Scholars struggle to interpret them

The pillars of Hazor’s tripartite building still stand at crisp attention.A barracks or a bazaar? Could it be a temple? Or maybe a stable? Perhaps a storehouse?

These are some of the suggestions regarding the function of a very important type of building that appears at one Israelite site after another for nearly 500 years.

Yet we don’t know for sure what these buildings were used for.

What’s more, even the most recent and most sophisticated techniques for understanding how structures at a site function have done little, if anything, to settle the matter among the squabbling archaeologists—of whom the writer is about to become one—who are debating the matter.

The building is called a tripartite pillared building. As its name suggests, it is a large, elongated rectangular structure divided roughly into thirds by two parallel rows of pillars that create a long central hall and two adjacent side halls.1

These buildings appear in ancient Israel from the late 12th to the early 11th centuries B.C. (the time of the Judges) down to the 8th or 7th century B.C. (the Divided Monarchy). Some scholars have suggested that the tripartite pillared style was uniquely Israelite,2 but a few of these buildings also appear at non-Israelite sites and they may have a non-Israelite origin.

On only one thing can scholars agree- These are very important public buildings, clearly distinguishable from domestic buildings by their size and construction. They are too big and their exterior walls are too thick to be houses.

More importantly, most of them are located within public sectors near city gates—often near public squares. All of them are close to the outer limits of the city. Moreover, they were all erected as independent structures; most dominated their surroundings. Thus, we can conclude that these tripartite pillared buildings served some important communal function.

Let us look more closely at the archaeological evidence to see if we can figure out what that function was. One problem is that there are considerable variations from one building to another. Almost nothing we can say is true for all of them. So, in the end, we will have to ask how the variations fit into whatever conclusions we reach. One possibility is that not all these buildings served the same purpose.

Another problem is that many tripartite pillared buildings were excavated long ago and the published reports fail to tell us much that we would like to know. Indeed, many of them were excavated in a way that would be considered inadequate or even unacceptable by today’s standards.

Consider some of the building variations-

First, they vary considerably in size. The smallest, at Tell Abu Hawam, measures roughly 25 feet wide and 35 feet long. The largest, one of several at Megiddo, measures about 40 by 88 feet. The proportion of width to length also varies, but the length varies far more than the width. The width ranges from about 23 feet (Tell Malhata) to 44 feet (Hazor); the length from about 35 feet (Tell Abu Hawam) to 88 feet (Megiddo).

The internal width of the side rooms (that is, excluding the outer walls and the pillars) varies little, however. They may be a little over 6 feet to about 9 feet wide; most are about 7.5 feet wide. The central hall is generally about the same width.

The floor of the side rooms is generally different from the central hall. Most of the side rooms are uniformly paved with unhewn stones or cobblestones. The central corridor is unpaved; the floor there consists of either chalk, beaten earth or clay. This uniformity of construction technique clearly indicates a designed functional difference between the center corridor, on the one hand, and the side rooms, on the other.

The distance between the pillars varies between about 2.5 and 5 feet. It is not always clear what was originally between the pillars. At Megiddo, the excavators found some shallow stone bins that they believed to be mangers. We will return to the significance of this evidence later. At other sites (Beth Shemesh and Tell Abu Huwam), the excavators found rubble fill between the pillars. Elsewhere (Hazor), they found platforms.

Unlike most public buildings of this period, which were approached by steps, the tripartite pillared buildings were approached by way of a gently inclined ramp.

There was usually one doorway, although in at least one case (Beer-Sheva) there were three, one at the end of each long room. The single doorway usually led into the center corridor, although in at least one case (Beth Shemesh) it appears to lead into a side room. The single doorway was thus usually in the short wall, but in one case (Hazor) it is in the long wall near the corner.

How did one pass from the central corridor to the side rooms? This is not always clear from the excavation reports, although it is likely that access to the side rooms was located at the end of the structures nearest the doorway. On the other hand, if the space between any two pillars was vacant, access could be obtained at that point.

None of the buildings’ roofs has survived. There are, however, only two possibilities. One is that only the side rooms were roofed and the central chamber was left open.3 However, this suggestion is untenable because this would have exposed the central hall in inclement weather.

