Raze of Glory:Interpreting Iconoclasm at Edfu and Dendera
Around 700 images published by the Institut Français d’Archéologie Orientale were examined to investigate the iconoclastic attacks at the Temple of Edfu and Dendera. For each scene, the degree of destruction was systematically recorded to reveal any potential prejudice against gender, appearance, or particular body parts. The selective mutilation of the feet, hands, and facial features can be attributed to practical factors as well as their symbolic significance. At the same time, there was no discrimination against gender or animal-headed deities at either Edfu or Dendera. The variations in hack marks observed at both sites were addressed with numerous illustrations. The frequent occurrence of multiple hack marks on singular scenes was interpreted as a choice with both practical and ritualistic implications. In addition to simplistic notions of religious hatred, intrinsic reward and personal gains played an equally salient role in motivating iconoclastic acts.
The Ptolemaic temples at Edfu and Dendera are among the best-preserved monuments in Egypt. Some have even described them as “virtually untouched,” as hardly any building material has been removed for reuse.1 But a modern day visitor to Edfu and Dendera would immediately notice the scars and chisel marks made by iconoclasts. As Sauer remarked, none of the relief would have survived if not for the sheer scale of the monuments and the enormous number of carvings.2
The Temple of Horus at Edfu lies on the west bank of the Nile, approximately 100 kilometers south of Luxor.3 The first temple to be commissioned during the Ptolemaic era, its construction began in 237 bce and was completed in 57 bce.4 The sanctuary is dedicated to the falcon deity Horus of [End Page 89] Behdet, who eventually became identified with the widely familiar Horus, son of Osiris.5
The Temple of Hathor at Dendera lies 60 kilometers north of Luxor on the west bank of the Nile and was constructed between 125 Bce and 60 ce.6 The temple is dedicated to the goddess Hathor, although Horus and Isis too have prominent roles.7 The two sites are connected by the Feast of the Joyous Union, when the statue of Hathor at Dendera was transported to Edfu to celebrate her marriage with Horus.8 Nevertheless, the distance between the two sites suggests that they should be approached as two separate loci of iconoclasm.
Like most iconoclastic acts, the image destruction seen at Edfu and Dendera are extremely difficult to date.9 Theories regarding the responsible agent revolve around several main groups: inhabitants of the temple after the decline of Ancient Egyptian religion might feel the need to remove the threat of the “evil eye,” whereas the domination of Christianity and Islam during subsequent centuries could have prompted the removal of pagan representations.10 It is most likely a religiously-motivated act, as a fear of the evil eye can hardly be accountable for the scale of destruction at both sites.
The responsible agent remains disputable, although the general consensus points towards the Coptic church. The Christian monks’ prominent role in temple iconoclasm is well-known, to the extent that monastic involvement can often be assumed even when there is no evidence for it.11 Image destruction at Edfu and Dendera often involved the lower registers—indeed, in certain areas the bottommost reliefs were the only damaged parts of the wall. Sauer12 proposed this as an indication that the attacks had taken place before the build-up of sand layers, which eventually buried the Temple of Dendera in its entirety until it was excavated at the turn of the twentieth century.13 In the absence of any viable alternative, the “sands of time” have become the most widely accepted dating method, making it improbable that the attacks were committed by Muslims after the Arab conquest. The interpretations in this work are based upon this assumption.14 [End Page 90]
Traditional Egyptian religion was already on the wane during the Roman period.15 Bagnall demonstrated the contraction of cults and temple constructions, as well as the absence of evidence for the continuity of festivals.16 By the third century, the temple economy had become obsolete and sacred scripts were virtually extinct.17 Edfu and Dendera’s locations in Upper Egypt might have delayed this process, but epigraphical records of pagan worship eventually ceased at Philae (“one of the last bastions of pagan worship”) around 456 ce.18
The exact end-date of Ancient Egyptian religion is widely disputed. Frankfurter maintains that the religion survived long after the third century, while “belief” could have persisted even after the end of temples and priests.19 Furthermore, there is nothing to stop traditional practices from being carried out in the temples even after the sixth century.20 At the same time, certain groups—much like modern day tourists—might continue to be attracted to them for a variety of reasons.21 Temples remained the de facto centers of the landscape, “the soul of the countryside.”22
Christian Attitudes towards Idol Worship
Christianity arrived in Egypt in the first century ce, and gained momentum when Emperor Constantine (reigned 306–337) began supporting it with state power and resources.23 The Roman Empire became increasingly intolerant of pagan cults during the fourth century, a process that reached a high point in Egypt in 391 with the destruction of the Serapeum at Alexandria.24 Most pagan sites that escaped complete annihilation were adapted as monastic residences, including the northern part of the Dendera temple complex.25
The prohibition against idol worship was grounded upon Judaic tradition, specifically in the Avodah Zarah.26 This was explicitly repeated throughout the Hebrew Bible: “thou shalt not make unto thee any graven image, or any likeness of any thing that is in heaven above, or that is on the earth beneath, or that is in the water under the earth; Thou shalt not bow down thyself to them, nor serve [End Page 91] them.”27 Idols are to be avoided and destroyed,28 and images carved in relief, like those at Edfu and Dendera, are forbidden as well.29 When confronted by idols, followers of Judaism—and in turn Christianity—are commanded, “thou shalt utterly overthrow them, and quite break down their images … ye shall destroy their altars, break their images, and cut down their groves.”30 Idols are also attacked on a philosophical basis. The divinity of objects like statues and images is questioned, for example, when Isaiah affirms that they are not gods, but merely “the work of men’s hands, wood and stone.”31 Idols are powerless as gods, for they “neither see, nor hear, nor eat, nor smell.”32
This is not to say that late antique Christians regarded pagan deities as non-existent, nor that they refused to associate idols with supernatural power. Instead they believed dangerous demons inhabited images and temples—despite their far inferior power compared to that of God and his saints, acts of worship and sacrifice could still endow them with strength sufficient to harm human souls.33 Destroying the temples and images became a means to prevent any further worship. At Edfu, some images of Horus were “Christianized” with the inscription of a cross or the name “Ephiphanios”—possibly the signature of a proud iconoclast.34 Indeed, episodes of image destruction played a prominent part in Coptic literature and biographies of saints as a deed to which followers should aspire.35 The best-known instances are those carried out by the abbot Shenoute in Upper Egypt during the fourth and fifth centuries.
