Yasur-Lau Assaf. Household Archaeology in Ancient Israel and Beyond

Подождите немного. Документ загружается.

362 michael d. press

these buildings are to be recognized as temples.

On the other hand,

discussions of the problems of identication can be found in the pre-

sentation of more controversial cultic loci (e.g., the Lachish “Solar

Shrine” [Tufnell 1953: 141–145] and Aharoni’s reconstructed cult

place at Beersheba [Aharoni 1972: 123–127; 1973: 13–18]). It appears,

therefore, that in the previous examples the cultic nature of the build-

ings was seen as self-evident.

To my knowledge, the rst systematic critique of cultic identi-

cations was made by Chester McCown (1950). McCown not only

reviewed specic cases of cultic identications (for instance, the

interpretation of the entire city of Megiddo Stratum V as a sacred

area by the excavators), but the more general tendency to identify any

unusual or unexplained artifacts or buildings as “cultic” in nature.

For McCown, most artifacts could not safely be identied as “cult

objects,” with the exception of the limestone horned “incense altar”

(1950: 210). He concluded that most other objects usually judged

to be cultic, such as pottery “model shrines,” censers, ceramic and

bronze stands, chalices, and cup-and-saucer combinations, could

just as easily have served secular functions, and were perhaps luxury

objects. Bronze and clay gurines, meanwhile, were determined to

have “religio-magical signicance” (1950: 214), but as part of “domes-

tic rites” (1950: 215). McCown even made some positive suggestions

for identifying cultic structures, specically based on a combination

of their plan, “cult remains” (special artifacts and animal bones), and

continuity of unique buildings in the area (1950: 208).

Questioning of cultic identications increased in the 1970s and

1980s. ese critiques tended to be particularist, focusing on individ-

ual sites. S. Yeivin (1973), for instance, reevaluated the identication

of temples and cultic buildings at Taʿanach, ʿAi, Arad, and Lachish.

Ottosson, in his survey of cult places (1980), reanalyzed Jerusalem

Cave 1 and even the Lachish Fosse Temples as secular contexts; he also

I do not necessarily mean to question the identication of any of these build-

ings as temples; I am merely remarking on a basic problem of methodology in these

publications.

It should be noted that the studies referred to above predated important theoreti-

cal work by Processualist archaeologists, particularly Renfrew’s inuential publication

of the Phylakopi sanctuary (1985). e impact of such work is evident in the archaeol-

ogy of cult in the southern Levant, as over the past two decades, publications of cult

places have more consistently addressed the identication problem, even if only in a

cursory manner (see below).

a problem of definition 363

underlined the need for a general critical attitude toward cultic iden-

tications, emphasizing the characteristic of a temenos as common to

temples (1980: 9–10). Perhaps most signicantly, M. D. Fowler’s 1979

dissertation formed the basis for a series of articles in the early 1980s

(1981, 1982a, 1982b, 1983, 1984a, 1984b, 1985a, 1985b) articulating a

relatively systematic critique of cultic identications. Part of this cri-

tique consisted of reevaluation of specic cultic contexts (e.g., several

at Tell el-Farʿah [Fowler 1981], the Shechem Migdal Temple [Fowler

1983], and the Taʿanach Cultic Structure [Fowler 1984a]). At the same

time, Fowler questioned some of the bases for cultic identications

more generally, in particular the inherently cultic nature of specic

artifact types. Much of Fowler’s work had, in fact, been anticipated

by McCown, including his skepticism toward unique objects; Fowler

did, however, investigate certain classes of artifacts more thoroughly,

notably gurines (1985a) and “incense burners” (1984b, 1985b).

In order to demonstrate the diculties in using artifacts for cultic

identication, I will briey discuss the problem of gurines. I have

chosen them for two reasons: rst, they are the artifact class most

commonly cited in the literature as an automatic indicator of cult (cf.

Fowler 1985a: 333); and, second, they are the artifacts with which I

have worked most closely. Typically, gurines have been interpreted

as cultic objects or amulets (for a survey of interpretations, see Kletter

1996: 73–81). Starting in the late 1960s, however, detailed surveys of

ethnographic literature on gurines—especially by Ucko (1968) and

Voigt (1983: esp. 186–187)—have illuminated the wide range of func-

tions that gurines can in fact have. us, Voigt (1983: 186) suggested

ve broad functional categories for gurines: cult gures, vehicles

of magic, didactic tools, toys, and representations of the deceased.