The more probable solution would be a clerestory design, that is, with the light coming from above.4 This was achieved by raising the roof over the central room above the level of the side rooms, thus creating an area for windows for ventilation and lighting in the raised side walls. The two rows of pillars inside the building supported the roofs of the side rooms and the raised roof of the central room. Raising the central roof allowed the passage of light and air to the side rooms, as well as to the central hall.

This roof construction is especially likely because these buildings sometimes come in sets with common adjacent wide walls. Although the excavators often found only one tripartite pillared building at a site, at other times they found sets. For example, at Beer-Sheva, there was a set of three. At Megiddo, there were two sets in different locations, one set of five and another of 12 (which can be broken down into subsets of five, five and two). Doorways, or even window openings in the end walls, would not have provided adequate light and air, so the roofs must have been constructed on the clerestory principle.

Scholars who suggest that the tripartite pillared building style was a uniquely Israelite point to the fact that they begin to appear just when the Israelites were emerging in the central hill country of Canaan and continue through much of the Israelite monarchy at sites into which Israelites expanded. On the other hand, the earliest examples were found at Tell Abu Huwam and Tell Qasile (from the Iron Age I period—1200–1000 B.C.) when the former was under Phoenician (or Canaanite) sovereignty and the latter was under Philistine sovereignty.5 In addition, a similar style building was unearthed at the eighth-century site of Bastam near Lake Urmia in Iran.6

The first tripartite pillared buildings were discovered at Tell el-Hesi by the American archaeologist Frederick J. Bliss in 1894. Bliss initially suggested that the structures may have been bazaars.7 The compartments between the pillars were thought to have served as individual stores. Bliss later rejected this proposal and submitted an alternate solution- The structures were barracks for soldiers. However, Bliss realized the many problems inherent in the barracks proposal and, ultimately, regarded the role of the tripartite pillared building as a mystery.

More recently, at least two other investigators (James Pritchard8 and Volkmar Fritz9) argued for the suggestion that Bliss abandoned—namely, that they were used for barracks. Several consideration, however, require us to reject this suggestion- First, no evidence exists in or near the tripartite pillared buildings of any arrangements for a kitchen. Nor have ovens or hearths been found. A study of barracks design from ancient Egypt to the modern American West indicates that some type of cooking area was absolutely essential.10 Another necessary feature of barracks design is the presence of granaries and storage areas within the structure. No such areas were available in the tripartite pillared buildings. A water supply is equally necessary for military housing, but no such resources have been detected in pillared buildings.

A German archaeologist, Ernst Sellin, working at the site of Ta’annak in 1903, discovered two rows of stone pillars that he interpreted as a “ritual pillared street.”11 Actually, what he had found was part of a tripartite pillared building.

When sets of tripartite pillared buildings were discovered at Megiddo, the University of Chicago excavators decided that the buildings had served as stables.12 According to this suggestion, the side aisles, or rooms, served as stalls for the horses; the central area was used to house carriages and to store food for the horses. Between some of the pillars, the archaeologists found shallow stone basins they interpreted as mangers from which the stabled horses were fed. Some of the pillars had holes in them at a height of about 3 feet; most of these holes went between the side aisles and the central area. These holes were interpreted as tethering holes for the horses. The paving of the side rooms was presumably to prevent the horses from slipping.

The interpretation that the Megiddo pillared complexes were stables was generally accepted by scholars for the next 30 years.13

Then, in the late 1960s and early 1970s, Israeli excavations at Beer-Sheva led by Yohanan Aharoni unearthed a series of tripartite pillared buildings near the city gate.14 The Beer-Sheva excavators concluded that the evidence “clearly indicates that [these buildings] served as storage [units] for various foodstuffs.”15 Unlike the buildings at Megiddo, at Beer-Sheva there were no stone basins that could be interpreted as mangers and no holes in the pillars that could be interpreted as tethering holes. The Beer-Sheva excavators relied instead on a vast amount of restorable pottery discovered in the side rooms of the units. In one side room, for example, they found 136 intact vessels, including bowls, cooking pots, storage jars, flasks, jugs, juglets and a strainer.

This mass of pottery clearly implied that the buildings had functioned as storehouses, according to the excavators.