The possible reasons for iconophobia and iconoclasm are numerous, and many have been well-researched. Religious texts often play a major role in affecting attitudes, but they rarely explain iconoclastic action in full. Further motivations will be examined in this chapter, in relation to the evidence from Edfu and Dendera.
It is clear that Egyptian ritual practice did not cease even during the period of Christian domination. Abandoned temples containing images remain a privileged site of interaction between human beings and the demonic gods.36 [End Page 92] Antony the Great was known to have spent “most of his times in the temples” during the fifth century.37 Early Christian literature talks about malevolent demons thought to inhabit images, and how the worship and sacrifice attracted by them helped maintain their malignant power.38 This recognition of the images’ potency fueled the need for iconoclastic behavior. In their cross-cultural overview of iconoclastic acts, McClanan and Johnson demonstrate how iconoclasts often possess some theory that relates images to the power they represent and seek to undermine that connection39 Thus despite biblical notions of the powerless idols,40 with their attacks iconoclasts were acknowledging that the myriad of relief images possessed significance and powers.
Image destruction can also act as punishment towards the human or deities represented—in ancient Egyptian contexts, the destruction of an image equaled the death of its referent.41 On the other hand, the destruction of pagan images can also be related to Christian concepts of purification and exorcism.42 Images can thus be “baptized” through the inscribing of a cross,43 an act which on some level can be interpreted as an acceptance of pagan idols into Christian life.44
More pragmatically, the obliteration of images serves as a statement against pagan worshippers. Much like vandalism, iconoclasm is fundamentally a public act aimed at the widest possible audience.45 The obliteration of pagan images serves to demonstrate the errors of pagan religion by countering the popular view that the destruction of idols would lead to dire consequences.46 The conversion of temple to church was depicted by Shenoute as “not merely the reorientation of a holy place but as a transfer from demonic evil to spiritual truth.”47 Ideally, this conversion of space would then lead to the conversion of people. This helps explain why the outline of the image is usually preserved, for this partial destruction allows observers to identify its original form.48 [End Page 93]
In the words of Rambelli and Reinders, “whereas interpretation can be widely varied, malleable and open-ended, physical evidence is relatively stable—all sides may agree that the icon’s head has come off, but not on what it means.”49 The goal of this work is to amass objective data through the systematic quantification of iconoclastic damage. One can hardly find a richer hoard of such physical evidence than at Edfu and Dendera.
Two types of data form the basis of this study. On-site observation provides an overview of the degree of destruction at the two sites. The second and most important sources are the publications made by the Institut Français d’Archéologie Orientale (IFAO). Regarded as “probably the most magisterial publication ever produced by a single Egyptologist,” Chassinat’s Le temple d’Edfou consists fourteen volumes totaling over 3,000 pages.50 This includes about 500 plates which serve as a photographic record of carved reliefs from Edfu. The plates were intended to record scenes on temple walls and, as a by-product, captured the effects of the iconoclastic attacks. More than 200 other plates from Dendera were made available with the subsequent publication of Le temple de Dendara.
Due to the sheer scale of the sites, the scenes in the publications are by no means exhaustive. Nevertheless, they represent a substantial sample size encompassing the various parts of the structures. As Chassinat was obviously more interested in the rituals depicted on the reliefs than the patterns of mutilation, there could be a concern that his archive may not be representative of the iconoclastic damage. Nevertheless, Chassinat did state that he strove to reproduce the liturgical scenes and texts regardless of the state of conservation.51 We therefore assume that the sampling strategy (though perhaps not completely random) does not discriminate against damaged scenes and is therefore suitable for use in this study. The on-site observations also offset potential deficiencies in the published sample.
Each of the published scenes was ordered by location and register: from an iconoclast’s perspective, scenes on the bottom are the most accessible, with the higher registers requiring extra determination and apparatus to reach. Practicality is not the only consideration—certain figures might be given more attention based on their gender or appearance.
Most interior walls at Edfu and Dendera contain two to four registers each (Fig. 1), in addition to the bottommost base (soubassement). Each register is around 2 meters in height, and the base about 1.5 meters. Today, the [End Page 94] average person, tool in hand, would barely reach the upper body of images on the first register at full stretch. That the soubassement at most areas of the temple were attacked suggests that the build-up of sands was hardly significant at the time. Therefore, iconoclasts would likely have required ladders or scaffolding just to reach the facial features of images on the first register.
