Fig. 1. Doll-figurine, single piece of textile and palm cording, with pebbles creating the head, 11 × 13 cm. KM 26416 (not previously published.)
    Fig. 1. Doll-figurine, single piece of textile and palm cording, with pebbles creating the head, 11 × 13 cm. KM 26416 (not previously published.)

    Among the thousands of objects attesting to daily-life activities in the Roman town of Karanis in Egypt are several figurines made of textile and papyrus, dating from the first through fourth centuries CE (fig. 1). The Record of Objects, which catalogued the material excavated during the University of Michigan campaigns at the site in the 1920s and 1930s, recorded these items as “dolls,” and they have been known as such ever since. Calling these objects dolls, however, is likely to provoke very different assumptions about what they were used for and why they were made in the first place. A reasonable interpretation might be that the figurines were toys used by children; but, in contrast, consider the sub-Saharan African wooden and beaded dolls in the University of Michigan Museum of Art, which were created for use in fertility rituals. In the case of the textile and papyrus dolls from Karanis, is it even possible to offer any kind of perspective on the use-context and meaning of these archaeological items? While definitive answers will always remain elusive, there are some clues worth investigating.

    Of the nineteen doll-figurines recovered from excavations at Karanis between 1924 and 1935, fifteen are in the collections of the Kelsey Museum, one resides in the Cairo Museum, and three more are noted in the Record of Objects books, but their location today is unknown (see table 1).[1] An archival excavation photograph (fig. 2) illustrates five of these textile dolls (at the right end of the second row from the top); also of note in this photograph are several wooden dolls (across the top row). In fact, several types of figurines—made of stone, wood, bone, and terracotta, in addition to textile and papyrus—were classified as dolls in the Record of Objects books. Some of these have since been relabeled (e.g., the terracotta figurines are now treated as “female orants”), but the point remains that the category of doll has been used broadly. The definition of the corpus of dolls discussed here is based primarily on the use of plant and animal fibers, but also of relevance are their size and certain iconographic elements. Most of them will be mentioned below with regard to their physical features, but there is space to provide a detailed review of archaeological contexts only for three of them (KM 3647, KM 7512, and KM 26413); these examples were selected because each comes from a different temporal horizon at Karanis, thus offering a stratigraphic cross-section. The discussion then turns to a comparison of the textile and papyrus figurines from Karanis with other known examples from elsewhere in Egypt. Thus, the analysis of physical features, the investigation of archaeological context, and the guidance of comparative material, all acting in concert, provide a helpful lens for exploring possible interpretations of the use and meaning of these artifacts—interpretations that consider the figurines as operating within the world of both adults and children.

    Location Accession Number Excavation Field Number Length(cm) Width(cm) Fibers Present Fabric Structure Additional Notes
    Kelsey Museum 3541 25-241C-A 5.5 3.5 Wool, linen Weft-faced plain weave with dovetail tapestry with weft-wrapping
    Kelsey Museum 3542 25-295-D 9 4 Wool Weft-faced plain weave; one fragment is balanced plain weave Sand present
    Kelsey Museum 3543 25-262-NII 7.5 8.5 Wool, hair(human?) Weft-faced plain weave
    Kelsey Museum 3647 25-200X-A 3 1.5 Papyrus, wool Yarns and papyrus strips
    Kelsey Museum 3648 25-247-A 6.5 3.5 Wool Weft-faced plain weave, interlinking sprang
    Kelsey Museum 3649 25-185D-A 4.5 2.5 Wool Weft-faced plain weave
    Kelsey Museum 3856 25-262-B 6 3 Papyrus, wool Yarn and papyrus
    Kelsey Museum 7506 28-B168K-Q 9 3 Wool Weft-faced plain weave
    Kelsey Museum 7508 28-B133A-D 11 6 Wool Weft-faced plain weave Excluded from this analysis
    Kelsey Museum 7512 27-C54A-A 9.5 5 Wool, hair(human?) Weft-faced plain weave, 1/1 interlinking sprang Mud as structural component of head
    Kelsey Museum 7513 27-C19-X 11 12 Wool Weft-faced plain weave, tapestry and weft-wrapping in one fragment Wood stick attached as arms
    Kelsey Museum 7523 28-157*-H 7.5 3 Wool Weft-faced plain weave, weft predominant plain weave in one fragment
    Kelsey Museum 10113 25-262-NI 7 3.5 Wool Weft-faced plain weave
    Kelsey Museum 26413 26-B36A-J 8 3 Wool Weft predominant plain weave
    Kelsey Museum 26415 30-B224C-A 8.5 5 Wool Weft-faced plain weave
    Kelsey Museum 26416 33-4017*-V 11 13 Wool, palm fiber cording, some linen stitching Weft-faced and weft predominant plain weave Small pebbles used to create the head
    Cairo Museum 54788 28-242*-CIII Unknown Unknown Unknown Unknown
    Unknown None 27-242B-T Unknown Unknown Unknown Unknown Noted only in Record of Objects books
    Unknown None 28-165*-R Unknown Unknown Unknown Unknown Noted only in Record of Objects books
    Unknown None 28-218*-M Unknown Unknown Unknown Unknown Noted only in Record of Objects books