Within these broad categories, there is much further possibility for

variety of function. is variety can be demonstrated by examining

a few specic examples from the Near East and Aegean, where we

have clear indications of function from contextual, iconographic, and

textual evidence:

1. Neo-Assyrian apkallu gurines: At a number of Assyrian sites (e.g.,

Nimrud, Khorsabad, Nineveh, and Assur), several types of clay gu-

rines (including a human gure with a bird’s head, a “bearded war-

rior” with spear painted on the body, and a bearded gure with a sh

hood on its head) have been found, typically under the oors of pal-

aces (Mallowan 1954; Oates 1959; Green 1983). Most correspond in

364 michael d. press

type to gures found on wall reliefs from Assyrian palaces, although

the two types of representations are generally not found together.

At the same time, a number of texts have been found describing the

manufacture of these gurines (Wiggermann 1992), matching the

excavated gurines in great detail. e texts indicate that the gu-

rines were apotropaic, used for protection of rooms of the palace.

2. Greek mourning gurines: From a few Mycenaean sites in the

twelh century, and again from the Late Geometric period and

especially the Archaic, we have a type of gurine representing a

woman holding her hands to the hand (and, less commonly, other

gestures such as hands to the cheek or the breast) (see Iakovidis

1966; Kurtz and Boardman 1971; Dothan 1982: 244). ese gu-

rines are always, or almost always, found in tombs. Similar gures

are depicted on other items from burials, such as Mycenaean lar-

nakes (e.g., Vermeule 1965; Iakovidis 1966: 46–49; Cavanagh and

Mee 1995: 46–51, 60–61, Figs. 1–10) and various forms of painted

pottery from the Late Geometric on (Zschietzschmann 1928; Board-

man 1955; Ahlberg 1971; Cavanagh and Mee 1995: 51–55). e ico-

nography of the painted scenes, together with the contexts, suggests

that the gurines are mourning gurines—used as a symbol of con-

tinuing mourning for the deceased aer burial. is conclusion is

further strengthened by references to mourning gestures, parallel-

ing those found in the gurines almost exactly, in Greek literature

(for surveys, see Mylonas 1963: 476; Ahlberg 1971: esp. 264–65;

Immerwahr 1995: 110).

3. Egyptian ushabtis: Made of various materials (clay, stone, wood,

or faience), these objects were common in tombs from the Middle

Kingdom to the Late Period. eir contexts, along with textual and

other iconographic evidence, make their function clear: they served

as stand-ins for the deceased, to perform manual labor in the aer-

life (see, e.g., D’Auria et al. 1992).

4. Egyptian faience gurines: While the nd contexts of these artifacts

are not as distinctive as those of the other objects above, their ico-

nography, occasional inscriptional evidence, and physical features

(their small size and, commonly, piercing) suggest a common func-

tion to most if not all examples—they were amulets to be worn

for protection (see, e.g., Herrmann 2003: esp. 2). Note that, while

deities such as Bes and Isis are represented, they do not function

as cultic; rather, the deities are chosen for their appropriateness for

protection.

a problem of definition 365

I have selected these examples because they are some of the best cases

known to me where the function of the gurines is clear, from a com-

bination of dierent types of evidence. In none of these cases can the

gurines be said to be strictly cultic. Instead, they serve as protective

gures, grave goods, servants—at most, they can be said to be religious

only in a broad sense (see below).

erefore, while gurines can in some instances be cultic (there are

of course cases where gurines are found as votive objects in sanctuaries

for example), such an interpretation can by no means be assumed. For

Mycenaean gurines, French (1981b: 173) has in fact concluded that

gurine function must be determined from context, and not vice versa

(cf. Tzonou-Herbst 2002: esp. 18; Nilsson 1968: 307–309). is con-

clusion has signicant implications, for gurines are the objects most

commonly assumed to be strictly cultic in nature. As a result, the need

for a more rigorous methodology for cultic identication is clear.

Suggestions toward a Solution

Given the above outlining of the problem—in short, the lack of a

clear methodology for identifying cultic activity (of any type, let alone

domestic cult specically)—we can now begin to address it. Originally,

in approaching this problem I had hoped to provide a more denitive

methodological statement. In recent years there has been a wealth of

attention given to the archaeology of religion and cult generally,

and

these provide a wide array of methodological approaches. At the same

time, this array is so wide that it becomes hard to single out an eective

method or set of methods, or even to be sure if scholars are addressing

the same problem. As a result, my goal here is simply to work toward

a solution, by making a set of methodological suggestions.

Dening Denitions

e rst suggestion, which serves as an initial step in the process, is to

dene basic terms. For the identication of cult, the three terms I will

e literature is too vast to cite in full here, but note especially a series of recent

conference volumes on the subject: e.g., Biehl et al. 2001; Insoll 2004; Kyriakidis

2007d; Barrowclough and Malone 2007; Rowan forthcoming.

