (1900)expansion

The plateau's gold is the reason there was anything here to build with, and nobody described the working of it until the sixteenth century, by which time the capital had been empty for a hundred years. João de Barros, printing in 1552 from reports sent home by men at Sofala, is the first written account of how it was actually done — and he says of the fields he is describing that they are the most ancient known in the country.

Water, and two hundred men

There is no deep mining in it. All the gold in the Manica field, fifty leagues west of Sofala, is dust, and the difficulty is that the land is dry: the earth has to be carried to water before it can be washed, so they dig hollows that fill in the winter, generally six or seven spans deep, because at twenty spans you hit rock.

Further inland, a hundred to two hundred leagues from the coast, the specks are larger — some in veins in the rock, some cleaned out by the winter torrents. There, Barros says, they dive in the still pools the rivers leave in summer and wash the mud they bring up. Where there are ponds, two hundred men gather, drain away half the water, and go through what is left.

That is placer work at scale: no shafts, no timbering, no pumps, and a labour force organised in hundreds for a season. It explains a great deal about the kind of state that grew on top of it. What such a polity has to be able to do is assemble people at the right time of year and take a share of what they bring up. It does not need an army, and there is no sign this one had one.

The credit trap

The most revealing paragraph in Barros is about how the coast got the metal out of the interior, and it is not by force.

He says the country was rich enough that a great quantity could be had if the people were covetous, but that they wanted gold so little that a man had to be very hungry before he would dig for it. So the Muslim traders who worked the interior used artifice to raise their appetite. They dressed the men and their wives in cloth, beads and trinkets until they were delighted with them; then gave them the whole lot on credit, and told them to go and dig, and if they came back within a certain time they could pay for it.

And by thus trusting them they oblige them to dig, and they are so trustworthy that they keep their word.

Read that from the plateau's side and the trade looks different from the way it usually does. The imports found in the excavations — the celadon, the glazed Persian wares, the beads by the tens of thousands — are not the proceeds of a gold trade. They are the advance against it. Someone handed them over first and came back for the metal.

What the state took

Barros does not say what the ruler's share was, and no source does. What the ground shows is where it accumulated: at one place, on one granite hill, for about three centuries.

A trade of this shape — seasonal, dispersed, run on credit extended by outsiders, in a commodity the diggers did not particularly want — puts the power with whoever can guarantee the outsiders access and guarantee the diggers their side of the bargain. That is a job for a court, not a garrison. The enclosures on this hill are the physical form of it: not walls to hold an army, but walls to say where the authority is.

And where it went out

East to Sofala, and from Sofala north on the monsoon. Barbosa, writing a generation later, says that before the Portuguese came the Muslim towns of Sofala, Cuama, Angoya and Mozambique were all subject to the king of Kilwa, who was the mightiest of them, and that there was great plenty of gold in that town because no ship went towards Sofala without first coming to that island.

That sentence is the far end of this one. The gold in it came off this plateau, was washed out of ponds by men in gangs of two hundred, and was paid for in advance with cloth.

The stone

Drafted from Pikirayi, African Archaeology Series (2001)peak

The granite of the Zimbabwe plateau exfoliates. Heated by the sun and cooled at night over long periods, it splits along planes parallel to the surface and sheds slabs, and a slab can be broken into blocks of roughly even thickness with a hammerstone.

That is the whole technical basis of the architecture. The builders did not quarry in the sense of cutting into a face; they collected what the rock was already producing, dressed it, and laid it in courses.

There is no mortar anywhere in the complex. Nothing holds the walls up but the weight of the stone and the accuracy of the coursing.

What that constrains

Everything about the architecture follows from a wall that cannot be tied together.

The walls are massive because they have to be — up to five metres thick at the base and battered inwards as they rise. There are no arches and no spanned openings of any width: entrances are narrow, and the ones that survive are either lintelled with short stone slabs or stepped. There are no upper floors of stone. The buildings people actually lived in were dhaka — puddled clay and gravel, thatched — set inside the stone enclosures rather than built of them.