    Table 1. Doll-figurines recovered from excavations at Karanis, 1924-1935.

    Physical Features

    The Egyptian desert environment has left us a rich legacy of organic archaeological materials, including cloth and other fibers used in textile arts and industry. Such organic remains are important attestations that augment our knowledge of the lives of the inhabitants of Karanis. Cloth was a valuable commodity that was reused and recycled throughout antiquity, resulting in textiles surviving to us with complex histories. The most common materials used in Egypt during the Roman period were wool and linen, but it is wool—along with various other plant and animal fibers—that appears to have predominated at Karanis (Wipszycka 1965, 37–42; Thomas 2001, 13). The textile figurines are consistent with this observation. With the exception of the two figurines crafted from papyrus, all of the textile dolls are made primarily of wool and consist of a weft-faced weave.[2] There are a couple of noteworthy elements that provide some variation from this fiber content and fabric structure: linen stitching is seen on KM 3541 and KM 26416 (fig. 1), palm cording is used for KM 26416, examples of tapestry are seen on both KM 3541 and KM 7513 (fig. 3), while KM 3648 and KM 7512 (fig. 4) each display an example of sprang (a weave that results in somewhat “elastic” textiles). Weavers often specialized in the production of particular kinds of cloth in bulk yardage since this often required different kinds of looms (Wipszycka 1965, 36; Thomas 2001, 14). Thus, to see a variation in fabric structure used among the dolls reinforces the idea that the cloth was recycled. Additionally, since several of the figurines contain a selection of fabric structures and fiber types, it is more likely the case that these dolls were the products of manufacture at the individual or household level.

    Fig. 2. Assortment of Karanis artifacts including textile dolls from expedition photograph (5.3681). In the second row, from right to left are:  KM 7513, KM 7512, Cairo Museum 54788, KM 7523, Km 7506. Sixth from right is KM 7508, excluded from this study; note the considerable size difference and lack of human characteristics.
    Fig. 2. Assortment of Karanis artifacts including textile dolls from expedition photograph (5.3681). In the second row, from right to left are: KM 7513, KM 7512, Cairo Museum 54788, KM 7523, Km 7506. Sixth from right is KM 7508, excluded from this study; note the considerable size difference and lack of human characteristics.
    Fig. 3. Large example of doll-figurine, with wooden stick arms, 11 × 12 cm. KM 7513.
    Fig. 3. Large example of doll-figurine, with wooden stick arms, 11 × 12 cm. KM 7513.
    Fig. 4. Doll-figurine illustrating eyes, mud-hair-sprang component, 9.5 × 5 cm. KM7512.
    Fig. 4. Doll-figurine illustrating eyes, mud-hair-sprang component, 9.5 × 5 cm. KM7512.
    Fig. 5. Small example of doll-figurine, papyrus, 3 × 1.5 cm. KM 3647.
    Fig. 5. Small example of doll-figurine, papyrus, 3 × 1.5 cm. KM 3647.

