*3.1. What Is the Context?*

Central to any archaeological inquiry is the question of context. In the case of explorations of the practice of Israelite religion, an immediate problem is that there was not any clear division between "secular" and "sacred" in the way that may exist in many post-Enlightenment modern societies today. Indeed, all practices were religious in one way or another and are represented in the household, public spaces, and tombs, as well as in shrines and temples. That said, such practices were sacred in different ways and seem to have existed on a continuum of "embeddedness" (Renfrew 1994), that is, a spectrum of variation in the degree to which such practices were embedded in the everyday activities of this world (such as eating a daily meal) or focused in a special way on experiencing the world of the supernatural (such as participating in a temple feast). Traditionally, architecture and artifacts have been the primary identifiers employed within such paradigms, but in recent decades, animal bones have been recognized as equally important (Horwitz 1999). Indeed, the most plausible identifications of cultic contexts engage a variety of data sets.

#### 3.1.1. What Type of Space Is It?

For households or other communal domestic spaces, evidence of religious practice is prevalent especially in so-called "cult corners" (cf. examples in Zevit 2001). Such spaces are usually identified by the high concentration of vessels that are generally associated with cultic practices, such as chalices, censers, and zoomorphic vessels, as well as architectural features such as low benches, which served as resting places for many of the vessels. Animal bone finds in these contexts also show variance with

nearby deposits, though, as will be discussed below, we suggest that greater care should be exercised in determining the nature of the deposit.

Tombs are also inherently ritual spaces, often with sequences conveyed in their layouts and associated epigraphic remains (Mandell and Smoak 2017). Animal bone finds are especially important in such contexts for distinguishing between tomb deposits that may be connected to these rituals and random deposits of non-primary accretions, as tombs were often secondarily used as dumping spots. Here, Horwitz (1987) distinguishes chance animal bone finds from intentional food offerings or sacrifices by evaluating a number of factors including: (1) the degree of association of the bones with a tomb or human remains; (2) a restricted range of species present (e.g., only sheep and goats); (3) evidence of intentional carcass element distribution (e.g., a lack of butchering refuse, such as phalanges, and the presence of heads typical of certain offerings [Horwitz 2001; Lev-Tov and Maher 2001]), that may include (4) articulated units (especially limb portions), and may exhibit (5) side preference; as well as (6) age-based, and (7) sex-based selection. Identification of primary concentrations is essential for determining the nature of the deposit and whether such remains represent a cultic meal *with* dead or a votive offering *for* the dead.

Temples and shrines are set apart by their dominant purpose of the veneration of the deity (or deities) and all of the associated activities, communal and restricted, that comprise worship. Here architectural features that follow particular patterns are often initial indicators (Holladay 1987; Zevit 2001; Mierse 2012), as well as the presence of an altar or a variety of altars. In rare cases, actual images or representative symbols of deities may be uncovered as well. Animal bones from temple contexts are often found in intentionally buried deposits resulting from votive deposits, favissa burials, and pits associated with sacrificial refuse or with the remains of sacred meals (e.g., Greer 2013). Major temples are comparatively rare in the southern Levant during the Iron Age (cf. Faust 2010, 2019), with the notable exceptions of Arad (Aharoni 1968; Aharoni 1993; Herzog et al. 1984; Herzog 1997), Moz.a (Kisilevitz 2015) and Dan (Biran 1994; Davis 2013; Greer 2013, 2017).

#### 3.1.2. Is It a Deposit?

A problem in a number of publications is the failure to determine the specific nature of the context from which the animal bone sample was recovered. Complications can arise even when the locus may be "secure," as defined by chronologically homogenous ceramic assemblages, clear site matrix relationships, and/or sharply defined architectural or stratigraphic features. Unlike ceramic remains, faunal remains recovered from "floors," for example, tell us less than most assume about any associated practices and may represent the activities of subsequent inhabitants when that floor was no longer used as a living space. This is due to the simple fact that animal bones are "trash," and people generally do not live in their trash. That is, it is unlikely that people, ancient or modern, would consume a portion of an animal, then dump its remains on the floor and leave them there while they rot and fill the space with their stench. Likewise, the discovery of an articulated animal carcass near an altar is neither evidence of a sacrifice left in situ, nor of a sacred meal during which the worshipers carefully picked the meat off each bone without disturbing the articulation. No, animal bones, whether resulting from a highly charged meal in a domestic space or a festival in a temple, were collected and discarded in communal middens, or saved and buried in intentional deposits. It is important to remember, too, that the find spots of these deposits should not necessarily be associated with the spaces of the eating events. Animal bones discovered outside of such concentrations are more likely surface scatter from disturbed deposits, perhaps used as fill and sometimes even as fuel, and are less helpful for approaching questions other than presence and absence of certain species across the stratum. Still, questions of presence and absence can play important roles in dietary differences based on religious prohibitions, rightly understood, as may be in the case of the pig mentioned above.

