Version classiqueVersion mobile

Les bâtisseurs sur les deux rives de la mer Rouge

Jean-François Breton

Chapter 7. Houses and palaces

Texte intégral

7.1. Identification of buildings

1All South Arabian cities comprise a number of buildings, all similar in size, aspect and building techniques. Most of the platforms are completely eroded or buried in debris: they have lost their original superstructures. Archaeological excavations at Shabwa (fig. 7.1), Tamna‘ and Raybûn (fig. 7.2) demonstrate that all these stone platforms were topped by a vertical wooden frame. The long-beams and cross pieces were carefully interlocked so as to construct a regular vertical frame, about 1.50 m high and 0.60 m-0.70 m wide. Two such frames, one on top of the other, supported a floor at a height of 3.00m. As it seems that the layout of the ground floor does not in itself allow one to distinguish a house from a temple, the nature of these buildings remains generally uncertain.

2The ground floor of the houses generally consists of a central corridor, with small lateral rooms, leading to a staircase located either along the axis of the corridor or perpendicular to it (fig. 7.3). The staircase leads to a first floor whose general organization is similar to the ground floor, except that some rooms may be wider. Whatever the number of floors, the house would be like a tower whose main function was the safety of its inhabitants. By contrast, the temples generally consist of a main cella (or one hypostyle hall and one or sometimes three cellae). The temples have an upper open-air terrace which could be reached by a staircase near the main door. But the building techniques of their stone platforms and of their wooden upper frames are quite similar to those of houses, so that it is hard to distinguish one from the other, except by detailed excavation of the stone platforms.

  • 1 Breton, Darles, Robin, Swauger 1997: 67-72.

3The existence of “religious” material (incense burners, votive stelae, altars, dedications...) may indicate a temple, but many houses (or palaces) also have private oratories. The main building at Tamna‘ (so-called “Tamna‘ Temple 1”) yielded some incense burners, bronze fragments of statues, five stone statues all with the same text (Jamme 874-875) and five statuettes dedicated to the “The Daughters of God” (Jamme 868-872); but dedication texts need not necessarily mean a religious building1.

  • 2 De Maigret 2005.

4Similar stone platform houses were excavated in Tamna‘, and their identification raises the question of the term “bayt”: house or temple? A. de Maigret2 points out some possible comparisons between the temple of Nakrah in Barâqish and houses in Tamna‘. The Nakrah temple is built on a base that appears to reproduce the masonry ground floors of the private houses. About 1.30 m above external ground level, it is composed of a thick outer wall containing a grid of large stones arranged at right angles to each other so as to act as a support for the four rows of three pillars forming the five aisles of the cella and the posts of the doors leading to the five rear cellae. Both the specific architectural layout and the general ground plan of this building seem the same as in the private houses. For de Maigret, the number of analogies increases if one takes into consideration the main entrance aligned with the central corridor, the staircase giving access to it, and the vertical rise of the inhabited portions over the upper surface of the base.

7.2. Tower houses

5Tower houses rise above a high stone platform or elevated podium, composed of a grid of stone walls demarcating rubble-filled compartments. A number of foundation-platforms of this kind have been partly excavated at Shabwa; they have sometimes been mistaken for tombs or for storerooms.

6It is possible that these solidly constructed foundations, such as those found beneath Shabwa houses, were not the result of a desire to create a firm base, but were mainly motivated by military considerations. Houses were multi-storeyed, and designed mainly for protecting their inhabitants. Their high stone platforms had no openings, and no easy access (only an L-shaped staircase); the ground floor had few ventilation holes, and the house was designed as a tower. The South Arabian houses are of a type where the principal pattern of circulation is vertical rather than horizontal. It is clear that such a tower-like development represents one possible basis for urban housing, rather like a modern skyscraper. The floors of the houses were made of a timber framework of regular dimensions (fig. 7.4 and 7.5). In the royal palace of Shabwa, each vertical frame was 1.50 m high, and two such frames together delimited the height of a floor (3 m). Windows and doors were inserted in the frames. Outside, these wooden frames were coated with mud plaster, or covered with decorated limestone slabs.

  • 3 Panel now kept at the Museum of the King Sa’ûd University, Riyadh. Routes d’Arabie 2010: 341.