    Although there is some consistency in the fiber type and fabric structure, the textile and papyrus figurines display much more variation in their size. On average, they measure about 7.5 cm in length and 5 cm in width, but these averages eclipse the two very small examples (the papyrus figurines, KM 3647 [fig. 5] and KM 3856) and one of the largest examples (KM 7513 [fig. 3]), which has wooden sticks attached to create the appearance of arms. The Karanis dolls are crafted on a small scale, but beyond this, what more can be made of their size?[3] It is reasonable to say that the size of objects, in many cases, is connected to their function, and it can also relate to their visibility, from the perspective of both the user and the viewer; in this regard, the Karanis dolls demand a certain amount of focused attention—perhaps on the part of a single individual—to be appreciated, whatever their function may have been. One further observation should also be made with regard to size: that is, not only are the Karanis dolls made on a small scale, but they could also be interpreted as “miniatures” when considered in connection to their iconographic elements.

    For all dolls within the corpus, there are recurrent iconographic details that are immediately evocative of human characteristics, thus supporting the notion that the figurines were meant to be, at least in some way, miniatures of the human form. The basic elements consist of several folded strips of cloth or papyrus fashioned around a firm textile nucleus, which are then bound together creating a form reminiscent of a head and limbless torso (KM 26415 [fig. 6]).[4] The clean folded edges comprise the front of the object, and the excess of the strips is left hanging. There are some variations: two figurines (KM 3649 and KM 7512 [fig. 4]) have a double textile nucleus, giving the impression of eyes; three others (KM 3647 [fig. 5], KM 7506, KM 26413 [fig. 7]) are bound, even excessively so, such that they are reminiscent of a baby in swaddling clothes; and one (KM 26416 [fig. 1]) consists of a single piece of textile enveloping a small mass of pebbles, which is then bound off by palm fibers to create a neck. In every example, the general scheme is followed such that some basic qualities of the human form are evoked: a front and back, a head, and bilateral symmetry.[5]

    Fig. 6. Doll-figurine illustrating basic compositional elements, 8.5 × 5 cm. KM 26415 (not previously published.)
    Fig. 6. Doll-figurine illustrating basic compositional elements, 8.5 × 5 cm. KM 26415 (not previously published.)
    Fig. 7. Swaddled doll-figurine, 8 × 3 cm. KM 26413.
    Fig. 7. Swaddled doll-figurine, 8 × 3 cm. KM 26413.

    In a few cases among the corpus of Karanis dolls, these elements are elaborated even further. A close examination of KM 7513 (fig. 3) reveals that the usual scheme was followed for the doll’s structure, but to it was fastened a stick for arms, and a hooded cloak was then draped over the arms and body, in effect disguising the core components. Another of the dolls (KM 7512 [fig. 4]), already with eyes instead of the single nucleus, also contains intriguing additions. A mass of mud mixed with hair is attached around the head feature and held together by a faded blue piece of textile on the back. Interestingly, the weave of this blue textile piece is sprang, the “elastic” fabric structure mentioned above; this technique produces a type of netting material that was used to make hairnets during the Coptic period in Egypt (Wild 1988, 50). The mud-hair-sprang component is attached to the inner portion by a piece of brown cord wrapped around the body. KM 3543 (fig. 8) has hair as well, falling into two plaits on either side of the head; it is secured around the neck with extra cord. Even though they are found on only a handful of the Karanis dolls, these unusual features make it difficult to deny a fundamental representational goal on the part of those who made these objects.

    Fig. 8. Doll-figurine with plaited hair, 7.5 × 8.5 cm. KM 3543 (not previously published.)
    Fig. 8. Doll-figurine with plaited hair, 7.5 × 8.5 cm. KM 3543 (not previously published.)

    At first, the alteration of red or blue strips of cloth, with either the papyrus reeds or undyed brown cloth to form the head, may appear to be just a decorative design. Yet, again, considering the small scale of these figurines and the human iconographic elements, the alternating colors do focus the viewer’s attention almost immediately to the “face” and “eye” features of the dolls. While the use of multiple strips of cloth or papyrus may indeed be simply a design choice (KM 26416 [fig. 1], by contrast, using a single piece of textile), this design leaves the impression of concentrism, highlighted further by the contrasting colors. It is worth mentioning that this same motif is displayed on one of the wooden dolls seen in figure 2 (third from right, top row). To be sure, this could easily be explained as coincidence, but it is worth considering that the textile and papyrus figurines may actually belong to a larger corpus of artifacts, including the wooden dolls, that overlap in significance and use.