So the walls are not buildings and were never roofs. They are enclosures: they mark, screen and separate, and what they contain is a set of clay houses and the space between them.

That is a different thing from the European castle a Victorian visitor saw, and correcting the misreading is part of what the fourth chapter is about.

The courses, and the chevron

The quality of the coursing varies systematically, and the archaeology uses it to phase the site.

The earliest work is rough, uncoursed and irregular. The best — on the outer wall of the Great Enclosure — is regular, level, and laid with a consistency that takes practice to achieve and a labour force to sustain. On one long stretch of that wall, near the top, there is a band of chevron decoration a dozen courses deep, made by laying blocks at alternating angles.

Nobody knows what the chevron means. It appears elsewhere in the Zimbabwe culture, and the readings offered — a serpent, water, fertility, a boundary marker — are guesses of the kind that are made when a symbol survives and its users do not.

The conical tower

Inside the Great Enclosure there is a solid stone cone, about ten metres high and five across at the base, tapering upwards.

It is not hollow. It has no entrance, no chamber, no inscription and no associated deposit. It is not a granary — the granaries on the site are clay and elsewhere. It is not a tomb; nothing has been found in it and it is solid.

The suggestions are that it represents a grain bin as a symbol of royal provision, or a phallic symbol, or a marker of authority with no specific referent. Pikirayi is candid that these are inferences from ethnography rather than readings of evidence.

A solid ten-metre tower that nobody can explain is a fair emblem for the whole place: unmistakably deliberate, unmistakably important, and mute.

And a hundred and fifty more

Great Zimbabwe is the largest of the stone enclosures on this plateau and it is not the only one. There are something over a hundred and fifty zimbabwes across Zimbabwe, Mozambique and Botswana — Khami, Danangombe, Naletale, Manyikeni and the rest — built in the same tradition, some before this one, most after it.

That is what archaeologists mean by the Zimbabwe culture, and it is the right unit. A tradition of six hundred years with many centres is what the evidence actually describes; one capital, however large, is a slice through it. The Swahili coast four hundred miles east has exactly the same shape and causes exactly the same trouble: a civilisation of thirty towns, and a habit of telling its story through whichever one was richest.

The rhinoceros on the hill

Drafted from Pikirayi, African Archaeology Series (2001)aside?What aside means

Great Zimbabwe did not begin from nothing around 1100. Two hundred kilometres to the south, at the confluence of the Limpopo and the Shashe, there was a capital doing the same things a century earlier, and what happened to its excavated grave goods in the twentieth century is a smaller version of what happened to the ruins north of the river.

Mapungubwe

From about 1075 to 1220 there was a capital on a sandstone hill above the Limpopo. It has terracing, a royal residence on the summit and commoner settlement below, imported glass beads and Chinese celadon, and gold.

The graves on the hilltop were excavated in 1932. One held a burial with a sceptre, a bowl, and a rhinoceros of gold foil tacked over a wooden core, about fifteen centimetres long. The rhinoceros is now one of South Africa's national symbols. For decades it was kept largely out of public view by governments whose account of the country's past it contradicted — the same treatment, in a different country at the same time, that was being given to the ruins north of the Limpopo.

Mapungubwe declines as Great Zimbabwe rises, and the conventional reading is a shift of the gold trade's centre of gravity northward. Whether the second grew out of the first, replaced it, or simply outlasted a climate shift is argued.

Why it matters that it came first

If the plateau's political tradition begins at Mapungubwe around 1075, then Great Zimbabwe is not a beginning and not a peak. It is the middle three hundred and fifty years of something that ran from the eleventh century into the nineteenth, moving its centre several times — south of the Limpopo, then the granite plateau, then the Zambezi escarpment and the western hills.

A reader who arrives at the ruins and takes them for the whole story has the shape wrong in a specific way. Nothing stopped here. A court was left, and the tradition moved, and it had already moved once before to get here.

Who built it

Drafted from Pikirayi, African Archaeology Series (2001)aside?What aside means

For most of a century the question of who built Great Zimbabwe was answered by people who had decided the answer in advance.