7The existence of at least two painted panels makes it possible to reconstruct the external appearance of these houses. The first, from Qaryat al-Faw3, depicts a tower house painted in vivid colors. From bottom to top it shows a stone platform, a stairway ascending to the ground floor, two doors (one with a human figure), a wooden storey with crossbars, a first row of windows with busts of human figures, another wooden storey, two upper rows of windows (a full row and a half row) with other busts, and terraces set back from one another with inscriptions, merlons, and ibexes at each angle (fig. 7.6).

  • 4 Al-Salami 2012: 93-95.

8The other image, at Jarf al-Yahûdî, some 50 km east of Sanaa, depicts at least four strong tower houses having storeys with wooden frameworks, topped either with merlons or with upper terraces set back from one another4. These houses contain towers which include high arches surmounted by (apparently) wooden storeys, topped with merlons.

7.2.1. Ground floor

9In Shabwa houses, the central corridor, parallel to the length of the building, ranges from 2.50 m wide (building 52), 2.10 m (Royal Palace), 2.10 m (building 72), 1.90 m (building 85); in Tamna‘, the corridor is 2.5 m wide (house B/A) and ca. 2 m wide (house B/C). As for the smaller houses of Mashgha, in houses I and Κ: 1.70 m-1.75 m wide, and in house J: 1.45 m. The width corresponds to the width of the stone foundation walls. The corridor opens on lateral rooms.

10In the Royal Palace of Shabwa (fig. 7.3), a central corridor, north to south, some 2.10 m wide and 22 m long divides the whole ground floor. At its southern end, a staircase leads to the upper floor(s). On both sides of the corridor, there are rooms of different sizes. Some are narrow (1.50 m to 2.20 m across): rooms n° 13, 14, 15, 16 on the NW side, and rooms n° 6, 7, 8, 9, 10 on the south side. Some rooms are divided into two smaller ones by a wall (rooms 3 and 4). Large-sized rooms, generally 4 m x 4 m (rooms n° 5 and 12), have a central mud brick pillar to support the ceiling beams.

  • 5 Bonnenfant 1989 and 1995.

11Doors with high thresholds, standing about 0.50 m to 1 m above the floor level and probably reached by a flight of stone or wooden steps, are found in many houses. Such high thresholds open onto small-sized rooms, sometimes containing basin(s) made of mud bricks or stone. Similar elevated thresholds opening onto small rooms were recorded in Yalâ (house A) and in Mashgha (houses I, J, and K). In Mashgha, thresholds are ca. 1.40 m above the floor, and were probably reached by wooden steps. Given the small size of the rooms on the ground floor, they were evidently not used for living but for storage of grain, probably with small windows to admit daylight. Traditional mud brick houses in the lowlands of Yemen have small rooms on the ground floor, used for the storage (of grain?) or goods; the second floor also has small rooms used for storage; and the third floor has a lavatory and larger rooms for living5.

7.2.2. Staircase

12The corridor ends in a staircase with flights of steps and landings (generally half-space landings extending across the width of the staircase). In the royal palace at Shabwa (building A), the flights of steps were inserted in the walls at both ends, one end in the central masonry core (2.50 m x 1.50 m) and the other in the wooden frame walls. Only the first three steps of the first flight, made of mud bricks, have survived; the first landing has vanished and the upper steps were also probably made of mud bricks with wooden supports. Raybûn houses show similar staircases of mud brick steps supported by beams and half-space landings. Building 72 at Shabwa has an exceptional stone staircase, probably of a late date. Vertical circulation inside the houses was thus quite elaborate, with regular flights and landings and with steps of a regular rise and width, and ascending to a considerable height in the main houses.

  • 6 Akopjan 1996.

13The city of Raybûn I has a minimum of 30 houses built on a stone platform with upper floor(s) made of vertical wooden frameworks, generally dated between the 5th BC-1st century BC, and destroyed about the end of the 1st century AD. About four of them were excavated by the Yemeni-Russian Mission in 1983-19956. House RAY-1-5, 10.30 m wide by 12.80 m long (about 125.40 sqm), stands upon a 4 m high stone platform (fig. 7.7). The platform is divided by four longitudinal walls framing the organization of the ground floor. The ground floor consists of four wide rooms on both sides of a central corridor. On the NW side, four rooms are connected by small doors. The corridor, 1.75 m wide, leads to the staircase, of which the first six stone steps are well preserved.