    Three Figurine Contexts at Karanis

    The archaeological contexts for the textile and papyrus figurines provide some clues about their use, but several initial cautionary remarks are necessary. While the arid conditions of the Egyptian deserts have preserved textiles for the archaeologist to study, most of the cloth finds from Karanis are nonetheless fragmentary in nature. Additionally, fourteen of the Karanis dolls were recovered from the most recent occupation phase at the site (Layer A, fourth through fifth centuries CE), and many of these contexts were characterized by the excavators as rubbish heaps (Thomas 2001, 22–23). Three dolls came from Layer B (mid-third through mid-fourth centuries CE) and two dolls from Layer C (mid-second through mid-third centuries CE); four of these contexts were houses, and one was a street (Husselman 1979, 7–32). While it may at first appear from a stratigraphic point of view that an investigation of the dolls in the B and C layers would be the more productive, such may not necessarily be the case. As the town of Karanis expanded and contracted in size over time, different areas were occupied continuously, while others were abandoned for various periods. Furthermore, when new construction was undertaken, this often occurred on top of former structures, sometimes incorporating them into the new ones (Boak and Peterson 1931, 43). Thus, it is likely that many of the B- and C-layer contexts could be considered “rubbish,” in much the same way as many A layer contexts are. This, however, does not preclude an investigation of archaeological context: the formation of rubbish deposits may still contain patterns of artifacts within them, indicating associations with particular activities, consumption habits, and identities (cf. Thomas 2001, 22–23; Rathje and Murphy 1992).

    KM 3647 (Layer A, Area 200). KM 3647 (fig. 5) was a surface find from a highly disturbed region of Karanis. The area was characterized by garbage (both ancient and modern), mined at one time for fertilizer, and also combed through in 1895–1896 by an expedition in search of papyrus documents (Boak and Peterson 1931; Wilson 1933). Thus, this papyrus figurine—along with the associated finds from this horizon—are clearly removed from their primary context of use. Even so, the assemblage may still maintain some coherence as a secondary deposit. In the case of KM 3647, associated finds include: pottery, glass, beads, a bronze ring, a spindle whorl, remains of a weaver’s comb, dozens of textile pieces, papyrus fragments representing a receipt for a household payment of grain, a bone amulet with concentric circles, and a faience figurine of Harpokrates (the son of Isis). All of these items would be consistent with a domestic context and, although the evidence is scanty, the assemblage does indicate the types of material possessions that individual households were likely to have.

    KM 26413 (Layer B, House 36). Archaeological work in 1926 unearthed the structure known as House B36. The previous season of excavation in 1924–1925 was the first and, as just noted in the case of Area 200, much of the ground explored by the archaeological team had already been highly disturbed. In search of more complete remains of the latest period of occupation, the excavators moved to a part of the site known as Area G. Contrary to their expectations, however, most of the structures in this portion of Karanis dated somewhat earlier, to the B-layer occupation (mid-third to mid-fourth centuries CE) (Boak and Peterson 1931, 9, 37).[6] House B36 was a small, one-room structure from which only a handful of artifacts were recorded, including (in addition to the figurine in question [KM 26413, fig. 7]): a threaded wooden needle, glass, a papyrus document, and an architectural terracotta fragment. In her recent publication on the textiles from Karanis, Thelma Thomas notes that the threaded wooden needle “is intriguing insofar as it could not have been used for sewing but served some other purpose, perhaps as a toy or prop meant to teach a child how to sew” (Thomas 2001, 26).[7] The assemblage here is modest, but it is once again helpful in visualizing the kinds of items that may have been a part of a household at Karanis.

    Fig. 9. Plan of houses C54, C55, and C75.
    Fig. 9. Plan of houses C54, C55, and C75.

    KM 7512 (Layer C, House 54). The elaborate textile doll KM 7512 (fig. 4) was recovered from a residential area of the city with several houses dating to the C level (mid-second to mid-third centuries CE). This part of the town was continuously occupied through C and B periods, and the multiple construction phases often reused materials from previous structures. House C54 may originally have been constructed as a kind of duplex unit, with the northern part of the house sharing a central courtyard with House C55 (fig. 9) (Boak and Peterson 1931, 63–64). Later during this period of occupation, House C55 was subsumed by its neighbor to the west, and C54 claimed the southern two rooms once previously a part of House C75. KM 7512 came from room A in House C54; this corresponds to the highest part of the fill and may constitute a second story. It is the only artifact catalogued for this room, however, and the distinction between upper and lower stories is unclear.