When the iconoclast enters one of Edfu or Dendera’s many chambers, he is greeted by lavish scenes of kings, deities, and hieroglyphs. But the first images to [End Page 95] catch his attention would certainly be those of the animal-headed gods: initially because the figure of a deity is always sculpted to face the same direction of those who enter the room.52 More importantly, the unusual appearances of the animal-headed deities would only have stirred his resentment towards paganism.
Acts of iconoclasm are usually highly selective,53 more so in Egyptian temples where comprehensive destruction is virtually unattainable. We might expect the same process of prioritization at Edfu and Dendera—be it a powerful deity, a particular gender, or a specific body part. For example, female figures were found to have been selectively attacked in Theban tombs occupied by monks.54 Women were strongly associated with the demonic in Coptic texts, as attested by Theophilus’s famous statement: “it is through women that the enemy makes war against the saints.”55
Damage done to specific body parts provides hints to the iconoclasts’ intention and identity. Image destruction in most cultures tends to focus on the face, with specific emphasis on the eyes.56 In ancient Egyptian contexts, the mouth and nose are crucial in maintaining the ka of the deceased.57 Dismembered bodies from Gerzeh most commonly had their feet, heads, and hands removed, possibly a ritual act to impede the movements of the deceased.58
Using the images in the IFAO volumes, this author recorded each individual in scenes with mutilation by its gender, form59 (that is, whether it was an animal-headed deity), as well as the body parts that were mutilated. In total, 704 plates were examined. Individuals from completely unscathed scenes were omitted to minimize the discrepancy in sampling strategies between the two sites.60 The tallied results display how frequently each body part has been attacked.61 [End Page 96]
On-site Observation: Locations of Destruction
The extent of damage is far greater at Dendera, although its rooftop and crypts have largely been spared. For both sites, attacks focused on the main areas of the temple, with less destruction on the surrounding chambers. One can almost trace the iconoclasts’ movements through the tripartite temple as they entered through the courtyard, then moved on to the hypostyle halls and finally branched out to the smaller chambers. The dark soot marks on the ceilings suggest that these were also the main areas that were inhabited by (possibly Coptic) occupants following the decline of Pharaonic religion.62
The most heavily damaged areas are usually well-illuminated sectors in the open air. The best example is perhaps the Wabit (20) at Dendera, where each body part of every image has been ruthlessly destroyed by iconoclasts. At Edfu, the Courtyard, Sun Court (13), and hypostyle halls were among the most heavily damaged areas due to the brightly-lit environment.
Also at Dendera (Fig. 2), Chambers 3 and 5 are noticeably the most heavily damaged amongst rooms adjacent to the Second Hypostyle Hall. It is hardly a coincidence that these are the only rooms with access to an external door, which illuminated an otherwise obscured area of the temple. Meanwhile, it is also striking that the chambers numbered 13 and 17 were hardly attacked as compared to neighboring rooms. These two chambers are inaccessible through the main passageway, which might have hampered the transport of ladders or scaffoldings.
In contrast, the Chapel of Min (1) at Edfu escaped image destruction despite the heavy presence of phallic symbolisms (Fig. 3). The fertility god Min is depicted with an erect phallus, so much so that he became associated with the Greek god Pan during Ptolemaic times.63 Perhaps the most reviled pagan god in Coptic literature, Min/Pan was wittily described as having “a [End Page 97] heart as hard as his shame” by the abbot Shenoute.64 It is remarkable that the iconoclasts should decide to spare this particular chamber, after having attacked areas of the temple less offensive to late antique sensibilities.
It is possible to offer alternate explanations for the phenomena at the Chapel of Min—for example, iconoclasts might have found the images so repulsive that getting close to them was simply undesirable. However, this is more likely a result of iconoclasts valuing accessibility and practicality above other factors. At Edfu the iconoclasts failed to extend their efforts to most of [End Page 98] the remote chambers; while even their tireless counterparts at Dendera too have forgone the poorly-lit rooms. Despite the colossal and uncompromising nature of the task, it was the most ordinary of impediments which ultimately shaped their choices.
[End Page 99]
The Published Sample
Although the damage at Dendera was far more severe, the two volumes captured samples with similar proportions of damaged scenes (Figs. 4 and 5). This was likely caused by the inclusion of the Dendera crypts, which escaped iconoclastic attack entirely. Undamaged scenes will be excluded from quantitative analysis.
Note that a number of the plates featured both damaged and unscathed individuals—the iconoclasts seem simply to have decided not to finish their [End Page 100] job on the specific scene (Fig. 6). Sauer proposed the explanation that certain representations on higher scenes might not be easily reachable on ladders.65 While this is certainly plausible for examples from Dendera, it does not explain why most scenes from Edfu that involved only partial destruction are located at ground level.66 Given the lengths to which the iconoclasts went in destroying sites of such scale, it is somewhat puzzling that such figures would have been left out because of carelessness.
Relief images and texts possess a very intimate relationship in ancient Egyptian culture,67 but this understanding is unlikely to be have been shared by the iconoclasts. It is safe to assume that they did not read hieroglyphs, as even Egyptian priests could no longer master the script by the third century ce.68 [End Page 101]
[End Page 102]
[End Page 103]
Iconoclasts probably perceived hieroglyphs as images rather than texts. The vast majority of destruction occurred on the bandeau texts immediately above the soubassement. Signs in the form of humans, animals (Fig. 7), and even inanimate objects (Fig. 8) have all been subjected to destruction in various degrees.