14House RAY-1-2, located south of the city, is 9.25 m by 10 m. On both sides of the central corridor, there are three rooms on the south side and two on the northern side (fig. 7.8). The upper walls show a regular wooden framework with inner mud filling. The outer side of the walls was entirely covered by thin limestone slabs decorated with central dressed chiselled panels.

15Some houses had a more complex design. At Shabwa, building n° 96 consists of two wings laid on both sides of a square open-air courtyard. Building n° 74 consists of a central courtyard with porticoes on two sides and rows of rooms behind them. As there is no staircase, it can be assumed that the building was only one storey high.

16In Makaynûn, building A (30 m x 25 m) consists of 40 rooms around a rectangular central courtyard. In its first chronological phase, the building comprised 11 series of two or four rooms, separated by thick double walls, each with a single door opening onto the courtyard. During its second phase, only 11 rectangular rooms survived, generally 7.70 m-8.00 m long by 4.20 m-5 m wide. The main gate remains unchanged, consisting in a long corridor with “teeth” narrowing the passageway. Not earlier than the 4th century BC, this could be either an “administrative” building or a “house of priests” (?).

7.2.3. Axumite houses

17Most Axumite houses belong to the above type: a stone podium with floors made in wood and mud bricks. In the wealthy houses of Axum, the upper stories consisted of regular wooden frameworks; however, due to erosion, fire or plundering, most of these have vanished completely.

  • 7 An fray 1963: 93-96.
  • 8 Anfray 1965: 51-54.

18Some houses have been excavated at Matarâ by F. Anfray (fig. 7.9 and 7.10). The main house (or “la construction principale”) consists of a podium (12.60 m by 11.20 m) some 2 meters high (fig. 7.10 and 7.11) with an internal grid of walls7. Some of the internal compartments have been excavated deeply, and some may have been used as storerooms. At Matarâ, the central building of House B (Anfray’s terminology: “le logis central”8) (fig. 7.12), 17.50 m square, also consists of a high platform with a regular grid of internal walls delimiting seven large compartments and four smaller ones, some 3 m deep. Three staircases lead to the upper floor (s) (fig. 7.13).

19At Axum, two areas of houses were excavated in 1995-1996 by D. W. Phillipson, one within and one outside the main built-up area. The houses were built of small roughly dressed blocks, without a massive podium and without extensive use of timber. The external area, to the north, preserves a complex of rectangular rooms and courtyards, subject to alterations and modifications. Dated 6th cent. AD, these houses seem to reflect a higher level of prosperity than at Matarâ.

7.3. Houses and their inscriptions

20Inscriptions are generally brief, both as regards the number of activities specified by the verbs and the enumeration of the parts of the houses. The building of a house is generally described as: “from the foundation to the top”. Stone platforms are clearly depicted in some texts as: “X and Y have laid the platform” (text DJE 12), or “has built and laid the foundations” (text from al-Harj). In one inscription from Shabwa (S/77/Mahdi) some rich landlords or merchants constructed and secured their house “Yafad” under the reign of Ili’adhYalût, king of Hadhramawt (3rd cent. AD.) (fig. 7.15).

  • 9 Jamme 1958: 186-187.
  • 10 Jamme 1958: 190-191.

21One text from House Yafash in Tamna‘ (Jamme 118) is interesting for its architectural depiction of the house: “and restored his house Yafash and its reception hall (?) (ms3wd) and its roof or gallery (s1qf), east from this house, and their two upper rooms (s2rḥt) and their terraces and their parapets (?), from the foundation to the top”9. Another text (Jamme 121), from Tamna‘, describes the wooden works (hṭb) and the upper rooms (s2rḥt) of the house Yarish10 (fig. 7.14).

  • 11 Maraqten 2013: 73-76

22Another text from Hinû az-Zurayr mentions “Tawab, son of Hâdirum of the family Lamasum”... (who) “constructed as well as he completed his house Mayfa’um and all its lower rooms and its upper rooms, its windows and its room terraces, together with full ownership rights” by Tawâb, son of Hâdirum (Text HZ Μ 1)11.