    Perhaps, then, it is useful to consider the artifacts recovered from this house’s room E, described as below room A. These include: a badly denuded terracotta figurine of Isis and Harpokrates, ivory dice, several kinds of faunal remains, a fragment of a muzzle, a wooden tethering stake, flint grinders, a whetstone, a fragment of a weaver’s comb, and a wooden doll. That there are several items related to the keeping of animals and the grinding of grain is not surprising since this room opened out into a courtyard area where such activities took place. It is, however, intriguing that KM 7512 is found in general association with the figurine of Isis and Harpokrates, the ivory dice, and the wooden doll. The terracotta figurine is interesting since there is a wall niche in the adjacent room, which has been posited to be both of a utilitarian nature and indicative of domestic ritual.[8] The dice and the wooden doll—if indeed this item could have been a plaything—are both suggestive of games and leisure activities. These possible associations are tantalizing in terms of the kinds of domestic activities with which the textile doll may be associated.

    The serious constraints that the stratigraphy at Karanis, as well as depositional and postdepositional processes, impose on proposed interpretations of the textile and papyrus figurines cannot be stressed enough. Even so, some very general observations are nonetheless worth noting. Most of the excavations at Karanis uncovered residential areas, although early exploration at the site did identify a temple precinct and a few burials (Boak and Peterson 1931, 2). Despite this, none of the dolls are known to have come from any context except those associated with residential areas; they were all found in houses, streets, and large areas without architecture. From the sample contexts described above, it is interesting that two of the dolls are generally associated with items that could be linked to children’s activities (the needle, the dice, and the wooden doll), and two more are associated both with Harpokrates figurines and with a wooden doll. While inferences about the meaning of the textile and papyrus figurines from this information alone could hardly be pressed further, turning to some comparative material can offer an additional lens through which to consider these peculiar artifacts.

    Comparative Material

    On the basis of their physical features alone, the Karanis figurines are a striking departure from most other textile objects known as dolls. A publication by R. M. Janssen describes a group of six textile dolls from various sites in Egypt, all dating from the first through fourth centuries CE—roughly the same timespan as the Karanis dolls (Janssen 1996). The most notable difference between Janssen’s group and the Karanis corpus is the size of the figurines: on average, the dolls from Karanis are nearly 12 cm smaller in height. Additionally, Janssen’s examples all have articulated limbs and, in some cases, show distinct facial features, breasts, and pubic areas. Three of these larger dolls were recovered from late nineteenth-century excavations of a Roman-period cemetery at Hawara, a site located in the Fayum oasis about two dozen kilometers south of Karanis (Petrie 1889, 12; Janssen 1996, 234). They are recorded as coming from female children’s graves that also included items such as miniature bottles, a miniature mirror, and textile-working accouterments. The association of these dolls with children is made clear by the context, but their resemblance to the Karanis figurines is nonetheless fairly weak.[9]

    There are, in fact, only two known examples of textile figurines that bear any real similarity to the Karanis dolls in both size and iconography. The first was excavated in 1905–1906 at Saqqara (northeast of the Fayum oasis) from a modest burial dating roughly to the seventh through ninth centuries CE; it was “that of a child, a bronze cross on the left arm, and in the hand, tightly clasped, a little rag doll” (fig. 10) (Quibell 1907, 12, pl. XXXVIII, no. 4). The second example—also dating to this later period—now resides in the Austrian Museum of Art and Industry (fig. 11) (Hoeber 1912, fig. 15). Its provenance is dubious, though it too may have come from a gravesite in Egypt, perhaps also Saqqara.[10] The construction technique of folded strips of textiles layered around each other to produce the head, and the further use of the technique to construct the eyes, clearly echo these same features on the Karanis figurines. At least for the first example, the association of this textile figurine with the world of the child is definitive because of the archaeological context. The same cannot, unfortunately, be said for the Karanis dolls, but their resemblance to these two examples is difficult to ignore.

