. 18
( 55 .)


of the horns of the altar, obtaining sanctuary from an avenger in hot pur-
suit, according to the evidence of Exod 21:13“14 and 1 Kgs 2:28. Then,
during the Deuteronomic reform in the seventh century b.c.e., in order to
prevent the loss of the institution of asylum when local altars were abol-
ished, Josiah appointed special cities of refuge (Deut 19:1“13). Wellhausen
concludes that the Priestly legislator af¬rmed this arrangement and speci-
¬ed six cities of refuge, three on each side of the Jordan (Num 35:9“34;
Josh 20:1“9).2 Since four of these cities were cultic sites known from other

1 JuliusWellhausen, Prolegomena to the History of Israel (preface by W. Robertson Smith;
foreword by Douglas A. Knight; 1885; reprint, Atlanta: Scholars Press, 1994), 33, 162, 375.
2 Wellhausen, Prolegomena, 162, asserts that Deut 4:41“43, which speci¬es three cities in Trans-

jordan, could not be original. He does recognize that Joshua 20 contained a Priestly kernel


biblical texts and the cities of refuge were also Levitical cities, it seemed ob-
vious to Wellhausen that the special status of the cities of refuge was linked
to the presence of an ancient altar. He recognized, to be sure, that not every
sacred site was a city of refuge. Some that became infamous as sites of pa-
gan worship, such as Bethel, Dan, Gilgal, and Beersheba, were intentionally
Those who have followed in Wellhausen™s wake have argued for minor
nuances in his scheme of development. The date of the establishment of the
cities of refuge has shifted back and forth. Max Lohr pushes the date back
to the reign of either David or Solomon and argues that the early monar-
chy sought to limit blood revenge by means of the cities of refuge.3 Roland
de Vaux is of the opinion that the cities of refuge dated back to the reign
of Solomon: He believes that although the texts that refer to them may be
Deuteronomic, one, Josh 20:1“9, rests on older traditions because the cities
it lists were under Israelite control only during Solomon™s reign.4 Milgrom
agrees with a Solomonic date and argues that the cities of refuge were es-
tablished by Solomon, who innovated a whole host of cultic procedures,
including a type of altar that precluded altar asylum.5
Others have stood ¬rm on a late-seventh-century Deuteronomic date.
N. M. Nicolsky holds that the cities of refuge did not exist before the Deutero-
nomic revolution of the latter seventh century.6 Henry McKeating argues
that the cities of refuge were established in the seventh century when the
authorities were still trying to regulate, not oust, clan-based justice.7
All these arguments are based on the assumption of institutional change
from altar asylum to cities of refuge, whether during the ¬rst century of the
monarchy or the Deuteronomic revolution. It is worthwhile, therefore, to
assess the evidence for radical institutional change.
First, we will examine the argument that in the earliest stage of legal
development in ancient Israel, a sacri¬cial altar served as a refuge for a
slayer: A killer would ¬‚ee to an altar and remain safe as long as he remained
there. This view has been based on texts recounting the ¬‚ight of Adonijah,
one of David™s sons, and the ¬‚ight of Joab, David™s army commander, to the
altar in the Tent-shrine during Solomon™s reign (1 Kgs 1:50“53; 2:28“34)
and on the statute on homicide in the Covenant Code (Exod 21:12“14).

surrounded by Deuteronomic retouching and explains that later additions to the Pentateuch
imitated Deuteronomic language (Prolegomena, 375). His refusal to recognize these additions
as products of Deuteronomy and thereby to redate Deuteronomy after P is, no doubt, due to
his belief in the late date of P. In his defense, it must be noted that passages in Deuteronomic
style are found in texts, such as Jeremiah, that are certainly not part of Deuteronomy.
3 Max Lohr, Das Asylwesen im Alten Testament (Halle: Max Niemeyer, 1930), 35.
4 De Vaux, Ancient Israel, 162“163.
5 Milgrom, Numbers, 506“507.
6 N. M. Nicolsky, “Das Asylrecht in Israel,” ZAW 48 (1930), 156“157.
7 McKeating, “Development of the Law of Homicide,” 64“66.