Due to their higher positioning, texts on the registers were rarely attacked. Iconoclasts were highly selective on the few occasions when they did erase hieroglyphs, a fact which offers insight into their priorities.69 The vast majority of hieroglyphs targeted were those in animal form, particularly bird signs.70 Anthropomorphic determinatives were sometimes singled out.71 Snake signs such as the cobra and horned viper were rarely attacked, probably due to their unobtrusive shape and size (Fig. 9).72 Hieroglyph erasures are usually limited to the bottom part of scenes (Fig. 10).73 Furthermore, the miniscule signs can be completely unreadable in darker areas of the temple, rendering their presence inconsequential to the iconoclasts. The effects of lighting on the process are clear from the fact that most hieroglyph erasures occurred on the well lit enclosure walls.
The Body Parts
The position and sheer size of the Egyptian headdress made this the least practical part to destroy. The crown is hardly a threatening part of the image, and was likely not regarded as a component of the body. Nevertheless, some did find it worthwhile to mutilate the crown,74 particularly at Dendera (Fig. 11).
The notion of decapitation is inappropriate for iconoclastic actions at both sites, as attacks were usually concentrated on the facial features and rarely extended to the neck. The face is a powerful and expressive feature of every image,75 making it unsurprising that it was the most-targeted body part at [End Page 104] both sites. However, it was not uncommon for iconoclasts to neglect this component in lieu of other parts of the body. At figure 12, for example, the facial features of the individual on the left were left untouched, although the area underneath his outstretched arms has been comprehensively mutilated.76 At Dendera the facial features were overlooked almost as often as the upper limbs. Signs of destruction can sometimes be found on the hair—but this might simply be a case of misplaced blows intended for the facial features, as the two are seldom found on the same figure.
In general, the upper and lower limbs seem to hold similar levels of significance for iconoclasts at both sites. The hands and the upper limbs can be treated as a single entity, but complacent iconoclasts often neglect the upper arm to avoid extra work. The same reason accounts for the higher frequency of mutilation of the feet as compared to that of the upper leg. Indeed, most of [End Page 105]
[End Page 106]
the iconoclasts at Edfu were content with removing just the hands and feet. The prevailing norm of face and limb erasure was also applied to nonhuman images, as for example at figure 13. This further highlights its popularity and efficiency as a method of disfigurement.
The Rest of the Body
Assessing destruction using binary data is admittedly a simplistic approach, as it does not account for the level of exertion. Images destroyed in a similar fashion might require very uneven amounts of labor—one iconoclast might meticulously chisel every inch of the figure, another might be content to scatter dispersed blows. While special significance can be assigned to the limbs and facial features, the rest of the body tends to be somewhat of an afterthought—it is large in area, with little symbolic value. In the absence of other feasible methods, this component represents the best measure of the iconoclasts’ effort level. As already suggested, those at Dendera were a far more dedicated bunch—they have mutilated this somewhat negligible component of the image four times as often as their Edfu counterparts. [End Page 107]
Comparison between genders produced similar results across all body parts. The frequency of attack is marginally higher for male individuals at Dendera, but slightly lower at Edfu (Fig. 14). There were examples in which female figures seem to have been singled out,77 but in other instances the opposite was true.78 Overall, the iconoclasts do not seem to have discriminated on the basis of gender. The only notable disparity is in crown destruction—while examples of crown erasures at Edfu are sparse, the sample from Dendera is significant enough to demonstrate that the crown of female individuals was attacked twice as frequently. In particular, the headdress of the goddess Hathor seems to have been a prime target for the Dendera iconoclasts.79 [End Page 108]
Whether we assume that the iconoclasts’ intention was to destroy pagan iconography or to nullify the magical threat of the reliefs, it seems reasonable that the animal-headed deities might have been singled out, owing to their unusual and supernatural appearances. However, across all body parts, the mutilation rate for such figures was very similar to the norm for other individuals (Fig. 15). In some instances, it was the animal-headed gods—amidst other figures—who came away unscathed from the iconoclasts’ attacks. In figure 16, for example, the falcon-headed god Horus was spared by the iconoclasts while the strictly anthropomorphic male figure on the left was attacked.80 There is definitely no indication that the iconoclasts were discriminatory against the appearance of individual figures. [End Page 109]
Height as a Limiting Factor
It should come as little surprise that the first register tends to be the most heavily-destroyed area in many of the rooms with in the two sites. While additional contraptions might still be necessary to reach the upper part of the first register, this could be achieved safely with relative ease. On the other hand, [End Page 110] reaching the registers above was a tall task indeed, with the potential for a fatal fall.81 Thus, in figure 17 the iconoclasts have painstakingly destroyed the scene on the third register (below), but did not manage to reach the register above. There are also various impediments that come with maneuvering at such heights—on a ladder, for example, the iconoclast’s reach would be limited by the need to maintain his balance, which could have resulted in certain body parts being left out.
Nevertheless, in instances where the iconoclasts have gone to the lengths of reaching the higher registers, destruction has occurred at a far higher rate and in a more thorough manner (Fig. 18). This perhaps implies that scaffoldings were employed to reach the higher reliefs, as indicated by documentary sources on the image destruction at Philae.82 Understandably, the only component with a higher destruction rate at the first register is the crown.