    Fig. 10. Textile doll, from child's grave at Saqqara. : Quibell 1907, pl. XXXVIII, no. 4
    Fig. 10. Textile doll, from child's grave at Saqqara.
    Quibell 1907, pl. XXXVIII, no. 4
    Fig. 11. Textile doll, provenance unknown.: Hoeber 1912, fig. 15
    Fig. 11. Textile doll, provenance unknown.
    Hoeber 1912, fig. 15

    One final example of comparative material for the Karanis corpus is worth mentioning, though it does not bear directly on the textile figurines. Instead, it relates to the previously mentioned possibility that the textile and papyrus dolls from Karanis—because of the use of the concentric circles motif—may be thematically (and perhaps even functionally) linked to the wooden dolls from the town. One example, now in a private collection, has a body made of wood and is draped in a woolen garment; carved into its wooden head are two circles clearly intended to be eyes (described by Fluck 2004, 399, pl. XXVI). This case alone suggests that it would be worthwhile to consider the full spectrum of Karanis dolls (textile, papyrus, stone, and bone) as a single corpus, and perhaps additional contextual research may reveal previously obscured associations among the domestic assemblages at the town.

    Interpreting the Karanis Dolls

    The previous sections have considered a “bottom-up” approach toward evaluating the textile and papyrus figurines from Karanis—beginning with a description of the artifacts, moving to an investigation of select archaeological contexts, then looking to comparative material—in an attempt to cull enough hints to converge on some possible interpretations of these objects. What follows is a speculative proposal, which focuses on the observations that the Karanis dolls are small-scale items, that they are iconic in their representation, that they appear largely in domestic contexts, and that they might relate to the world of the child. This proposal is offered only as another step up in an interpretive process.

    Let us suppose that the Karanis figurines are amulets. The size range of the dolls is much more consistent with the size of amulets, which were small-scale, often hand-held objects created for the protection of an individual. Their use was pervasive in Dynastic Egypt and, while they were utilized especially in funerary ritual, they also served the living as apotropaic measures against misfortune and disease (scarabs are perhaps the best-known Dynastic Egyptian amulets). This amulet tradition persisted through the Graeco-Roman periods in Egypt, often invoking through its iconography the aid of deities from the Egyptian pantheon. The Kelsey Museum’s collection houses more than 250 amulets from Karanis alone, several of which are representations of such gods as Serapis, Osiris, Isis, Horus, and Harpokrates. These amulets also take the form of dangerous animals—crocodiles and lions, for instance—as if to harness their ferocity on behalf of the protection of an individual. While amulets were often made of particular kinds of stone or metal, they were also manufactured from organic materials as well (Aune 1997).

    Just as amulets could be clear in their presentation, by taking the form of a deity or by employing a magical inscription, so too they could be ambiguous, such that they are nearly impossible for the archaeologist to identify in terms of their function and meaning. If indeed the Karanis textile and papyrus figurines are amulets, then another look at their iconography may be helpful. In this context, it is striking to note the similarity of the Egyptian tit and sa amulets with the Karanis dolls. The sa-sign, or ankh, embodied protective forces, and the large loop at the top of the symbol almost seems to correspond to the head of the Karanis dolls. The tit, or “girdle of Isis,” in amulet form also shares structural characteristics with the dolls: “the standard form of this amulet is an open loop of material from whose bound lower end hangs a long sash flanked by two folded loops” (Andrews 1994, 6, 40, 44–45, fig. 45). In addition to an iconic evocation of the Egyptian ankh, the Karanis dolls also carry the essence of human iconography in their features—almost a literal prefiguration of a human, perhaps of a child to come (Nel and Leibhammer 1998, 223).

    From the comparative examples of dolls discussed here, it is clear that some types of textiles dolls were associated with the world of the child. Again, hypothetically, suppose that the Karanis figurines were amulets and that they were crafted for the protection of unborn and newborn children. Infant mortality was a profound concern in Roman Egypt: an infant had a one in three chance of dying before the first birthday, although after this first year the chances of survival to adulthood increased (Bagnall and Frier 1994, 77, 100). Any measures taken, either in the form of community or personal ritual, could help allay the anxieties associated with childbirth and the earliest period of a child’s life. If a child did indeed die, then, as in the case of the child’s grave at Saqqara, the protective textile doll amulet was buried too.