This interpretation overlooks the fact that Adonijah is not portrayed as
ever having killed another person. He seeks sanctuary because he is a political
offender. He attempted to claim the throne during David™s reign, and as a
direct result of this act, Solomon was proclaimed by David as the rightful
heir. Adonijah, now fearing for his life, ¬‚ees to the altar and holds on to its
“horns.” Solomon promises that no harm will come to Adonijah as long as
he behaves. (Of course, Adonijah cannot resist and asks for one of David™s
concubines, a request that questions Solomon™s legitimacy, another political
misdeed, and he is ¬nally put to death.) When Adonijah ¬‚ees to a sanctuary,
he is seeking refuge from his political predicament, not from committing a
Joab™s ¬‚ight has similarly been misinterpreted. Although his act has been
interpreted as seeking sanctuary from the penalty for homicide, the text
re¬‚ects a different motive for his ¬‚ight. Joab had sided with Adonijah,
Solomon™s rival for the throne (1 Kgs 2:28), and when he hears the news
that Solomon has executed Adonijah and has dismissed Abiathar, a sup-
porter of Adonijah, from his post as priest, Joab realizes that it is political
payback and takes refuge at the Tent-shrine. It does him little good: He is
taken from the sacred precincts and killed. Joab™s seeking sanctuary at an
altar was politically motivated. It was not based on evading criminal culpa-
bility for a killing. Altar asylum protects from political intrigue, not from
the punishment for homicide.8
Seeking respite from political enemies is re¬‚ected in a curious episode
in Nehemiah™s memoirs.9 During the Persian period, a prophet advises
Nehemiah to protect himself from those sent to kill him in the night by
taking shelter in the Temple (Neh 6:10“13). Nehemiah refuses, object-
ing that a person in his position should not show cowardice by taking

8 …ke Viberg, Symbols of Law: A Contextual Analysis of Legal Symbolic Acts in the Old Testa-
ment (Coniectanea Biblical Old Testament Series 34; Stockholm: Almqvist & Wiksell, 1992),
122ff., recognizes that Adonijah is not seeking asylum for homicide and that his action has
no apparent relation to asylum for homicide. Viberg also realizes that Joab™s case was prob-
lematic for two reasons: 1) Why would Joab have sought asylum since the type of homicide
he committed would have allowed him to be removed from the altar? 2) If Joab committed
culpable homicide, why would there have been a problem with removing him from the altar?
Viberg solves this problem by arguing that Joab was appealing for some form of general asylum
whereas Solomon was applying Exod 21:12“14, albeit incorrectly, since Joab was struck down
at the altar without any indication that he was removed. Of course, the answer may be simpler:
Joab thought that seeking asylum at the altar would or might protect him, and Solomon either
thought it did not or simply killed Joab anyway. The narrator™s framing of Joab™s offense as the
shedding of the blood of war in peacetime makes Joab™s actions appear illegal. However, taking
vengeance on others for what they did in war is considered acceptable, according to Judg 8:
9 The use of the Temple as a sanctuary from political machinations is attested for the First

Temple period as well. Joash is kept hidden from his mother Athaliah within the Temple after
she has killed everyone else of royal stock (2 Kgs 11:2“3).