The sequence

Karl Mauch reached the site in 1871 and concluded it was built by Phoenicians — the Queen of Sheba's palace, on the reasoning that a beam of cedar he took to be Lebanese could not be local. It was local sandalwood.

Theodore Bent dug the site in 1891 for the British South Africa Company, surveyed it accurately, published good plans, and concluded that the builders were Arabians and that Africans could have had no part in it except as slaves.

The same company funded excavation by Richard Hall in 1902. Hall cleared the deposits down to bedrock across large areas of the site, searching for the ancient foreign layer and destroying the stratigraphy that would have dated what was actually there. He published his conclusion that it was Phoenician or Sabaean.

David Randall-MacIver excavated in 1905 and said it was African and medieval. Gertrude Caton-Thompson excavated in 1929, using a stratigraphic method Hall had not used and which could no longer be used in the areas he had cleared, and said the same: African, medieval, built by the ancestors of the people living there.

She was right, and the professional argument was substantially over from that point. The political one was not.

What was done with it afterwards

The Rhodesian government of the 1960s and 1970s required that official guides, museum displays and publications present the site's origins as unproven or foreign. Archaeologists working in the country were placed under pressure not to publish the African attribution. Peter Garlake, the country's senior inspector of monuments, left the post and the country over it.

The state that did this took its name — Rhodesia — from Rhodes, whose company had funded the excavation that destroyed the evidence. The state that replaced it in 1980 took its name from the site.

A familiar problem with an unusual motive

Plenty of these kingdoms are known mainly through accounts written by people with a stake in them. Songhai is known largely from two chronicles written by the scholarly families its founder had persecuted. Almost everything about the first Bulgarian empire's opening two centuries was written by Byzantines. Ghana's destruction by the Almoravids, a conquest that probably never happened, survived for a century on the strength of being a tidy story.

In each of those the interested party was a rival or a successor, and the distortion was old. Here it was a modern colonial state that needed the ruins on its territory not to have been built by the people it was governing, and it had the archaeology, the museum labels and the civil service to say so with.

What it cost

Hall's clearance removed the stratified deposits from large parts of the site, so the dating is coarser and the sequence thinner than it should be. Whatever those layers held about how the place was organised, fed and governed is simply gone.

We know Great Zimbabwe lived on gold and on the trade that carried it, because the workings and the imports survive being dug out badly. How it raised its authority, how it fed itself, and how power passed from one holder to the next are blank — and some part of them might not have been, if the first serious excavation had been looking for what was there rather than for who else could have put it there.

A well-accepted fact

Drafted from Bent, The Ruined Cities of Mashonaland: Being a Record of Excavation and Exploration in 1891 (1895)aside?What aside means

The theory that Great Zimbabwe was built by foreigners was not folklore. It was assembled by a competent excavator who published his reasoning, and the book is still in print because it shows the whole operation being carried out in the open.

Theodore Bent dug at Zimbabwe in 1891, the year after the British South Africa Company occupied the country, under the Company's auspices. He surveyed the site accurately with R. M. W. Swan, produced plans that are still consulted, and catalogued the finds. Then he explained what they meant.

The argument

It runs through astronomy. Swan's measurements of the curves and orientation of the enclosures established, Bent says, that the builders were well versed in geometry and studied the heavens carefully. He proposes that the great elliptical enclosure is a Sabaean temple of the god Almaqah, the points of comparison being very strong; Professor Sayce contributes a parallel with the prehistoric temples of Malta. A Roman coin of Antoninus Pius, found seventy feet down an old mine shaft near Umtali, is offered as a valuable link in the chain of evidence.

Then the conclusion, in the plainest form he gives it:

It is, however, very valuable confirmatory evidence when taken with the other points, that the builders were of a Semitic race and of Arabian origin, and quite excludes the possibility of any negroid race having had more to do with their construction than as the slaves of a race of higher cultivation; for it is a well-accepted fact that the negroid brain never could be capable of taking the initiative in work of such intricate nature.

The last clause is not an inference from anything he dug up. It is the premise, stated as a fact everybody accepts, doing the work the evidence could not do.