    Considering these realities of child mortality, it is probably no coincidence that there was a robust community honoring the child-god Harpokrates at Karanis. Ample testimony to this cult is offered by the numerous terracotta figurines representing the deity found in both residential and temple contexts at the town and by two domestic wall paintings, one of Harpokates alone and one depicting him nursing at the breast of Isis, his mother (Allen 1985, 137, fig. 13; Boak and Peterson 1931, 34). These two deities are the focus of F. Dunand’s study of popular religion in Roman Egypt (1979), in which he establishes the widespread devotion to the various manifestations of these deities. He argues that invoking symbols of fertility was especially important to rural populations, such as that of Karanis, both for agricultural production and for human reproduction. The nature and characteristics of the Karanis textile and papyrus figurines dovetail nicely with such ritual beliefs and practices, and it is worth repeating that two of the dolls whose contexts were described above were found in association with terracotta figurines depicting Harpokrates and Isis.

    To be sure, these final comments intimating a connection between the corpus of Karanis dolls discussed here and the cult of Harpokrates are tenuous at best. The truth of the matter is that whatever the significance of these unusual artifacts may have been, it will remain largely unknown. The analysis offered here, nevertheless, does seem to point in the direction of the child’s world. It may be that these figurines were playthings; amulets they may have been; perhaps, in fact, they had a multidimensional meaning and function, encompassing the worlds of leisure and ritual. In any case, the textile and papyrus figurines from Karanis represent a tradition unparalleled elsewhere in Roman Egypt.


      1. Several dolls have been included in other publications: KM 26413 (Gazda 1983, fig. 52; Maguire, Maguire, and Duncan-Flowers 1989, fig. 146); KM 26413 can also be seen with KM 3647 in Thomas 2001, figs. 37–38; finally, KM 3543, KM 7512, and KM 7513 all appear in Fluck 2004, pls. 7–9.return to text

      2. A technical analysis of the fiber content and fabric structure for all of the dolls discussed here along with color images are now available online through the Kelsey Museum Artifacts Database. The project to provide this information for all of the textile holdings in the Kelsey Museum has been a result of efforts by Thelma Thomas, Kate Carras, and Jane Batcheller.return to text

      3. It is important to bear in mind here that several of the dolls have probably not survived fully intact. Nevertheless, the fact that there appear to be some figurines more fully intact than others suggests that the general size range remained on the smaller end of the scale.return to text

      4. One doll (KM 3542) had lost the binding that secured the object’s structure, making it possible to probe the core in order to examine its contents; this revealed only dirt and sand, which is likely to be at the core of the other dolls’ heads and was probably used to lend structural integrity to the facial impression.return to text

      5. It is particularly interesting that the human characteristics of these dolls seem to satisfy only the bare minimum of signals that are cognitively evocative of the human form and notions of agency. These signals—facial cues, eye spots, and symmetry—have a deep evolutionary history and can be triggered by the simplest suggestion (Gell 1998; Gombrich 1973; Washburn 1999).return to text

      6. The B-level chronology in Area G is inconsistent, and the assignment of dates was made on the basis of papyri, ostraca, and coins. Some of the houses, including B36, date to what is generally regarded as the B-layer period (i.e., mid-third to mid-fourth centuries CE). Other parts of Area G, however, date to a slightly earlier period (second to early third centuries CE). The authors also note how one corner of House B36 is set back in a curve, which they hypothesize was done to avoid blocking one of the streets on which this structure sits. return to text

      7. She also explains that the cotton cloth through which the needle is threaded is printed with a patterning process that she notes is unattested in Egypt during the Roman or Byzantine periods and probably belongs to a later medieval occupation at the site. This observation highlights how difficult it is to sort out chronologies at Karanis. return to text

      8. Husselman (1979, 47) identifies this as a cupboard or utilitarian niche; Gottry (1995, 33–36) discusses these types of unembellished niches in the context of domestic niches in Dynastic Egyptian traditions.return to text

      9. Fluck (2004) also describes several dolls falling in between the types published by Janssen (1996) and the Karanis figurines.return to text

      10. Fluck (2004, 387–388) explains how this example was perhaps previously known under a different inventory number.return to text

      Works Cited

      Allen, M. L. 1985. “The Terracotta Figurines from Karanis: A Study of Technique, Style, and Chronology in Fayoumic Coroplasts.” PhD dissertation, University of Michigan.

      Andrews, C. 1994. Amulets of Ancient Egypt. Austin: University of Texas Press.