¬‚ight.10 As far as he is concerned, sanctuary asylum still operated in
his day.
The theory that altar asylum was abrogated is also based on the reference
to homicide in the Covenant Code, and it must be stated at the outset that the
Covenant Code™s statute is enigmatic. The evidence for the changeover from
altar asylum to cities of refuge is found in the early dating of the Covenant
Code and its reference to taking refuge at an altar. This evidence is in fact
Scholarly consensus identi¬es “the Covenant Code,” Exod 21:1“23:19,
as an independent collection of laws incorporated in a larger literary unit,
the Book of the Covenant. Although nineteenth-century scholarship was
divided on whether the Covenant Code belonged to the Pentateuchal sources
J11 or E,12 there was a wide consensus during the past century that the
Covenant Code was independent of J and E13 and was a product of the
tribal period. Since the Covenant Code does not contain any references to
the monarchy and re¬‚ects the circumstances of small landholders, the period
of the judges when the Israelites were settled on patrimonial estates in their
tribal territories seemed appropriate.14 The tribal period was considered

10 Nehemiah does not object that entering the Temple precincts or the inner sanctuary would
make him liable to be killed (Milgrom, Numbers, 507). Nehemiah does enter the Temple
precincts later in the narrative in order to remove the furnishings of Tobiah™s room (Neh 13:8).
Those who were not cultic personnel were barred from the area of the altar in both the First and
Second Temple. When concerns arose about the apparent misappropriation of money brought
into the First Temple, the solution was to set aside a box for the collection on the right side
of the altar (2 Kgs 12:5“12). Since the laity did not have access to the area of the altar, the
Temple guards, who were priests, were made responsible for taking the money from the laity
and depositing it in the collection box. 2 Chr 26:16“20 explains Uzziah™s leprosy as a result of
his encroachment upon the prerogative of the priests to offer incense upon the altar. However,
according to 1 Kings 2, Adonijah and Joab do enter the area of the altar and take hold of it,
contradicting this prohibition. The fact that the area of the altar was forbidden to the laity in
the First and Second Temples may not hold true for pre-Temple shrines, like the Tent-shrine in
pre-Temple Solomonic Jerusalem.
11 Wellhausen, Prolegomena to the History of Israel, 12“13, 392, considers it to be an expression

of a legislative element of the Jahwist, a document of the main part of the Assyrian period,
ca. 800 b.c.e.
12 Adolf Julicher, who argues for the relationship of the Covenant Code to E, is referred to
in Bruno Bantsch, Das Bundesbuch, Ex. XX 22“XXIII 33 (Halle: Max Niemeyer, 1892),
13 Bantsch was the ¬rst to argue for the independence of the Covenant Code from the docu-
mentary sources (Das Bundesbuch, 73). He holds that the individual !yfp`m were known to J at
the end of the ninth century, while the collection of them was known to E in the middle of the
eighth century (Das Bundesbuch, 122). He also holds the opinion that the !yrbd were a product
of the prophetic movement of the mid-eighth century (Das Bundesbuch, 121).
14 Cf. Paul, Studies in the Book of the Covenant, 44; Childs, The Book of Exodus, 456.

Ludger Schwienhorst-Schonberger, Das Bundesbuch (BZAW 111; Berlin: Walter de Gruyter,
1990), 268, distinguishes between the casuistic law of the eleventh“tenth centuries, later put
together as a book in the ninth“eighth centuries, and then placed as part of divine law in the

to be the least socially differentiated period in Israelite development. Legal
thought and institutions were at their least developed as well. According to
this opinion, the dating of the Covenant Code to the tribal period would
be the launching point for extrapolating earlier stages in the adjudication of
The consensus on the date of the Covenant Code has been undermined
in recent years. One line of attack on a tribal period for its dating has fo-
cused on recognizing reformist elements within the Code. Arguments for an
eighth-century date utilize three proofs highlighting a reformist tendency:
1) The amalgam of literary forms that makes up the entire complex of the
Covenant Code suggests the dissolution or deliberate combination of for-
merly separate legal traditions.15 2) The presence of laws protecting the
poor and regulating slavery for debt is evidence for a date well into the
monarchy because there would be no need for them during the poor eco-
nomic conditions of the prestate period.16 Slavery gained signi¬cance only
during the later monarchic era. 3) The Covenant Code in its attention to
the alien presupposes widespread population shifts, which would ¬t well
with the serious refugee problem dealt to Judah after the fall of the northern
However, the proofs for an eighth-century date are ¬‚awed. While it is true
that the laws in the Covenant Code appear to have barely been reworked
from their original formulation because texts of varying literary form had
been placed side by side without any attempt to make thematic or linguistic
unity, it is, however, possible to offer many different points in Israelite history
when the combination of formerly separate legal traditions could have oc-
curred. Other considerations are also equivocal. First, laws addressing social
issues appear in all the legal corpora of the Bible and would be applicable
in almost any period of Israelite history. Second, the social conditions that
would incur debt slavery were prevalent during most of the period of the