The cemetery

The most instructive passage in the book is about a negative result.

Bent's party searched for the builders' graves and found none. What they did find was African remains in abundance, and a small cemetery of about twenty bodies under piled stone ten miles off, also African. From which:

Consequently we came to the conclusion that the ancient inhabitants, who formed but a garrison in this country, were in the habit of removing their dead to some safer place.

He supports it with a parallel from the Bahrain mounds and the Persian custom of carrying the dead to Kerbela and Mashhad.

Read forwards, that is a reasonable piece of comparative argument. Read for what it is doing, it is the conclusion consuming the evidence: the only human remains on the site belong to the local population, and the response is not to wonder whether the local population built it but to invent a funerary practice that explains the absence of anybody else.

What he would not claim

Bent is not a crank, which is exactly why the book mattered.

On Ophir he refuses to commit. Mashonaland may have been the land of Ophir or it may not; it may have been Punt or it may not; the two may be identical or not; there is not enough evidence, as far as I can see, to build up any theory on these points which will satisfy the more critical investigation to which subjects of this kind are submitted in the present day. All that can be established, he says, is that the ancient Arabians got a great deal of gold from this country.

A man who declines the biblical identification, publishes his plans, and says plainly where the evidence runs out is a much more persuasive witness than an enthusiast. The restraint is what carried the rest of it.

What it was for

The timing is not incidental and does not need to be alleged: Bent says who sent him. A settler company had taken a country the previous year, and its expedition reported that the ruins in it had been built by Arabians and that the people living among them could not have done it except as slaves.

Everything that followed — the Company-funded clearance of 1902, the official line held by the Rhodesian state for most of a century, the pressure on archaeologists not to publish what they had found — was built on a foundation Bent laid in one season with a theodolite and a spade.

And the foundation is visible in the book, because he was honest enough to show his working. He declined the biblical identification, said where the evidence ran out, published his plans, and then wrote down that the local population could not have built it because the negroid brain was incapable of initiative, as a fact everybody accepted. The careful parts and the assumption are on the same pages, by the same hand, and the assumption is the only thing holding the conclusion up.

Left

Drafted from Pikirayi, African Archaeology Series (2001)end

Nobody sacked Great Zimbabwe. There is no destruction layer, no burning, no weapons in the deposits, no defensive rebuilding of the walls and no siege works outside them. Some time in the fifteenth century the people who lived there went somewhere else, and the walls stayed up because nobody had any reason to pull them down.

That is an unusual way for a capital to end and it is worth being precise about what is known.

What is known

The date is a range and not a year. Radiocarbon and the ceramic sequence put the abandonment somewhere in the fifteenth century, over some decades rather than in a season, and the excavation that would have tightened it was destroyed in 1902 by a man digging for Phoenicians.

Pikirayi's synthesis sets out three candidates and does not choose between them, which is the correct posture.

The land. A concentration of people and cattle in one district for three hundred years exhausts what is under it. Soil fertility falls, the firewood within carrying distance is gone, grazing degrades, and the point arrives at which the settlement costs more to feed than the site is worth. This is the ordinary fate of pre-industrial capitals that do not sit on a river delta, and this one sits on granite.

The trade. The gold trade's centre of gravity moved north, towards the Zambezi, and the state that grew there — Mutapa — took the traffic with it. A capital whose whole revenue came from being where the trade was routed stops being viable when the routing changes, and nothing about the routing was guaranteed.

The climate. Possibly. The evidence is thinner and the argument is made with more hedging.

None of the three excludes the others, and the honest reading is that a place that had been marginal in three ways at once stopped being worth the trouble at roughly the same time in all three.

What did not happen

The people did not disappear, and the tradition did not stop.

They went on living across the plateau and went on building the same architecture at other places — Khami to the west, Danangombe after it, and something over a hundred and fifty smaller stone enclosures across what is now Zimbabwe, Mozambique and Botswana. Some of the later masonry is better than anything at the capital, with decorative panelling the older walls do not have, and on a smaller scale.