23As shown by the excavations, these houses were rarely decorated. Their platform is laid in coursed masonry, and only the wealthiest houses had a platform in finely dressed ashlar masonry. The royal palace of Shabwa shows a 5m high stone platform with fine pointed dressed headers and stretchers, with polished margins all around. Its wooden framework delimits regular panels which were covered by thin limestone slabs, with pointed dressed and chiselled rectangles. Openings with wooden frames were used as windows or doors. In some cases, alabaster was used in the windows.

7.3.1. Tower-houses

24The original height of these ancient houses is unknown. The few surviving walls rise no higher than two meters. In an effort to resolve the question, archaeologists examined the dedicatory inscriptions created at the time of construction of these houses. From Tamna‘ as from Shabwa no inscription provides details about the height of the houses. There is now a “general trend” to reconstruct two floors in average houses, and more in exceptional houses like the so-called “palaces”.

7.4. Palaces in South Arabia

25There is speculation regarding the importance of the courtyard in the design of the so-called “palaces”. A special category of “courtyard houses” has been proposed. Courtyards seem to play a very significant role in house functional layout, but also in temples. In fact, design formulae appear to be few in number. The simplest form of temple had an open-air courtyard with rooms at one end; in later periods more elaborate temples were constructed with a high central building and a lower courtyard surrounded by porticoes. The same formula seems to have been used for sanctuaries such as the Bar’ân temple in Mâ’rib, and palaces like the so-called “Harîb palace” (or Tamna‘ Temple 1) and the palace of Shabwa. The principles governing the layout of these buildings have not received much detailed architectural consideration.

26The designation of “courtyard house” is, however, not very explicit. If it means anything it should mean a house where a number of rooms open onto a central courtyard.

7.4.1. General layout

27The general layout of these constructions consists of a central building (building A), with a high-rise stone platform, dominating a central courtyard almost surrounded by another U-shaped building (building B). Access to the courtyard could be either by passages between these two buildings or by a gate opening onto the courtyard.

28The Bar’an temple in Mâ’rib is designed this way. This is a monumental building, ca. 62 m (north to south) by 75 m (east to west) (fig. 7.16). On the east side stands a stone platform, the socalled “Building A”, 19.5 m by 27.50 m, ca. 4 m high, consisting of four integrated temples (on the same platform) built between the 9th and 5th century BC. At its western entrance was a monumental propylon with six monolithic pillars, 8.30 m high, probably the highest in Arabia. On the west, there is a U-shaped building (“Building B”), having a wide portico on its three sides, with benches and altars. The courtyard below (36.5 m by 31.50 m), four meters below the ground floor of building A, is entered from the west. In its early period, the building was east-west oriented, from the upper cella in the east to the western gate.

  • 12 Strangely, these benches contain animal bones.

29The difference between a temple and an ordinary building inheres in the different layout and architectural organisation of the central building (A). The Bar’an building A has a monumental propylon leading to a small porch opening onto a central courtyard with porticoes on three sides. In the centre of this small courtyard was a rectangular adyton containing a life-size bronze idol (perhaps a bull), two stone altars, and a statue of the donor (fig. 7.17). The lateral porticoes consist in two rows of monolithic pillars supporting stone architraves and a stone roofing (as in the Masâjid temple or the ‘Athar dhû-Risaf temple in as-Sawdâ’). As no staircase was designed, there was no way to reach a terrace that might have been used to perform worship. It should therefore be concluded that the layout is of a one-floor building with a courtyard, porticoes and cella, clearly dedicated to the cult of ‘Almaqah. The size of the cella allowed only a few people to enter. The majority of the public would have worshipped in the lower courtyard (with the well) and in the porticoes (with altars and hollow benches12).

7.4.2. The “main building” at Tamna‘

  • 13 Breton, Darles, Robin & Swauger 1997.

30The “Main building” (TT1 or Harîb Palace) at Tamna‘, oriented on an east-west axis, can be divided into two main sections13 (fig. 7.18). The eastern building (A) consists of a massive undressed masonry platform, ca. 10.30 m high, with large recessed niches; it is a trapezoid ca. 32.50 m on its eastern side and 23.30 m on the south. Building B, on the west, is U-shaped, ca. 36.70 m (north-south) and 24 m (east-west), and delimits a rectangular courtyard. Irregular in width, it comprises three rows of rooms, with a large portico on all three sides, and a main entrance on the north (fig. 7.19). The paved courtyard, rectangular in plan (17.10 m x 12.25 m), is ca. 1 m below the ground floor of building A. The ground floor of building A was reached through a monumental propylon, with flanking steps, made of a high platform (7.75 m x 3.30 m) supporting 8 stone pillars.