The stage was thus set for a major scholarly debate that has continued to this day- Were these tripartite pillared buildings storehouses or stables?

Even before the Beer-Sheva excavators published the report of these “storehouses,” the Megiddo stable interpretation was substantially undermined by a major article written by Professor James B. Pritchard of the University of Pennsylvania, who concluded that these buildings were not, and could not be, stables.16 Pritchard’s arguments were powerful, but they did not go unchallenged. The leading Israeli archaeologist of his day, Yigael Yadin, countered with some equally powerful arguments that the buildings were indeed stables.17

More recently, John S. Holladay of the University of Toronto published a lengthy, well-researched article in which he too concluded that the buildings were stables.18

A still-later study by Larry G. Herr, rejects both of these arguments and concludes that the buildings were really market places—going back to Bliss’s suggestion of bazaars in which the stalls constituted stores for different vendors. Herr bases his argument in large part on modern Philippine parallels.19 His argument has not met with much acceptance, however; the major debate continues to be whether they were storehouses or stables.

My own conclusion is that they were storehouses.

On method of analysis, however, I agree with Holladay that you cannot necessarily determine the function of a building by the artifacts found in it. Small-find evidence may merely represent the final use to which the building was put before its final destruction or abandonment. For the investigator attempting to determine the function for which the building was originally designed, small-find evidence may even be misleading. Thus, we cannot rely on the pottery found in a few of these buildings (e.g., Beer-Sheva) to determine the purpose for which these buildings were designed.

Nevertheless, the pottery does suggest a use to which the buildings could be adapted. Clearly they could have been used as storehouses, and indeed were used that way—or at least some of them were in their final phase.

I find it significant that no small finds of equine nature were found in any of the tripartite pillared buildings. Holladay argues that “one properly should expect very little either in terms of occupational accumulation or in the destruction debris of the ‘stables’ themselves which would be specifically related to either horses or chariots.”20 The reasons he gives for that vacuum are the necessary cleanliness of a stable, and the clearing out of all equipment when the horses and chariots of the garrison were taken into the city’s final battle. Yet it is hard to believe that not one sliver of evidence of horse and chariot trappings would survive the destruction of the many tripartite pillared buildings that have been excavated—if they were indeed stables.

Holladay’s conclusion depends principally on an analysis of the buildings’ design to determine its function. I think his method is right, his conclusion is wrong. The original design of a building does accurately reflect its intended function.

Holladay argues that the structural features of the buildings reflect a design meant to accommodate horses. I follow the same method and conclude that the form of the buildings actually reflects their intended function as storage facilities. Holladay considers the site, the arrangement of the building, the provision for ventilation and lighting, the floors, the dimensions, etc., and concludes that they are all appropriate for stables, especially considering the mangers and tethering holes.

A principal difficulty in interpreting the plan of these buildings as stables is that the side aisles are not wide enough to allow the removal of individual horses along the back of the stalls. This difficulty was already pointed out by Pritchard in 1970. Holladay argues, however, that it would have been possible to remove an individual horse without removing all the other horses blocking it on the way to the doorway by having each horse closely tethered “so that his chest was touching the manger” and “each groom could turn and hold his horse between 40 degrees to 90 degrees in his stall.” Holladay concludes that in “one way or another it would easily have been possible to lead any particular horse out of the stable without first having cleared all other animals standing between him and the entrance.”21 Holladay may be right, but this procedure would certainly require a large number of support personnel to run such an operation.22 The Israelites could have saved themselves vast amounts of manpower and money if they had only been smart enough to build wider stables! The fact of the matter is that the tripartite pillared buildings were not designed to allow for the easy removal of individual horses from the side aisles.

Another problem with the “stable” argument is the site of these tripartite pillared buildings. Stables should be located in an area isolated from the population of the city.23 Most of the tripartite pillared buildings, however, were in the middle of the public sector of the city, near the hub of civic activity.