366 michael d. press

be concerned with are religion, ritual, and cult. Commonly, in studies

of religion of ancient Israel, one or more of these terms is insuciently

dened. In addition, I believe that many archaeological analyses and

methodological discussions of “cultic” contexts are weakened speci-

cally by their unclear or contradictory denitions of these terms (see

below; cf. Darcque 2005: 294).

For example, Dever has written a series of articles (e.g., 1983, 1987,

1994) and a book (2005) on the subject of ancient Israelite religion

yet has little general discussion of methodological issues. Neverthe-

less, he does at times provide a theoretical treatment of religion (esp.

2005: 1–31; also 1983: 572). In these treatments, he briey summarizes

denitions of “religion,” based on those of anthropologists and biblical

scholars (especially Z. Zevit), and highlights key concepts (Dever 2005:

1–2). Dever, however, draws a simple dichotomy between “theology”

and “cult,” paralleling his distinction between “belief” and “practice”

(and other dichotomies, such as “ocial” vs. “popular” religion) (Dever

2005: 5–6). Besides limiting cult to “popular” religion, and divorcing

theology and belief from it, Dever leaves no room for any type of reli-

gious ritual (a term he does not dene) beyond cult.

Zevit provides an extended discussion of religion and an attempt at

denition (2001: esp. 11–17). Nevertheless, despite dealing with meth-

odological issues about “cult practice” and “cult places” (2001: 81–84),

and dening a set of terms related to cult (cult place, cult room, cult

corner, cult cave, cult complex, cult center, and cult site; 2001: 124–125),

nowhere does Zevit dene “cult” itself (or ritual).

Before continuing, it will be necessary for me to briey dene, or

indicate what I mean by, the terms religion, ritual, and cult. It will

of course be impossible to arrive at “denitive” denitions of these

terms. In fact, in highlighting this denitional problem, some inu-

ential scholars (for example in ritual studies) have argued against

the focus on denitions altogether (Goody 1961, 1977; Bell 2007). I

believe, however, that such explicit denitions are helpful, and even

necessary, for two basic reasons: First, they are necessary in each

individual study, so that the reader can understand what the author

means. Terms such as those under consideration here can be especially

susceptible to multiple meanings, and so the use of general “pointers”

at least (cf. Rappaport 1999: 23), if not precise denitions, becomes

essential. Second, on a more general level, we as scholars need a com-

mon vocabulary, as opposed to having our own personal sets of de-

nitions, so that we can have a dialogue, so that if we debate whether

a problem of definition 367

something is “cultic,” we can be certain that we are discussing the

same thing—that we are having a substantive discussion rather than

simply a semantic one (cf. Kyriakidis 2007a: 289). (At the same time,

Bell’s comments [2007: 283] on the futility of trying to arrive at a

common denition are to some extent valid; I would therefore view

the rst of the two reasons above as more fundamental.) Moreover,

these denitions should not merely be academic exercises, but valid

working statements with some analytical utility in the specic study

in question.

For the purposes of this study, I will dene religion by repeating the

denition of Zevit (2001: 15), who modies that of Cavanagh (1978:

16–19), for “Israelite religions”: “the varied, symbolic expressions

of, and appropriate responses to the deities and powers that groups

or communities deliberately armed as being of unrestricted value

to them within their worldview.” I do not necessarily agree entirely

with this (or any other) specic denition of religion, but use it for

its “pointing” value with concepts such as deities/powers, symbolic

expressions and responses, and worldview. As for ritual, I will fol-

low Rappaport’s denition as “the performance of more or less invari-

ant sequences of formal acts and utterances not entirely encoded by

the performers” (1999: 24). While Bell, as noted above, has advocated

against an actual denition of “ritual,” she has formulated six charac-

teristics of “ritual-like activities” (1997: 38), which follow those sug-

gested by Rappaport’s denition: formalism, traditionalism, disciplined

invariance, rule- governance, sacral symbolism, and performance.

While Bell, among others, has suggested that ritual is principally, if not

exclusively, religious in nature (1997: 94; 2007: 278), many others have

suggested that ritual can describe both religious and secular behavior

(e.g., Rappaport 1999: 25; cf. Kyriakidis 2007b: 6 n. 2). In addition,

even among religious ritual there is a wide variety of practice. Cult,

then, would best be limited in scope not just specically to religious

ritual, but to ritual involved with worship of a deity.