Other Observations: Untimely Ends
Although the iconoclasts appear to have been quite scrupulous in their selection of body parts for mutilation, there were occasions when they seem to have been rather careless. In figure 19, for example, iconoclasts have left out the feet and one arm of the female figure to the left.83 Sometimes identical individuals within the same scene were mutilated to very different degrees, or even ignored entirely. In figure 20, the individual in the middle was left completely undamaged despite its identical appearance to other figures on the same scene.84 While destruction of the face and limbs was the overriding feature of both sites, it is questionable how much symbolic significance can be attributed to these body parts, given how frequently they have been nonchalantly overlooked. For example, distinct features such as the bovine ears of Hathor85 and the eyes86 have often been preserved. There seems to be a certain degree of flexibility in the iconoclasts’ choices, and the omission of any particular component did not elicit concern.
The garments worn by depicted individuals also played a part in the iconoclasts’ decisions. The long dresses worn by women and goddesses87 conceal [End Page 111]
[End Page 112]
the lower limbs, often saving these body parts from mutilation. In figure 21, the lower limbs of the deity (right) and king (middle) have been entirely destroyed, but the long dress worn by the female individual (left) meant that only her feet were singled out by the iconoclasts. Attacks on clothed parts of the body (as well as the crown) may have been perceived as redundant, whether for practical or cultural reasons. It also suggests that iconoclasts did not transfer their hostility against pagan worship towards the Pharaonic culture it was linked to.88
The “castration” of male individuals usually occurs separately from the mutilation of other body parts. The best examples can be found at the Chapel of [End Page 113]
[End Page 114]
[End Page 115]
Min at Edfu (Fig. 22).89 However, the castrations were limited to the first register, and no damage was made on the other body parts of any individual. The phalli were gouged out in their entirety, leaving a deep, distinctive gash unlike the small and dense chisel blows made by iconoclasts. For this reason, it is imperative to treat this as damage of a completely different nature.
The agents behind the castrations were probably individuals intending to “harvest” phalli for use as an aphrodisiac,90 resulting in depressions known as fertility gouges or pilgrim’s gouges.91 At Dendera, where depictions of Min are harder to come by, some even resorted to using the non-erect phalli of mortal men.92 Dating the castrations in relation to the iconoclastic attacks is problematic, as the two acts might be essentially unconnected.
If anything, the iconoclasts seem to have been reluctant to partake in the removal of phalli.93 On a scene from Edfu where most body parts have been attacked in familiar fashion, the erect phallus of Min was inexplicably ignored.94 This disinclination on the part of the iconoclasts can also be observed in scenes depicting female nudity.95
The appetite for phallic images was great—to the extent that the Egyptian apron (worn by men underneath a kilt) was sometimes mistaken for a phallus and removed (Fig. 23). Although the apron was a traditional garment dating back to the Old Kingdom, Egyptians of Late Antiquity might no longer have recognized or understood the nature of such costumes.96 Like examples of “true” castrations, the phallic projects between the legs of the figures were usually gouged out instead of being meticulously chiseled. False castrations are observed mainly at Edfu, although similar examples can also be found at Dendera.97
Unexpected Hack Marks
In some instances we see blows scattered over a relief away from any targeted individuals.98 It may be that those are hard-to-reach areas of the scene where precision is unattainable. The erratic nature of the desecration marks is perhaps reminiscent of the story of Anoub, the desert father who performed a [End Page 116]
[End Page 117]
The iconoclasts’ fixation with anthropomorphic images as well as certain body parts produced a degree of uniformity that is easily recognizable at both sites. However, unexpected attacks on inanimate objects are not entirely rare. Examples of such objects include scepters,100 cartouches,101 and plants.
More often than not, the ankh signs held by figures have escaped erasure by the iconoclasts.102 The preservation of the ankh seems to have been intentional, as [End Page 118] iconoclasts often took care not to damage the sign while mutilating the hand where it was held.103 A lot has been written about the possible relationship between the ankh and the development of the Coptic cross, particularly its similarity with the Christian Staurogram.104 This again hints at the involvement of Christian agents, as iconoclasm committed following the Muslim Conquest is unlikely to have spared an object that resembled Coptic iconography so closely.
Examining the Hack Marks
Egyptian temples are an iconoclast’s worst nightmare—whereas Graeco–Roman temples can be emptied with relative ease by the removal of statues, every inch of an Egyptian temple is covered in pagan images and hieroglyphs.105 Theoretically, the two-dimensional reliefs may be of secondary importance in regard to the three-dimensional sculptures that offer the direct opportunity for worship.106 In any case it was the cult statue in the naos that was most prized by iconoclasts, the “image d’action”107 which could be humiliated in public before being dramatically destroyed in a blaze of glory.108 In contrast, those tasked with mutilating the carved reliefs of the temple would find their undertakings far less rewarding and much more laborious. Most hack marks found at Edfu and Dendera are very fine and concentrated, with hundreds of blows required to mutilate each individual image.
A great variety of tool marks can often be found on the same scene109 and sometimes even on the same individual figure.110 In figure 25, for example, small puncture holes were made on the male individual, whereas long lacerations were used to mutilate the goddess on the right. Individual factors such as fatigue, change of tools, and change of postures could have played a part in producing such patterns, but they can hardly account for the range of variation we often observe on the same relief. The different marks were probably made by separate individuals, perhaps with the use of differing tools.