Holladay has suggested that if the buildings had once been used as stables, the underlying soil should reflect the chemical characteristics of enrichment from horse urine. According to one study cited by Holladay, a working horse excretes liquid wastes containing on the average 1.7 grams of phosphoric anhydride per day.24 Following Holladay’s suggestion, the excavator of Lachish, David Ussishkin, conducted a series of tests for heightened phosphorus content of the soils under a building that had a plan very similar to the Megiddo “stables.” Holladay admits the results were “less than conclusive.”25 According to the excavator, Holladay concedes that “no significant differences could be detected between samples from the area of interest and test samples from other locations.”26

The basic functional requirements of a storehouse are to protect the grain from excessive moisture intake; to control the temperature of the grain; to inhibit the activity of microorganisms within the grain; to prevent insects and rodents from infesting the grain; to allow for easy transport, handling and storage of the grain; and to provide access to the grain for needed conditioning. These are unchanging demands, whether in ancient Israel, ancient Egypt or the modern United States.27 Good storage practices encompass all these requirements. The tripartite pillared buildings were designed to meet these requirements.

We will consider several factors-28

• Site. A storehouse should be located in a central, preferably public area, and it should be independent of other structures. It should be built on moderately high ground with natural drainage, since groundwater can quickly destroy cereal crops. The subsoil upon which the structure is built, moreover, should be well drained, ensuring protection against dampness and ground moisture. In addition, the building should be oriented to receive free access to air currents for ventilation.

• Arrangement. “The rectangular [storehouse/crib]” notes a grain-storage specialist, H. J. Barre, “is perhaps the most common type used, largely because of its simplicity of construction and its effectiveness in drying [cereals].”29 The great Biblical archaeologist William F. Albright also recognized the importance of the parallel rectangular rooms- “The long narrow design, like that of modern American farm granaries, helped to keep grain from mouldering.”30 Rectangular storehouses often appear as double crib versions, maintaining a “driveway” through the center off the structures with approximately the same dimensions as the cribs themselves.31 For instance, the military granaries of Rome were designed with long magazines that held grain in bins arranged laterally on either side of a central gangway.

• Access. Granaries designed as long hangars invariably have entrances in the end walls. Discussing the many rectangular storehouses (horrea) of Rome, one commentator states, “All the evidence, however, leads to one conclusion, that the means of access in military granaries was always at the ends and not in the sides.”32

• Roofing. Two major factors control its design- illumination and ventilation. The roof should provide maximum ventilation because stored grain is prone to overheating, especially in a climate like Israel’s. An effective ventilation system will lower the humidity and cool the grain.33 Sustained direct sunlight must be avoided since internal combustion of grain is a common occurrence.34 The best illumination is indirect sunlight. The clerestory principle of roofing is particularly appropriate for a series of rectangular storehouses that share common sidewalls because ventilation and illumination cannot be provided along the sides of the structures.

• Flooring. “The principal functions of a [storehouse] floor,” H. J. Barre observes, “are to hold grain [weight], to provide protection from ground moisture, and to prevent entry of rodents.”35 Storehouse floors are often raised above ground for this purpose; such floors are called suspensurae36 and often consist of stone, concrete, cobblestones and even crushed sherds.37

• Proportions. The dimensions of rectangular storehouses vary, but usually the width of a granary varies much less than its length.38 In general, it would not be misleading to suggest an average proportion of 3-1 for the length-to-width ratio of rectangular storehouses.

It seems clear that the tripartite pillared buildings meet these design requirements especially well-

1. Tripartite pillared buildings consist of a standard rectangular storehouse design with two elongated “crib” rooms and a center aisle, or gangway.

2. The side aisles, where the grain was kept, had thick, paved flooring. The center aisle did not need to be paved because it did not hold grain, but it was clean and dust free.

3. The walls of the buildings—both the outer walls and presumably the dividers between the side rooms and the central hall (assuming the grain was stored in bulk rather than in jars)—were strongly constructed to withhold the lateral pressure of the grain weight.

4. The best roofing design, especially for storehouses in rows sharing common side walls, is the clerestory construction. It allows for maximum ventilation and adequate illumination without direct sunlight. This is the type of roof structure that was almost certainly used in these buildings.

5. The surface areas in and around the storehouse need to be well drained, since grain is prone to destruction caused by wet weather conditions. The subsoil should have good drainage as well to prevent moisture seepage from below ground. Apparently this condition was fulfilled in the tripartite pillared buildings.