In trying to identify cultic practice in the archaeological record, we

can use the above denitions and concepts to suggest at least two types

of behavior to look for (i.e., behavior whose material correlates we

should look for) archaeologically:

1. Habitual activity (as opposed to one-time activity). is type of

behavior is suggested by the characteristics of invariance and for-

malism noted above. We might therefore expect an accumulation

368 michael d. press

of similar types of artifacts in a cultic context, as well as for the

same cluster of artifacts and features to be repeated in dierent

cultic contexts within a cultural area.

2. Actual act of worship. Cultic practice, based on the denition

above, would only be a subset of all religious practice or ritual.

erefore, something may be reective of religious ideals, a symbol

or reminder, or be related to observance of life-cycle events (such

as death), and not be cultic.

Dening Analogies

In his archaeological analysis of ancient Israelite religion, Holladay

writes: “All archaeological interpretation necessarily proceeds, in one

manner or another, on the basis of ethnographic analogy” (1987: 250).

While much archaeological argument might not be characterized as

“ethnographic,” I would otherwise agree completely with Holladay’s

statement. In any analysis that we as archaeologists perform, we are

constantly comparing artifacts and features to something else; but to

what? Is the basis of our analogy ethnographic? Is it ethnohistori-

cal? Is it anecdotal? Is it based on something from the specic cul-

ture in question, or from cultures proximate in space or time, or is

it drawn entirely from the imagination? In my view, such analogies

must be made explicit, so that their validity can be evaluated by other

scholars—and by the authors themselves.

Dening Methods

In order to arrive at suggestions toward an actual method, we must

rst survey (briey) the methods that have already been suggested for

identifying cult. Unlike the earlier work on the archaeology of religion

of the southern Levant, discussed above, several studies over the last

two decades have made such statements of method. ese statements

can be categorized in two broad groups: those using a hypothetico-

deductive method, and those using an inductive method. To simplify

somewhat, the former use a list of characteristics identied in cult

generally and test them on a specic case; the latter eschew any such

generalizations, but instead try to identify cult on a case-by-case basis

through knowledge of the specic culture.

a problem of definition 369

Hypothetico-deductive Methods

Coogan (1987) made one of the earliest explicit statements of a method

for identifying a cultic context, at least for southern Levantine archae-

ology. He proposed four basic criteria (when textual evidence is lack-

ing): isolation, exotic materials, continuity, and parallels (1987: 2–3).

Coogan also suggested that the criteria be used in terms of probabil-

ity: the more criteria satised by a given context, the likelier it would

be that the context is cultic. In fact, there is nothing really new in

Coogan’s criteria themselves. As mentioned above, McCown (1950:

208) briey highlighted the characteristics of “cult remains,” continu-

ity, and plan (if not explicitly parallels in plan); and Ottosson (1980:

9–10) emphasized a temenos, separating the sacred from the secular,

as a key characteristic of temples. In addition, these criteria in some

form were used to make most of the identications of cultic contexts

in the archaeological record; the explicit, systematic statement of these

criteria, however, marks Coogan’s study as dierent.

Nevertheless, there is a fundamental problem with Coogan’s presen-

tation: while the criteria are explicit, their basis is not; nor is the deni-

tion of the term “cult” stated. At certain points Coogan gives examples

of cultic contexts for dierent characteristics, but these are in fact

oen drawn from archaeology—in other words, their cultic character

is generally not certain, but has only been determined through the use

of such criteria, which Coogan is now using to prove cultic nature of

contexts. For the criterion of continuity, Coogan does provide modern,

proven examples from Damascus and Jerusalem. Overall, the number

of examples is not extensive, and oen Coogan uses them in a circular

fashion—both as examples to prove a specic criterion, and then as

illustrations of the overall set of criteria.

In his publication of the Phylakopi sanctuary, Renfrew (1985) pro-

vides the most detailed and comprehensive review of the problems

of cultic identication. e study is notable both for its introductory

essay, which attempts to establish a framework for identifying and

understanding cult archaeologically (1985: 11–26), and for an evalu-

ation of the proposed sanctuary according to this framework aer a

presentation of the evidence (1985: 361–391). Beyond these elements,

Renfrew’s Phylakopi publication is signicant for the inuence it has

had on discussion of the nature of cult and its identication in the

archaeological record. Signicant studies of cultic contexts have fol-

lowed Renfrew’s framework, on mainland Greece (Piladis-Williams

370 michael d. press

1998; Morgan 1999; Moore and Taylour 1999; Konsolaki 2002), Crete

(Mersereau 1993: 17–18; Prent 2005), and Cyprus (Bolger 1992: 150–

152; Smith 1997; Webb 1999: 13–15 for discussion, but not direct

application). For the Levant, the only direct applications of Renfrew’s

framework (to my knowledge) are Levy’s for the site of Gilat (2006)

and (to a lesser extent) Zevit’s survey of Iron Age cult places in Israel

(2001: 125–247); discussions or references of varying lengths, how-

ever, have been provided by Gilmour (1995: 5–6), Frick (2000: 12–14),

Zevit (2001: 82–83), and Dothan (2003: 191).