      Aune, D. E. 1997. “Amulets.” In The Oxford Encyclopedia of Archaeology in the Near East, ed. E. M. Meyers, 113–115. Oxford: Oxford University Press.

      Bagnall, R. S., and B. W. Frier. 1994. The Demography of Roman Egypt. Cambridge: Cambridge University Press.

      Boak, A. E. R., and E. E. Peterson. 1931. Karanis: Topographical and Architectural Report of Excavations during the Seasons 1924–1928. University of Michigan Studies, Humanistic Series 25. Ann Arbor: University of Michigan Press.

      Dunand, F. 1979. Religion populaire en Égypte romaine. Leiden: E. J. Brill.

      Fluck. C. 2004. “Ägyptische Puppen aus römischer bis früharabischer Zeit.” In Gedenkschrift Ulrike Horak (P. Horak), ed. R. Pintaudi, 385–400. Papyrologica Florentina 34. Florence: Edizioni Gonnelli.

      Gazda, E. 1983. Karanis: An Egyptian Town in Roman Times. Ann Arbor: Kelsey Museum of Archaeology. (Revised and updated edition, Kelsey Museum Publication 1, 2004.)

      Gell, A. 1998. Art and Agency: An Anthropological Theory. Oxford: Clarendon Press.

      Gombrich, E. H. 1973. “Illusion and Art.” In Illusion in Nature and Art, ed. R. Gregory and E. H. Gombrich, 193–243. London: Duckworth.

      Gottry, H. 1995. “Domestic Religion in Graeco-Roman Karanis: Origins, Theories, and New Approaches.” Undergraduate honors thesis in Classical Archaeology, University of Michigan.

      Hoeber, F. 1912. “Alte Puppen.” Kunst and Kunsthandwerk 15:416–417.

      Husselman, E. M. 1979. Karanis Excavations of the University of Michigan in Egypt, 1928–1935. Topography and Architecture: A Summary of the Reports of the Director, Enoch E. Peterson. Kelsey Museum Studies 5. Ann Arbor: University of Michigan Press.

      Janssen, R. M. 1996. “Soft Toys from Egypt.” In Archaeological Research in Roman Egypt: The Proceedings of the Seventeenth Classical Colloquium of the Department of Greek and Roman Antiquities, British Museum, Held on 1–4 December, 1993, ed. D. M. Bailey, 231–239. Journal of Roman Archaeology suppl.19. Ann Arbor: Journal of Roman Archaeology.

      Maguire, E. D., H. P. Maguire, and M. J. Duncan-Flowers. 1989. Art and Holy Powers in the Early Christian House. Illinois Byzantine Studies 2. Urbana: University of Illinois Press.

      Nel, K., and N. Leibhammer. 1998. “Evocations of the Child.” In Evocations of the Child: Fertility Figures of the Southern African Region, catalogue of an exhibition held at the Johannesburg Art Gallery Exhibit, ed. E. Dell, S. Shoolman, and E. Burroughs, 219–223. Johannesburg: Johannesburg Art Gallery; Cape Town: Human and Rousseau.

      Petrie, W. M. F. 1889. Hawara, Biahmu and Arsinoe. London: Field & Tuer.

      Quibell, J. E. 1907. Excavations at Saqqara 1905–1906. Cairo: Imprimerie de l’Institut Français d’Archéologie Orientale.

      Rathje, W. L., and C. Murphy. 1992. Rubbish! The Archaeology of Garbage. New York: Harper Collins.

      Thomas, T. K. 2001. Textiles from Karanis, Egypt, in the Kelsey Museum of Archaeology: Artifacts of Everyday Life. Ann Arbor: Kelsey Museum of Archaeology.

      Washburn, D. 1992. “Perceptual Anthropology: The Cultural Salience of Symmetry.” American Anthropologist 101.3:547–562.

      Wild, J. P. 1988. Textiles in Archaeology. Aylesbury: Shire Archaeology.

      Wilson, L. M. 1933. Ancient Textiles from Egypt in the University of Michigan Collection. University of Michigan Studies, Humanistic Series 31. Ann Arbor: University of Michigan Press.

      Wipszycka, E. 1965. L’Industrie textile dans l’Égypte romaine. Archiwum Filologiczne 9. Wrocław: Zakład Narodowy im. Ossolinskich.