Bones discarded in communal middens result in high concretions of faunal material in spaces defined by architectural features, from simple pits to slots in between walls from previous strata, as well as in silos and tombs and other abandoned features that are used secondarily as dumping spots. In that these bones, once religiously charged by their association in cultic activities, are now mixed with the bones discarded from everyday meals (or less religiously charged meals), their usefulness for reconstructing details of those cultic activities is minimal. Bones from concentrated deposits, however, are rich with potential for recovering details about ritual practices and should be the focus of inquiry.

#### 3.1.3. What Type of Deposit Is It?

Even when deposits are identified, far too often such are simply identified as "sacrificial remains" when, in fact, they may represent various types of religiously charged deposits with very different characteristics. Identifying these differences is especially important if one seeks to speculate regarding the cognitive dimensions of the significance of these deposits for ancient peoples.

Deposits resulting from burnt offerings, for example, will consist of fine ash and, if any elements can even be identified, they will be intensely charred and denatured often obscuring their analysis. Such finds are only rarely reported in archaeological reports of the southern Levant, though they have been identified in Aegean archaeology (Forstenpointner 2003). In all likelihood, such deposits are often missed or misidentified in excavation.

Votive deposits, as symbolic offerings that were not consumed, are most readily identified by their articulation with other bone elements. That is, the discovery of a fully articulated limb of a particular animal would render it highly unlikely that it had been consumed in that most elements become disarticulated in the butchering process. Further, even those units that retain their articulation in the butchering process become separated in the final stages of preparation prior to cooking, such as in the disjointing and chopping undertaken to fit portions into a cooking pot, and even more so in the cooking process itself, typically an extended stewing. Such votive portions are most typical in tombs and may be viewed as symbolic portions for feeding the dead, but have also been discovered buried in jars or in pits in the floors of sanctuaries often exhibiting a side preference (Davis 2008; Greer forthcoming).

Most animal bone finds, as mentioned above, are the remains of eating events and when discovered in religiously charged contexts they may be confidently associated with sacred feasting. In contexts where there is an altar structure or other paraphernalia associated with the slaughter of animals, such as in a temple, one may further suggest that such are the remains of sacrificial meals. In contrast to the remains of burnt offering and votive deposits, these bones are often from the meaty portions of the animal, though not exclusively, and exhibit postmortem modifications such as cut marks and breakage patterns associated with the disarticulation and chopping of the portions into smaller units for cooking.

Considering the difference in the type of religiously charged deposits here discussed, we suggest that greater precision should be applied to potential designations of excavated remains as each exhibits different characteristics. Here, too, we suggest greater clarity in speaking about "sacrifice," in that most sacrifices are evidenced in the form of sacred feasts.

## 3.1.4. Is the Context Yahwistic?

Associating deposits, whether from burnt offerings, votive offerings, or the remains of sacred meals, with Yahweh-centric worship is a complicated endeavor due to many of the factors discussed in Section 2.1. Our strongest evidence comes from specific mention of Yahweh in any associated epigraphic remains, such as one finds at the gate shrine of the desert caravansary site of Kuntillet 'Ajrud or in certain tomb inscriptions such as Khirbet el-Qom (Dever 2005). Such may also be suggested by names with Yahwistic theophoric elements that are discovered in close association with other finds that suggest a temple context, such as at Arad (Aharoni 1968; Herzog et al. 1984) or at Tel Dan (Biran 1994), yet this is not as straight forward. Iconography represented in certain artifacts found at an installation, too, may be associated with Yahwistic imagery, most notably, the images of the Taanach cult stands that are often connected to Yahweh and (the) Asherah (Taylor 1993; Zevit 2001; Dever 2005).

The style of particular structures or artifacts and their measurements, such as the field stone altar at Arad (Herzog et al. 1984, p. 11; Zevit 2001, pp. 169–70), have also been associated with Yahwistic worship in the context of other finds. So, too, the presence of an altar kit, including what may be a bowl for catching sacrificial blood (Greer 2010), as well as the proportions of architectural features (Davis 2013; Greer 2013), and iconographic features (Ackerman 2013) at Tel Dan have been associated with Yahwistic worship. Similar associations have been suggested in the discovery of horned altars, such as at Beersheva (Zevit 2001, pp. 171–74). There are, however, two potential problems with this second category of identification: (1) they are often based on associations with biblical texts (an issue that will be discussed below), and (2) without further epigraphic association with Yahweh, such features may be shared by other non-Yahwistic sanctuaries as well, as exemplified in the architecture of Ain Dara (Monson 2000) or variations of horned altars associated with Philistine and Aegean contexts (Gitin 2002; Sala 2018).