In the treatments mentioned above, Renfrew’s framework is gen-

erally applied without comment, or else modied for application on

types of cultic sites other than temples. It would therefore be worth-

while to evaluate that framework briey. roughout his introduc-

tory essay, Renfrew emphasizes explicitness: for him, “each step in

the argument must be open to examination” (1985: 11). He provides

clear denitions of the three key terms identied above: religion, ritual,

and cult. For Renfrew, ritual can be both religious and secular; cult

is specically related to the ritual of religious worship (1985: 14–15).

e heart of Renfrew’s analysis is his formulation of “aspects of sacred

ritual” (1985: 16–17), involving “frequent features of cult practice,”

considered “on a cross-cultural basis”: the asymmetrical relationship

of worship, oerings, a special location (either natural or built), and

a focus of attention. He then produces a set of 18 behavioral corre-

lates of these concepts (1985: 18–19), which might be reected in the

archaeological record. roughout, Renfrew is cautious, noting vari-

ous potential aws in his framework: for instance, he observes that

these correlates could identify not only cultic activity, but also secular

ritual, or even nonritual symbolic systems (1985: 20). is caution and

self-criticism are an additional characteristic that recommends Ren-

frew’s presentation.

Ultimately, however, Renfrew’s discussion exhibits some of the

same problems found in other discussions. One is that his deni-

tions, while explicit, are not consistently applied. For instance, while

Renfrew’s denitions suggest dierentiation between cultic and other

types of religious ritual, his subsequent discussion conates the two,

suggesting that all religious ritual is cultic (cf. Darcque 2005: 294).

us, the Agia Triada Sarcophagus, while interpreted by Renfrew as

funerary ritual and not worship, is still labeled as cultic “in the broad

sense” (1985: 25). is conclusion therefore leads to confusion in both

his correlates and his analysis, as it is unclear exactly what he means

a problem of definition 371

by arguing for something as “cultic.” In addition, while Renfrew sug-

gests that the features of cult that he outlines are found frequently

and cross-culturally, he in fact gives no explicit basis or examples for

these features. He routinely characterizes these features as occurring

“[i]n most societies” or as “commonly used,” but nowhere does he

justify such statements. In terms of both of the suggestions I have

made so far toward a solution for identifying cult—explicit denitions

and explicit analogies—Renfrew’s discussion, while admirable, is ulti-

mately lacking.

Inductive Methods

Some scholars have criticized Renfrew from a Post-Processualist per-

spective. Even among these, however, there is common acknowledg-

ment that Renfrew’s study is an advancement over typical discussions

of identifying cult (e.g., Vella 2000: 29–30; Darcque 2005: 294). Some

have rejected Renfrew’s “trait-list approach” (Begg 1991: 9) and his

attempt to generate overgeneralized rules (Begg 1991: 9–10; Hodder

1992: 152; Vella 2000: 30). us, in critiquing Renfrew’s framework,

such scholars have emphasized the need to focus instead on the evidence

within the context of the particular culture in question (Begg 1991: 10;

Hodder 1992: 152; Darcque 2005: 295). As an alternative to Renfrew’s

framework, Vella, building on the work of Hodder, has advocated one

based on “a logical process wherein there is a double movement from

data to interpretation and back again, where an argument is cumula-

tive” (2000: 30). is is similar to the hermeneutic method advocated

for an archaeological application by Schloen (2001), as well as to the

art-historical method of Erwin Panofsky (e.g., Panofsky 1955).

Warren, meanwhile, has suggested that Renfrew’s analysis is ulti-

mately not deductive but in fact inductive (1986: 155). A similar obser-

vation has been made by Hodder (1992: 152). While Renfrew tries to

set up objective, general criteria by which Phylakopi can be recognized

as a sanctuary, he must in the end resort to comparison with other

similar structures from the Late Bronze Aegean—what Hodder labels

as a “hermeneutic position” (1992: 152). Hodder has further argued

that most archaeological argument is in fact not “positivist hypothesis-

testing” at all, but rather hermeneutics (1992: 213). As an alternative,

Hodder proposes his own model for dealing with the potential vicious-

ness of the hermeneutic circle (1992: 213–240). In the end, however, I

nd it hard to detect serious dierences between Hodder’s model and