6. Entrances are almost always in the end walls of the center aisle. There is no need for windows in the building.

7. These storehouses were located in a fortified, public area, where the major economic activity of the community occurred. It was also advantageous for them to be sited, as most of them were, near a city gate so that the produce from the field could be easily transported to the facility.39

Based on these attributes—proper storage practices and the fundamental functional requirements of safe grain storage—it seems likely that the Israelite tripartite pillared buildings served as storehouses during the Iron Age.

It remains only to consider what may seem to be counterevidence- the mangers and the tethering holes.

Whether these shallow basins actually served as mangers has been questioned, although I think they probably did.40 But they are certainly not uniformly present in the tripartite pillared buildings. Even at Megiddo they are relatively scarce. A few mangers between the pillars do not suggest that all of these buildings were stables. The mangers were probably used in a few storehouses in conjunction with the tethering facilities to feed the pack animals as they waited to be loaded or unloaded.

And that suggests the function of the few tethering holes that were found in the pillars. They constituted short-term tethering facilities. Pack animals moving in and out of the building would have been tied there while being loaded or unloaded.

1. For a brief overview, see Oded Borowski, Agriculture in Iron Age Israel (Winona Lake, IN- Eisenbrauns, 1987); Zev Herzog, “The Storehouses,” in Beer-Sheva I, ed. Yohanan Aharoni (Tel Aviv- Tel Aviv Univ./Inst. of Archaeology, 1973), pp. 23–30; Yigal Shiloh, “The Four-Room House. Its Situation and Function in the Israelite City,” Israel Exploration Journal 20 (1970), pp. 180–190.

2. Herzog, “The Storehouses,” p. 25; Shiloh, “The Four-Room House,” pp. 183–184.

3. R.S. Lamon and G.M. Shipton, Megiddo 1 (Chicago- Oriental Institute Publications, 1939), p. 39.

4. Lamon and Shipton, Megiddo 1, p. 35.

5. Joseph Callaway and Lawrence E. Stager have both suggested that the archaeological evidence from the 12th/11th centuries B.C. may reflect a cultural expansion from the coastal plain eastward into the hill country. The appearance of tripartite pillared buildings first on the coast may support that thesis.

6. See W. Kleiss, “Ausgrabungen in der Urartaischen Festung Bastam (Rusahinili) 1969,” Archaeologische Mitteilungen aus Iran (AMI) N.F. III (1970), pp. 7–65; “Ausgrabungen … 1970,” AMI N.F. III (1972), pp. 7–68; “Die urartaischen Anlagen in Bastam nach der Grabung 1975,” AMI N.F. 7 (1974), pp. 107–114. It has also been suggested that the design may have originated in Egypt, although the evidence is rather weak. See also John S. Holladay, “The Stables of Ancient Israel,” in The Archaeology of Jordan and Other Studies, ed. L.T. Geraty and L.G. Herr (Berrien Springs, Ml- Andrews University Press, 1986), p. 111. Some very good scholars have recently concurred with Holladay’s conclusions—for example, Stager, “Archaeology of the Family in Ancient Israel,” Bulletin of the American School of Oriental Research (BASOR) 260 (1985), pp. 1–35.

7. Frederick J. Bliss, A Mound of Many Cities (London- Palestine Exploration Fund, 1894), pp. 95–96

8. James B. Pritchard, “The Megiddo Stables- A Reassessment,” in Essays in Honor of Nelson Glueck, Near Eastern Archaeology in the Twentieth Century, ed. James A. Sanders (Garden City- Doubleday, 1970), p. 274.

9. Volkmar Fritz, “Bestimmung und Herkunft des Pfeilerhauses in Israel,” Zeitschrift des deutschen Palästina-Vereins 93 (1977), pp. 30–45.

10. See, for example, A Badawy, A History of Egyptian Architecture (Berkeley- Univ. of California Press, 1966), pp. 198–230; Ramsay MacMullen, Soldier and Civilian in the Later Roman Empire (Cambridge, MA- Harvard University Press, 1963). See also, Col. George Grogham, Army Life on the Western Frontier (Norman, OK- Univ. of Oklahoma Press, 1958).