Less helpful are determinations of Yahweh-centric worship made purely on mention of a location that is associated with Yahweh in the Bible in that sites were often the home of various peoples with differing religious allegiances, as the texts themselves describe, and this varied over time. Another problem is that the dating and composition of biblical texts is not straightforward and this can complicate the association of certain sites with Yahweh. Such may be the case with the Mount Ebal site that clearly exhibits evidence of cultic activity (Hawkins 2012), along with evidence of feasting (Horwitz 1986), but seems unlikely to be associated with the altar of Joshua even if historicity is granted in that our earliest textual traditions, evidenced later in the Samaritan Pentateuch and confirmed against the Masoretic text by evidence from readings from the Old Greek traditions and a Qumran fragment, place this altar on Gerizim (cf. Knoppers 2013, pp. 202–3, with references). Determining the religious affiliation of a site based on the presence or absence of certain animal bone remains, such as the lack of pig bones in cult areas at Qeiyafa (Garfinkel 2017, though listing other factors as well), is also insufficient in that worship sites associated with other deities, such as the gate shrine at Geshur, also exhibited these same sacrificial avoidances (Fisher 2005; Greer 2013, p. 100 n. 10), not to mention the complexities associated with the absence of pig remains (Hesse and Wapnish 1997; 1998). While such sites may well have been Yahweh-centric, there is not enough evidence from animal bones alone to determine their affiliation and our strongest cases can be made when evidence from multiple sources—epigraphic, artifactual, architectural, and biblical—is compounded in addition to the zooarchaeological evidence under consideration.

#### *3.2. How are Bones Excavated?*

Once a context is established, standard methods of excavation, as they pertain to animal bones, usually entail hand collecting any visible remains of bones and placing them in paper bags marked by area, locus, square, and collection date to be analyzed later by a zooarchaeologist. Many digs now also mandate sieving, either wet or dry, and/or flotation protocols, but usually only as a random sampling technique (one out or every five buckets, for example) or in areas that have been identified as particularly important for one reason or another (a midden, for example). This lack of comprehensive sieving, of course, biases the collection toward larger species—e.g., sheep, goats, cattle, deer, pig, etc.—and significantly underrepresents smaller species that may be present in a particular sample. Such random sampling can, however, indicate presence and absence of fish and birds and certain microfauna (i.e., rodents and reptiles) and even with comprehensive sieving, element counts should not be used to calculate percentages of species due to the inevitable loss via natural taphonomic processes, discussed below, that disproportionately underrepresent smaller species.

Other problems at the collection stage pertain to the selection bias of the excavators, often volunteers with little experience, who may single out certain bones that are larger, such as a whole metapodial, or more interesting to the untrained eye, such as an astragalus. They may further miss the relationship of articulated bones and the importance of associating them together even if in different baskets or even loci. Another problem is that bone elements may be destroyed by a pick or trowel in excavation resulting in the collection of only pieces of shattered bone, sometimes even spread between buckets, or the complete obliteration of diagnostic features of a particular element, thus significantly skewing element counts. With large digs that have been in the field for years, an added dimension

is potential staff turnover in which case one cannot be confident that the same protocol in collection technique was applied equally across the sample.

While these are all very real problems in recovering a sample, a simple counter measure is to maintain the presence of a trained zooarchaeologist on site who can equip volunteers and provide oversight to collection methods. A further advantage to such presence is interaction among the staff onsite regarding particular questions of context that may be examined through faunal analysis; such cannot occur when zooarchaologists analyze remains without an understanding of the context and, conversely, excavators may make different decisions in the field with a more complete understanding of the animal bone remains.

#### *3.3. How Are Bones Analyzed?*

As in all archaeological research, samples are analyzed with particular questions in mind. That said, regardless of questions posed, certain features of faunal specimens are typically assessed due to their applicability to a wide range of questions. Such include the anatomical identification, orientation, and size of individual elements, the various taxa from which the specimens likely derive, the age and sex if it can be determined, and the nature, location, and frequency of postmortem modifications to the bone (Hesse and Wapnish 1985; Davis 1987a; O'Connor 2008; Reitz and Wing 2008; Gifford-Gonzalez 2018).

#### 3.3.1. What Is Recorded and How Is It Determined?

The initial identification takes place on the elemental level, in that related taxa and to a large extent all vertebrate mammals, share anatomical similarities (Davis 1987a). The closer the relationship, the more similar the anatomy. Once the element is identified and sided, either left, right, or indeterminate, identification is made regarding the taxa.

Identification of element and taxon is made by a visual comparison of the morphological features of a specimen, often supplemented by metrical data (von den Driesch 1976), with elements in comparative collections derived from excavated and/or modern remains, as well as with elements depicted in various standard manuals (Sisson et al. 1975; Schmid 1972; cf., Niven et al. 2009). While the anatomy of certain domestic species has changed some over the millennia in terms of size due to commercial animal farming, it is not so much as to obscure identification.