  • 14 Breton-Darles-Robinx& Swauger 1997.

31The eastern “Building A” shows neither a central courtyard with porticoes nor a central adyton nor a podium. According to archaeological records of the excavations in the 1950s, the ground floor consists of a central corridor with lateral rooms. These rooms, very narrow (averaging 1 m wide), are separated by 1 m thick walls. The purpose of building these walls so close to each other was to support the heavy weight of the upper floor(s). These rooms may have been used as storerooms. The ground floor also shows a number of rectangular stone bases, about 1 m square. According to the drawings by J. L. Swauger14, some of these bases were built over the walls of the platform, in parallel rows to the central corridor. In this particular case, the ground floor may comprise some pillared hall(s), but probably not a single hypostyle hall. The erosion of the stone platform has caused both the accesses to the rooms and the staircases to disappear. It is thus impossible to know if there was just one axial staircase, or several staircases located in the corners of the building.

32The remains of the walls on the ground floor suggest an organization with many rooms, some very small and some wider on the western side. There is a central corridor (east-west) probably leading to a rear staircase, but lateral staircases can also be presumed to have existed in the angles. This layout seems very similar to that of the Shabwa palace (see below). The Shabwa “Building A” consists in many storerooms on the ground floor and wider rooms for living on the upper floors, while building Β seems also to have been used for storerooms on the ground floor. It is also similar to Yeha/Grat Be’al Cuebri main building where the ground floor has many rooms of a small size on both sides of the central corridor. As the platform is reinforced by “angle bastions”, it can be supposed that one or more staircases were located inside the bastions, as at the Endâ Mîkâ’êl and Tâ’akhâ Mâryâm palaces in Axum.

7.4.3. The “Royal Palace at Shabwa” (fig. 7.21)

  • 15 Seigne 1992: 110-131.

33The general layout of the “royal palace” at Shabwa consists in a central building (“Building A”), 22.30 m x 19.80 m, with a 6.50 m high-rise stone platform, on which wooden framed floors were built. “Building B”, on its northern side, U-shaped, ca. 32.50 m (north to south) by 38.5 m (east to west), encloses a rectangular courtyard15. This building, ca. 5 m wide, comprises three series of rooms of various dimensions and porticoes, a single one in front of the west and east sides, and a double one in front of the northern side. The north and south aisles, flanking Building B, create two narrow passageways (ca. 0.90 m and 1.20 m wide) between Buildings A and B. The paved courtyard, ca. 3.80 m below the ground floor of building A, includes a monumental staircase to enter building A.

34Building A shows a central corridor opening onto a number of lateral rooms. At its southern end, a staircase leads to the upper floors. The ground floor has no hypostyle cella, no lateral porticoes, no podium, and no religious installations (altars etc.). The ground floor rooms are small in size, without any openings: they were storage rooms (and not living rooms). Bigger rooms were located on the upper floors, with wooden windows to admit light.

7.5. Palaces in Tigray: from Yeha to Axum

  • 16 Anfray 1997: 8-11.

35Grat-Beal Guebri/Yeha is a major building in Tigray with extensive dimensions (48.50 m wide) and a high-rise podium (8 m) (fig. 7.23). On the southern side, a large staircase leads up to the level of a massive platform. The monumental entrance lobby consists of a front line with six horizontal bases, ca. 2.10-2.30 m long by ca. 1.20 m-1.30 m wide, supporting rectangular pillars (ca. 0.80 m-0.95 m in section), and a rear line, on the north, with only two preserved bases16. These two bases, 1.18 m apart, seem to mark the entry of the building. Due to the poor state of preservation of the remains, reconstruction of the entrance lobby is hypothetical, but, in any event, one can affirm that this entrance system belongs to the typology of South Arabian buildings, such as the Nakrah temple in Barâqish, the extra-muros temple in Ma‘in, and the main building in Tamna‘.