A number of erasure techniques has also been observed by Roth at Hatshepsut’s mortuary temple at Deir el-Bahri. It was suggested that this [End Page 119] resulted from several programs of destruction with different goals, carried out in multiple stages: “literate scribes marked the hieroglyphs that were to be deleted with scratching; the silhouettes were chipped in a first pass meant to level the raised relief; and the rectangular roughening blended the relief with the background surface, which was then smoothed and finally recarved.”111 However, such a sophisticated approach would not be required at Edfu and Dendera, where the sole objective was the obliteration of pagan images. While different tool marks are to be expected with the involvement of multiple personnel during a large-scale attack, this does not explain the variety often present on a single scene. Therefore, the different hack marks at Edfu and Dendera are likely the result of multiple iconoclasts working on the same scene—possibly even side-by-side.112
[End Page 120]
A non-exhaustive list of the different hack marks observed at Edfu and Dendera can be found at: muse.jhu.edu/article/618918/file/supp01.pdf. In general, they can be loosely classified into four groups, the first two being the most common by far:
a). Rounded punctures
The rounded marks were made using a pointed tool, likely accompanied by a mallet. These are very easily identified, with only slight variations in roundness and depth. Each blow produces only a small dent on the image, making this possibly the most laborious method. Oblong marks can also be made by using the same tools at an oblique angle.113
b). Long scratches
Long scratches are very commonly found at both sites and were most probably made using chisels or adzes. Usually this produces downward, diagonal marks at about 45°, but vertical or horizontal marks can also be found. The use of a mallet would help, which means it might be possible to deduce handedness and changes in posture based on the direction of diagonal scratches (Fig. 26).
c). Ragged, irregular lacerations
These are characterised by deep and violent lacerations, often resulting in grotesque images. Compared to the hundreds of uniform blows required by the previous two methods, this was undoubtedly a much more powerful and thorough way of effecting mutilation. However, the ragged marks are very uncommon at both sites, and usually found only on raised reliefs.
Some of these marks can be a combination or variation of the ones explained above, but the high density and irregularity makes identification problematic.
Iconoclastic actions are often tagged with a negative notion of irrevocable loss. However, destruction also created new meaning and practices—take the toppling of the Berlin Wall or the statue of Saddam Hussein at Firdos Square, which many would remember as the creation of new eras rather than mindless destruction.114 As one landmark falls, a landmark moment is often established.
Destruction is a cultural and religious activity capable of transforming the status of the agents involved.115 Men can be immortalized by their [End Page 121]
[End Page 122]
creations, but history also remembers those who destroyed. Early Coptic literature is filled with formulaic stories of iconoclastic acts: “an influential holy man leads a group (usually monks) to a place or object of Ancient Egyptian worship (an idol or temple), and destroys it. Afterwards, many people are converted to Christianity.”116 For highly-ranked members of the monastery, image destruction is a legacy to uphold. Shenoute defended his violent attacks on pagan temples as merely complying with the laws of Theodosius’s successors to destroy pagan monuments that the latter did not manage to eliminate during his lifetime.117
Subsequently, in the Life of Moses of Abydos, the dying Shenoute was made to predict that the soon-to-be-born Moses would carry on his work to abolish pagan practices and overturn their temples. Moses did not disappoint, courageously leading an all-night session of anti-demonic prayer to drive out the pagan god Bes from an abandoned temple. When the Bishop Macarius was captured by pagans and about to be sacrificed, Besa (another successor of Shenoute’s) was alerted by a vision sent by the abbot from his deathbed. He arrived in the nick of time, opened the locked temple doors miraculously and rescued the bishop, before setting fire to the temple.118 Every detail of Besa’s heroic adventure mirrored Shenoute’s famous raid of the house of Gesios, an Egyptian Libanius believed to be a worshipper of pagan idols.119
While all glory belongs to the leader, the others involved may also stand to gain from their exploits. If—as the literature suggests—the iconoclasts were members of the monastery, image destruction could be considered part of an ascetic lifestyle that ultimately led to personal salvation. Asceticism became prominent in Christian Egypt during the third and fourth centuries and usually involved extreme forms of withdrawal from society.120 We hear of monks subjected to punishing levels of manual labor and sufferings, which Shenoute regarded as the most powerful means of defeating the devil and gaining salvation.121 Image destruction could also serve as a means of achieving Christian martyrdom.122
Physical encounters with demons are a rare occurrence in Coptic literature, even for someone of Shenoute’s stature. Given that most of the monks’ conflicts with the demonic consisted merely of thoughts and inclinations, iconoclastic [End Page 123] attacks on the pagan images became the closest form of physical combat with demons in both figurative and literal senses.123 Through these symbolic battles the monk could assume the role of a “fighter” or “combatant”—a far cry from his monastic life, which hardly conformed to traditional Roman notions of masculinity.124
By the fifth century, attacks on the demonic pagan gods had become an important part of the monks’ identity.125 As the headquarters for conspiring demons, the temples provide the perfect settings for monastic struggle and triumph.126 For individual monks, the temples marked the start of a journey filled with adversities, with the eventual end-point of self-mastery.127
In a period when dissent and divisions among Christian groups were common, iconoclastic attacks served to unite the Coptic community.128 Temples came to provide the setting for holy men cleverly conspiring to create an in-group identity, all through the construct of a common enemy in the demonic pagan gods. Thereupon, it is hardly surprising that the spread of Christianity in Egypt was fueled directly by the destruction of pagan monuments.129 This fascinating dynamic made the pagan gods an indispensable element of Coptic monasticism, the perfect adversary in a voluntary struggle.130
A Sensory Experience
The dramatic stories of Shenoute and his followers paint a very different picture from the mundane chiseling at Edfu and Dendera. Indeed, the enclosed setting of an Egyptian temple rather contradicts iconoclasm’s need for publicity and performativity.131 Yet it would be unreasonable to perceive these iconoclastic acts as sterile, monotonous events. On the contrary, it is likely that they would have produced a stimulating atmosphere more akin to liturgical rites.