Typically, common domesticates, especially sheep, goat, and cattle, dominate the assemblages of the southern Levant, though deer and pig occur in significant numbers in certain regions during certain periods. Of these, sheep and goats are notoriously difficult to differentiate though this can be done with some degree of confidence for certain elements depending on the experience of the zooarchaeologist (Boessneck 1969; Zeder and Lapham 2010). Various bird and fish species are also quite frequent, as are remains from other animals that were not typically consumed, such as donkey and dog. Elements that cannot usually be identified at the species level (e.g., vertebrae, long bone shaft fragments, and rib fragments) are grouped in larger categories based on size and are not typically used for calculating species percentages.

Age at death is estimated by osteological and dental maturity based on the state of fusion in the former (Silver 1969; Zeder 2006) and on the eruption and attrition of teeth in the latter (Payne 1973; Grant 1982; Greenfield and Arnold 2008). The sex of the animal can be determined by certain elements such as the pelvis (Greenfield 2006a), though due to the fragmentary and limited nature of such indicative elements sex-based comparisons are only common for large samples.

Postmortem modifications to the bone, such as cut and chop marks, are also recorded and may be associated with the activities of slaughtering, processing, preparation, and consumption, often based on ethnographic parallels (cf. Binford 1978, 1981; Klenck 1995; Grantham 1995, 2000). Associated practices are also reconstructed through experimental archaeology that seeks to replicate ancient practices in ways that produce similar remains to what was excavated, as has been applied especially for Greek sacrifice (Forstenpointner et al. 2013). Modifications due to gnawing from dogs or rodents (Marean and Spencer 1991), weathering (Behrensmeyer 1978), or burning (Shipman et al. 1984) are

also recorded, as well as any deformations of the bone due to sickness and/or injury (Siegel 1976; Baker and Brothwell 1980).

Recent scientific advances have transformed the field in terms of providing new data sets for exploration (Gifford-Gonzalez 2018). The most common of these applications include advances in microscopy applied to differentiate between different cut marks and the technologies that produced them (cf. Shipman 1981a; Greenfield 1999, 2006b), isotopic analysis applied to determine grazing contexts and migration patterns (Gifford-Gonzalez 2018, pp. 503–10; Arnold 2018), and studies of ancient DNA applied to bring greater precision to species identification (Gifford-Gonzalez 2018, pp. 511–30), among others.

#### 3.3.2. How Are Bones Counted and Compared?

Before bones can be counted and compared, one must first determine how well the excavated sample represents what was deposited in the past, i.e., understanding and accounting for the taphonomic processes that contribute to the creation of certain deposits and the modification and eventual destruction of their contents over time (Shipman 1981b; Hesse and Wapnish 1985; Lyman 1994). Such differences must then be evaluated as to the statistical probability that the observed difference was created by intentional processes and may not be explained by random selection (Lyman 2008).

Once taphonomic realities have been considered, bones are then quantified, and discussion continues about the best way to count bones that is both reliable, i.e., replicable, and valid, i.e., that it measures what it aims to measure (O'Connor 2008, pp. 54–67). The two most frequently employed procedures are those that base their analysis on raw counts of identified specimens, recorded as NISP (number of identified specimens), and those that seek to estimate the minimum number of individual animals present in a sample, recorded as MNI (minimum number of individuals). There are strengths and weaknesses to both in that each assumes a different context (Hesse and Wapnish 1985, pp. 112–16; Gifford-Gonzalez 2018, pp. 385–412): NISP assumes that the specimens derive from a large number of animals that have all been processed at different times and places, and deposited in mixed contexts (such as in a midden), thus rendering it unlikely that two or more bones derive from the same animal; MNI assumes the opposite scenario, i.e., that the bones are likely to have come from a small number of animals killed and deposited at the same time (such as a tomb). The once-common practice of weighing bone material is no longer employed in most workflows for faunal specialists due to inherent problems with its validity (Chaplin 1971; Casteel 1978).

Assessing the frequency of bone element representation in certain analytic categories has also been important in the application of zooarchaeology to the study of Israelite religion. Such analyses of carcass part distribution—that is, distributions based on differences between the representation of bones associated with certain groups of elements, such as the "head" (crania and teeth), "trunk" (vertabrae and ribs), "limbs" (scapulae, humerii, radii, ulnae, femorae, tibae, and long bone shafts), and "feet" (phalanges)—can indicate the syntax of butchering procedures, and has been applied to cultic settings (Hesse et al. 2012). Of special importance has been the difference between deposits that indicate a dominance of right or left sided portions in cultic contexts across the ancient Mediterranean (e.g., Davis 1987b, 2008; Marom and Zuckerman 2012; Greer forthcoming).