  • 17 Anfray 1997: 17.

36Behind the entrance, the massive platform consists of a grid of stone walls demarcating rubble-filled compartments. Many cellular platforms built to support the superstructures were carefully excavated by F. Anfray, some down to 3.40 m17. The walls of these compartments show rough rubble, earth-clay mortar and many pieces of wood. The elevated platform is similar to typical platforms of South Arabian buildings.

  • 18 Krencker 1913: 107-117.

37The art of building is represented in Axum in a highly significant way to judge by the surviving remains. The monuments of Tâ’khâ Mâryâm, ‘Enda Sem’ôn and ‘Enda Mikâ’el were partly excavated in March-April 1906 by the DAE18. They were called palaces only because of their large dimensions.

7.5.1. General layout

38The Tâ’akhâ Mâryâm palace is 120 m long by 80 m wide, with a central building of 24 m on a side. The most impressive palace was ‘Enda Sem’ôn, whose external dimensions are unknown due to lack of a complete excavation, but its main building is roughly 35 m on a side. The central building of ‘Enda Mikâ’el is ca. 27 m on a side (fig. 7.24). The square central pavilions are similar although slightly different in internal organisation. They consist of a square with large angular projections. Two symmetrical staircases lead to the ground floor. A huge porch leads to an axial corridor, opening on lateral halls (or hypostyle halls): ‘Enda Sem’ôn has two main hypostyle halls (ca. 20 m by 10 m) with 4 rows of 7 pillars each. The projections contain smaller halls and staircases.

39Another podium of impressive size can be inferred from sections still visible under the terrace on which Maryam Tsion church was built. Destroyed in the 16th century AD and rebuilt in the 17th century, the presence of the church prevents the examination of the remains of the ancient monument; but its platform, visible at the east and the south, indicates that it was one of the largest buildings of Axum.

7.5.2. Dongur Palace

  • 19 Anfray 1972.

40In Axum, a building of which only the platform remain, offers a typical example of the traditional building technique19 (fig. 7.25). It is located at the western periphery of the ancient city, in the Dongur locality. The surface occupies 3,000 sqm. It forms a quadrilateral of 57 m by 56.50 m. Forty rooms, ordered in two asymmetrical housing units, are grouped around a central building of 18 m on a side. This central pavilion (there remains only its platform and approximately 1 m of the upper walls) included 7 rooms at the ground floor. A monumental staircase of 7 steps gave access to it in the east. Two other double-flight staircases are found on the other sides.

41The stepped platform, ca. 1.80 m high, is built in coursed rubble masonry bonded with mortar. A course of large squared blocks runs all around its top. In the angles, quoining of dressed blocks reinforces the structure. The recesses give the podium a very characteristic trapezoidal aspect. Evidence for at least one upper storey is clearly supplied by stairs and stairwells. Courts separated the main building from the surrounding secondary dwellings.

7.5.3. Reconstructions of two palaces: ‘Enda Mikâ’el and Shabwa (ca. 3rd cent. AD)

42There is good reason to believe that many buildings in Axum were constructed in the same style that is depicted in stone relief on the great stelae. D. Krencker emphasizes that the northernmost stela shows a narrowed-down version, reduced in scale, of a real built-up façade. He applied his theory to the ‘Enda Mikâ’el palace, and some other architects have applied the same theory to the Shabwa royal palace in Yemen. This northernmost stela possesses three storeys.

  • 20 Buxton & Matthews 1974: fig. 8-17.

43The multi-storeyed monoliths show, on the ground floor, high ceilings, probably higher than those of the first and second floors, corresponding to large and impressive halls in ‘Enda Mikâ’el, and to small dark rooms in the Shabwa palace Building A with small lower windows (fig. 7.26 and 7.27)20. The stelae show that the best rooms of the buildings were on the upper floors. Windows of the top three storeys have elaborate grills of miniature-arch design. In Shabwa palace (Building B), the central mullioned windows are located on the first floor, as are the finest decorative elements, namely the coffered ceiling. No doubt the second upper floor had wider rooms: a wooden four (or two) pillar system would have been enough to support the second upper floor or the flat roof above. In South Arabia, the debate turns around the number of floors which the houses (or palaces) of Shabwa and Tamna‘ had: a second (or a third) upper floor could have been the rule.