So the fifteenth century did not end a civilisation. It relocated a court.

Why that is hard to see

Because the capital is enormous and the successors are not, and because a visitor standing inside the Great Enclosure is looking at the biggest thing in the tradition and naturally reads it as the tradition's whole life.

The scale is also what made the ruins available for a different story altogether. A large abandoned stone complex with no inscription, no chronicle and no living court to contradict anybody was an invitation to whoever came along next with a theory — and the people who came along next had one.

Symbaoe

Drafted from Theal, Records of South-Eastern Africa, volume VI (1900)afterlife

The first written description of a zimbabwe was printed at Lisbon in 1552, by a man who never went to Africa, from reports sent back by Portuguese officers at Sofala who had not seen it either. It is one of the more remarkable passages in sixteenth-century travel writing, and everything in it is second-hand.

What Barros was told

In the mines of Toroa, otherwise called the kingdom of Butua, ruled by a prince who was a vassal of Benomotapa — the most ancient workings known in that country, all of them in the open plain —

in the midst of which there is a square fortress, of masonry within and without, built of stones of marvellous size, and there appears to be no mortar joining them. The wall is more than twenty-five spans in width, and the height is not so great considering the width.

Then, above the door, an inscription which some learned Muslim merchants who went there could not read, and could not even say what the character was. The building is nearly ringed by hills, and on the hills are others made the same way, without mortar, "and one of them is a tower more than twelve fathoms high."

Anyone who has stood inside the Great Enclosure will recognise every element: the massive battered wall, wider than it is high in proportion, the hill complex above, and the conical tower. The "inscription above the door" is almost certainly the chevron frieze, described by men who assumed a decorated band on a monumental wall must be writing.

The word

The most valuable sentence is the one about the name.

The natives of the country call all these edifices Symbaoe, which according to their language signifies court, for every place where Benomotapa may be is so called; and they say that being royal property all the king's other dwellings have this name.

So zimbabwe was not the name of a ruin. It was a live administrative term for a royal residence, in use across the plateau a century after the great one was abandoned, and the Portuguese heard it applied to buildings still being lived in. Barbosa, writing about 1518, says there was a great town called Zimbaoche fifteen or twenty days inland where the king of Benametapa often stayed, with many houses of wood and straw.

Barros goes further: each of these places was kept by a nobleman with charge of it in the manner of a chief warden, called the Symbacayo, keeper of the Symbaoe, and some of the king's wives were always resident there in his care.

That is an institution, not a relic. A hilltop stone court with a resident custodian and a household of royal women was the ordinary furniture of government on this plateau, and Great Zimbabwe was the largest example of a type that outlived it by centuries.

Two admissions in one paragraph

Barros then reports two things his own informants told him, and puts them next to each other without appearing to notice.

The first: nobody in the country knew when or by whom the buildings had been raised, since the people had no writing, "but they say they are the work of the devil, for in comparison with their power and knowledge it does not seem possible to them that they should be the work of man."

The second: Vicente Pegado, captain of Sofala, showed some Muslim visitors the Portuguese fortress there — its windows, its arches — so that they could compare it with the stonework of the inland edifice. They told him the two could not be compared, for smoothness and perfection.

The Portuguese fort lost. That is a Portuguese source, printed under royal privilege, recording that its own new fortification was judged inferior workmanship to a wall built without mortar by people the same book calls heathens.

Three and a half centuries later a man would stand in front of the same wall and conclude that Africans could not have built it.

What the plateau did next

The court kept moving and kept working. North on the Zambezi escarpment, Mutapa, which the Portuguese reached, traded with, tried to convert and eventually undermined — the first state here that outsiders wrote about while it was alive. West, Khami, and after it the Torwa and the Rozvi at Danangombe and Naletale, building the same walls better.

The gold went on going to Sofala and being paid for in advance with Gujarati cloth, exactly as before, until the Portuguese took Sofala and began buying at the source.

And the word survived all of it. In 1980 a country took its name from a building whose builders left no name of their own, and whose name in the sixteenth century simply meant the place where the king is.