11. Ernst Sellin, Tell Ta’annak (Vienna- Denkschriften der Kaiserlichen Academie der Wissenschaften, 1904), p. 104.

12. P.L.O. Guy, New Light from Armageddon (Chicago- Oriental Institute Communications, 1931), p. 37.

13. During those years the function of the buildings was rarely called into question, although their date was seriously debated. See J.W. Crowfoot, “Megiddo A Review,” Palestine Exploration Quarterly 1940, pp. 268–276.

14. Reported by Herzog, “The Storehouses,” pp. 23–30.

15. Herzog, “Storehouses,” pp. 23–30.

16. Pritchard, “The Megiddo Stables.”

17. Yigael Yadin, “The Megiddo Stables,” in Magnalia Dei- The Mighty Acts of God, ed. F.M. Cross, W.E. Lemke and P.D. Miller (Garden City, NY- Doubleday, 1976), pp. 249–252. For a summary of this debate, see “Megiddo Stables or Storehouses?” BAR 02-03.

18. Holladay, “The Stables of Ancient Israel,” p. 11.

19. Larry G. Herr, “Tripartite Pillared Buildings and the Market Place in Iron Age Palestine,” BASOR 272 (1988), pp. 47–67.

20. Holladay, “The Stables,” p. 109, n. 5.

21. Holladay, “The Stables,” p. 123.

22. As Herr has pointed out; see “Tripartite Pillared Buildings,” p. 55.

23. Herr, “Tripartite Pillared Buildings,” p. 56.

24. Holladay, “The Stables,” p. 155.

25. Holladay, “The Stables,” p. 158.

26. Holladay, “The Stables, p. 158.

27. There is a vast body of literature on the subject of grain storage practices. One of the better works, and a text that I rely upon heavily, is Storage of Cereal Grains and Their Products, 2nd ed., J. Anderson and A. Alcock (Minneapolis, MN- Jones Press, 1969).

28. These categories are taken from Holladay, “The Stables of Ancient Israel,” in order to show that a different conclusion may be drawn based upon similar data.

29. See H.J. Barre, “Country Storage of Grain,” in Storage of Cereal Grains, pp. 334–335.

30. William F. Albright, “The Excavations of Tell Beit Mirsim III- The Iron Age,” Annual of the American Schools of Oriental Research 21–22 (New Haven- ASOR, 1943), p. 24.

31. Barre, “Country Storage of Grain,” p. 335 and fig. 12.

32. G. Rickman, Roman Granaries and Store Buildings (Cambridge- Cambridge Univ. Press, 1971), p. 234.

33. See R.E. Hamilton, C.J. Lynde, and R.K. Larmour, “Flour Storage in Bulk,” in Storage of Cereal Grains, p. 443.

34. See L.V. Barton, “Effect of Different Storage Conditions on the Germination of Seeds of Cinchona Ledgeriana Moens.,” Contributions to the Boyce Thompson Institute 15 (1947), pp. 1–10.

35. Barre, “Country Storage of Grain,” p. 293–297.

36. Rickman, Roman Granaries and Store Buildings, pp. 293–297.

37. This practice is known as opus signinum. An example of this flooring method can be found in the construction of the Grandi Horrea of Ostia. See G. Calza, “Gli horrea tra il tevere e el decumano, nel centro di Ostia antica,” Notzie degli Scavi di Antichita (1921), p. 371.

38. The storehouses at Masada, Ostia and Rome are primary examples of the standard rectangular granary width.

39. Herr’s argument that storehouses should be located in non-public areas because of security reasons is weak (p. 53). Even a cursory glance at ancient urban planning in Rome, Ostia, Egypt and elsewhere indicates that many storehouses were located in public areas. See Rickman, Roman Granaries and Store Buildings.

40. Stager, “The Archaeology of the Family,” pp. 1–35, points out that the stables at the post-lron Age site of Kurnub support the belief that these basins served as mangers. Abraham Negev, the excavator of Kurnub, describes the basins in the stables- “The western and eastern walls of the central hall contain doors and four ‘arched windows.’ Mangers built into the sills of the ‘windows’ indicate that the elongated rooms served as stables.” See “Kurnub,” in Encyclopedia of Archaeological Excavations in the Holy Land, ed. M. Avi-Yonah and E. Stern (Englewood Cliffs, NJ- Prentice-Hall, 1977), vol. 3, pp. 722–735.