44The giant stela (n° 1) shows projecting courses above each of the upper storeys. Such courses of almost horizontal slabs occur in ancient Ethiopian buildings (Debra Damo) and in more recent ones, their purpose being to throw rainwater clear of the walls. Similar courses of schist slabs have been found in Shabwa palace.

45In the ‘Enda Mikâ’el palace, D. Krencker determined that the thick walls found at each corner of the platform were foundations of stairways, thereby deducing that the building had at least one upper storey (fig. 7.24). The Tamna‘ Main building (or TT1) had four similar corners possibly with stairways in each of them (or at least two on the eastern side). The presence of an upper floor in this building constitutes possible evidence that it was indeed a house (or an administrative building) and not a temple. From the impressive height of the platform, one can infer that the original height of the Main building at Tamna‘ must have been considerable. But the houses of the Market place at Tamna‘ probably had two upper floors only.

  • 21 Buxton & Matthews 1974: fig. 17.

46The roofs of all these houses were flat. Since the occupants of the palaces would certainly have used these roof-terraces, we may assume that they were surrounded with parapets, pierced at their base by drainage holes which probably led through to projecting gargoyles. Some zoomorphic stone waterspouts have been found at Axum, and gutters with bull’s head spout at Shabwa. The question remains how the tops of the corner-towers would have been finished off, and it seems probable that they would have been covered by a slightly sloping flat roof (Shabwa). In ‘Enda Mikâ’el, one possibility is a perforated parapet with a purely decorative arch-design (fig. 7.27) (and not crenellations as supposed by D. Krencker)21.

  • 22 Buxton & Matthews 1974: fig. 2 and 3.

47The plan of Enda Mikâ’el, as revised by Buxton-Matthews22, shows that the palace may well have been divided into two sets of rooms, each set with its own staircase (fig. 7.24). The presence of two symmetrical entries suggests that there was such a division. Each suite would have comprised, on the ground floor, a high divided hall, a smaller hall and a corner room (besides the stairway). The ground floor with its rougher architecture, small windows and lack of privacy was presumably used by guards or servants, and the larger halls by the ping (?). The plan of the Shabwa palace (Building A) is different because of its single entry, and its long corridor leading to the single stairway. But the small size of the rooms and the poor illumination argue either for guards’ living quarters or for storerooms.

48In both buildings (Enda Mi’kael and Shabwa palace) the private quarters of the household would have occupied the upper floors. In Axum, rooms would have been devoted to dining and the reception of important guests, and the open roof would have provided a pleasant extension. At Shabwa royal palace, rooms could be devoted to the household (with the kitchens inside as in traditional Yemeni houses), but there is no indication of any separation of the floors by gender (as in modern Yemen).

49In the architecture of South Arabia and Tigray, it is very probable that the possession of a large house reflected the social status of the owner and the prestige of the family. The central pavilions, raised on high podia, isolated by the courtyards surrounding them, and approached by massive flights of steps, may well have been an expression of the special position of the rulers translated into architectural terms.


1 Breton, Darles, Robin, Swauger 1997: 67-72.

2 De Maigret 2005.

3 Panel now kept at the Museum of the King Sa’ûd University, Riyadh. Routes d’Arabie 2010: 341.

4 Al-Salami 2012: 93-95.

5 Bonnenfant 1989 and 1995.

6 Akopjan 1996.

7 An fray 1963: 93-96.

8 Anfray 1965: 51-54.

9 Jamme 1958: 186-187.

10 Jamme 1958: 190-191.

11 Maraqten 2013: 73-76

12 Strangely, these benches contain animal bones.

13 Breton, Darles, Robin & Swauger 1997.

14 Breton-Darles-Robinx& Swauger 1997.

15 Seigne 1992: 110-131.

16 Anfray 1997: 8-11.

17 Anfray 1997: 17.

18 Krencker 1913: 107-117.

19 Anfray 1972.

20 Buxton & Matthews 1974: fig. 8-17.

21 Buxton & Matthews 1974: fig. 17.

22 Buxton & Matthews 1974: fig. 2 and 3.

© Centre français des études éthiopiennes, 2015

Conditions d’utilisation :

Cette publication numérique est issue d’un traitement automatique par reconnaissance optique de caractères.
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search