Like all rituals, Christian liturgical celebrations were multimedia events in which the liturgical environment, architecture, sounds, and gestures all [End Page 124] played prominent roles.132 The architectural setting of a dim, secluded cult temple certainly provided a highly unusual sensory environment. Perhaps the chiseling was accompanied by utterance—the use of spells and sermons to attack and neutralize demonic images has been well-documented in Coptic literature.133 Meanwhile, the variation of hack marks on single scenes suggests that a number of iconoclasts were working simultaneously, probably in close proximity.134 In such an environment, the sound of metal on sandstone could have produced effects not dissimilar to the acoustic dimension of lithic production, described by some as “unique and memorable” or even musical.135 Regardless of the tedious nature of the task, those involved would have been part of an extraordinary sensory experience befitting its ability to alter the status of each participant. It was, for all intents and purposes, a rite of passage.
The perpetual Western bias towards the visual sense has been widely acknowledged.136 Indeed, this study has been guilty of placing great emphasis on the tool marks—mere visual remnants of what might be a far more elaborate event. The persistent use of terms such as “images” and “scenes” undermines the temple’s richness in sensory aspects, whether when they were still in use for worship (incense, processions, reading of texts) or when the iconoclastic attacks took place.
Even our visual impression is highly defective, due to the fact that color has faded away from the vast majority of the reliefs. In ancient times, with colors on pagan images still intact, the visual effects of image destruction would have been much more powerful. The blue bracelets and anklets worn by relief figures would have acted as cues to iconoclasts as to where blows should start and end. It is for this reason that we often see hack marks that do not extend beyond the wrists or ankles with great uniformity.
An emphasis on the ritualistic aspects of iconoclasm would help explain some of the more surprising findings of this study. Under Coptic contexts, the lack of prejudice against gender and form is indeed quite extraordinary. Stories and rhetoric often employ women to signify demonic conflict, with the female body representing seduction or resistance.137 Even female monks were given less power than their male counterparts.138 On the other hand, wild and [End Page 125] fantastic animals were strongly associated with demons due to their malice and ferocity.139 Arena combats with animals were frequently used by ancient authors as a metaphor for struggles against demons in the quest for virtue.140
If one could deemphasize the visual outcome of iconoclastic acts (how many scenes were destroyed, which depictions were being targeted), perhaps it was the process of their creation that is of utmost significance. Consequently, the attacked images need not be particularly “demonic.” Indeed, the data from Edfu and Dendera suggest that iconoclasts disregarded gender and appearance completely, making seemingly random choices on which images to erase. Prominent body parts that were inexplicably left out might have coincided with the end of a ceremony—if image mutilation is merely part of a ritualistic act, the “completion” of each scene becomes rather inconsequential.
It has already been noted at Karnak and other major sanctuaries that iconoclastic attacks often concentrated on the head and limbs of images.141 This pattern can now be tangibly demonstrated using the data from Edfu and Dendera. The limbs and facial features are the most expressive and threatening components of the body, but the singling out of these body parts was also based on practical considerations. Ideally, iconoclasts would prefer to erase each anthropomorphic image in its entirety, as they have done in certain parts of Dendera.142 Surprisingly, there was a lack of prejudice towards elements which are inimical to the iconoclasts’ ideologies, such as female figures, animal-headed gods, and the ithyphallic Min.
The main areas of the temples sustained the heaviest damage, a result of prioritizing illumination and accessibility. Height was a limiting factor, albeit not an insurmountable one—otherwise iconoclasts would have been better off concentrating on all scenes on the first register. Indeed, the thoroughness of image destruction on higher registers suggests the widespread application of sophisticated contraptions such as scaffolding. Destruction probably came to a premature halt at both sites—given adequate time, iconoclasts going the lengths to reach the higher registers would have been unlikely to overlook the inner chambers of Edfu, likewise the higher floors and roof at Dendera.
The iconoclastic acts at Edfu and Dendera are ritualistic not only in the sense that they are highly repetitive tasks, but also because of their ability [End Page 126] to transform the identity and status of the participants. While conventional explanations of halting pagan practice, countering the threat of “magical” images, and the conversion of people have undoubtedly been involved, they were hardly the be-all and end-all from the iconoclasts’ viewpoints.
The intrinsic value of image destruction was to be found beyond the surface evidence. It would be unwise to reduce the image destruction at Edfu and Dendera to solely a result of religious intolerance. Pagan images were far from being a mere nuisance to Coptic monasticism—a lot can be gained from their presence (and subsequent destruction), on both religious and personal levels. The variances in hack marks, abrupt ends, as well as the indifference towards gender and appearance suggest the possibility that the iconoclastic acts were ritualistic in the strictest sense. This is not to suggest that every chipped wall in Egypt had been part of a full-fledged liturgical celebration. More often than not, a destroyed image is merely the result of loosely channeled contempt and aggression. It is the scale and density that differentiates Edfu and Dendera from most iconoclastic evidence, providing room for analyses and interpretations. In archaeological terms it is analogous to comparing flints to megaliths.
The sensory aspects of iconoclastic attacks should not be ignored: sounds, colors, and the architectural environment are all integral components in temple iconoclasm. Hence the tool marks and disfigured images we now see are secondary products of a far more complex practice. For iconoclasts at Edfu and Dendera, image destruction was merely a means to an end—it was the participation process that constituted the ultimate goal.
I am grateful to Penny Wilson for her input and assistance. I thank Andrew Millard and the anonymous reviewers for their insightful comments, and the Birley Bursary for funding my field-work in Egypt. Heartfelt thanks also to Noel Lenski for his patience and editorial expertise. All remaining errors are mine.
27. Exod 20.4–6 (King James Version)
28. Exod 34.17; Num 33.52; Deut 7.5, 7.25, and 12.3 (King James Version).
30. Exod 23.24 and 34.12–14 (King James Version).
48. “A particularly complex and potent form of harm to sacred objects is disfiguring, because the disfigured object functions as a sign of both its power and powerlessness. In some cases damaged objects are kept in view, a display of the damage itself. If the object were totally destroyed or attacked to the point of being unrecognizable, it would be less useful in the construction of a ‘profanophany’” (Rambelli and Reinders 2007, 20).
53. For instance, during the sack of Nineveh, reliefs at the Palace of Sennacherib were mutilated to various degrees based on the individuals depicted. Damage is most severe to the king, followed by the crown prince and his two attendants, while the rest of the human figures were only slightly erased, see May 2012, 188–89.
59. A lay person would find little discrepancy between Amun and a pharaoh, or Isis and a queen. On the other hand, iconoclasts would find the falcon-headed Horus or crocodile-headed Sobek very distinctive indeed.
60. A good example would be scenes from the Dendera crypts, all of which escaped iconoclastic destruction. Hence a 50% frequency for facial features does not mean that half of all individuals at the site had their face mutilated, as the value for the entire sample or site will be substantially lower.
61. The fluid nature of the subject means that a set of criteria has to be established to minimize the inevitable intra-observer error.
• The crown is regarded as destroyed if 50% or more of the entire crown area was destroyed.
• Facial features are marked as attacked if more than 50% of the facial area has been mutilated.
• Regarding hands and feet, due to the small areas of these components, partial mutilation is almost never a problem. A special note is made where only one of the hands/feet was destroyed, but those are nevertheless coded as having been attacked during the final analysis (the same applies for the rest of the limbs).
• Regarding upper and lower limbs, the entirety of the limbs has to be mutilated to be considered attacked. This is necessary because hack marks on the hands and feet can often spill over to the limbs. The large area taken up by the limbs also meant that the mutilations are sometimes only half-completed, perhaps due to complacency—in those instances they are judged as not destroyed.
• As for the rest of the body, complete mutilation is required to be considered targeted, as spillover hack marks can make the recording procedure highly problematic.
66. Examples of scenes on the first register or soubassement with both damaged and undamaged individuals: Edfu pls. 260, 380, 411, 428, 429, 434, 576, and 651.
69. In contrast, non-selective erasure of hieroglyphs can be observed at Dendera pls. 620 and 624.
70. Other examples: Edfu pls. 215, 326, 492, 494, 497, 499, 501, 503, 505, 511, 512.
71. Examples: Edfu pls. 225, 499, 501, 550.
72. Snake signs were found to be mutilated at Edfu pls. 559, 646, 647.
73. Limitation of hieroglyph destruction to the lower part of the scene at Edfu pls. 213, 223, 495, 562, 660, 663. On rare instances this has occurred on the upper part of a relief, with the bottom half unscathed, e.g. Edfu pl. 507.
74. Some examples at Edfu pl. 650; Dendera pls. 53, 264.
76. Other examples where facial features escaped erasure include Edfu pls. 242, 346.
77. Dendera pls. 163, 184.
78. Dendera pl. 660.
79. For examples, see Dendera pls. 53, 264.
80. See a similar example at Edfu pl. 411.
83. For more examples, see Dendera pls. 57, 99, 648; Edfu pls. 236, 254, 589.
84. Further examples at Edfu pl. 229, Dendera pl. 598.
89. Similar examples can be found at: Edfu pls. 279, 331, 334, and 335.
92. See examples at Dendera pls. 342, 345, 407, 410, 533, and 538.
95. Dendera pl. 186.
97. Dendera pl. 170; Edfu pls. 462, 463, 471, and 502.
98. Examples of erratic and/or misplaced blows: Edfu pls. 67, 278, 500, 504, 518, and 519; Dendera pls. 182, 598.
100. Dendera pl. 188.
101. Dendera pl. 620.
102. E.g. Edfu pls. 213, 215, 327, and 560. Erased ankh signs can be observed at Edfu pls. 224, 553, and 554, although iconoclasts were inconsistent and left some of the ankh signs on the same scene intact.
103. Edfu pls. 213, 215, and 327.
110. For instance, hack marks found on the legs would be very different from those on the torso. Examples: Dendera pls. 78, 97, 151, and 166; Edfu pls. 327, 339.
112. This might have been necessitated by the use of scaffolding, which could accommodate only one scene at any given moment.
130. Brakke 2006, 226: “The desires of bishops and others to decisively defeat the demons/gods sat at odds with the monks’ need to engage the demons in a prolonged struggle to overcome their own deficiencies. Shenoute could not let his conquered and dismembered Satan die: the devil had to live on in his breaths, the evil thoughts that provided the material for the Christian’s efforts to gain virtue.”
134. It is perhaps not far-fetched to picture the zealots working alongside each other, chanting in unison with each rhythmic blow.
142. The Wabit being the best example.