eighth“seventh centuries. Henri Cazelles, “L™Auteur du Code de l™Alliance,” RB 52 (1945),
188, argues for a slightly earlier date, holding that Moses is the direct source.
15 Frank Crusemann, The Torah: Theology and Social History of Old Testament Law (trans.
Allan W. Mahnke; Edinburgh: T & T Clark, 1996), 165“169; Rainer Albertz, A History of
Israelite Religion in the Old Testament Period (trans. John Bowden; 1992; reprint, Louisville,
Kentucky: Westminster/John Knox, 1994), 1.183“184.
16 One problem with this claim is the references to “slave” in biblical literature depicting

the premonarchic period. Crusemann argues that the term db[ in Joshua, Judges, and 1 and
2 Samuel was most often used as a polite designation for a subordinate speaking to a superior
and, therefore, the use of the term for “slave” must date from the later centuries of the monar-
chy, not the tribal period nor the beginning of the monarchy (The Torah, 152). The real con¬‚ict
during the early monarchy, according to Crusemann, was between the king and the people, not
between free and slave, according to the evidence of 1 Sam 8:16“17. However, there were slaves
even during the premonarchic period, and the word used for them was db[ “ there is no other
term in Hebrew.
17 Albertz, A History of Israelite Religion, 1.183, 336, n. 166.

monarchy as re¬‚ected in both historical texts (e.g., 2 Kgs 4:1“7) and in the
social criticism of the prophets.18 Third, poor economic conditions were not
the only circumstances under which a person might be sold into slavery. A
thief who could not pay the penalty for his theft would become a slave: The
thief™s illegal act and his inability to pay, presumably because he was poor,
were the circumstances that caused him to be sold into slavery. Fourth, the
absence of a king could be taken as surprising if the Deuteronomic laws,
more ¬rmly situated in the monarchic period, were replete with references
to the royal establishment, but Deuteronomy makes only a few references to
the institution of kingship (Deut 17:14“20). Fifth, the agrarian nature of the
society re¬‚ected in the Covenant Code cannot serve as proof of an early date
because Israelite society persisted in remaining agriculturally based in pat-
rimonial estates throughout the First Temple period.19 Lastly, the Covenant
Code does not exhibit archaic language, a characteristic expected of early
texts and the only sure proof of an early date.20 In general, it is dif¬cult to
determine a speci¬c date for the Covenant Code. It appears, then, that its
date cannot be chronologically set within any speci¬c time during the First
Temple period.
The second issue is whether the refuge mentioned in the Covenant Code
is an altar. In the statute, there are two references to places involved in the
adjudication of homicide, one of which appears to be a safe haven, the other
a place from which a killer can be taken. Exodus 21:13“14 reads, “If [the
killer] did not do it by design, but God caused it to meet his hand, I will assign
you a place to which he can ¬‚ee. When a man schemes against another and
kills him treacherously, you shall take him from my very altar to be put to
death.” The referent of “a place to which he can ¬‚ee” in verse 13 is the
crux of the matter. Does it refer to the altar mentioned in the following
verse? Some have argued that the implication of taking away a killer even
from God™s altar is that the place of asylum in Exod 21:13 is identical to
the altar mentioned in Exod 21:14.21 However, if the statute in the ¬rst
verse was referring to the “altar,” why did it not simply state “altar”?22 The
implication, thus, is that !˜qt and jA“¿y were distinct. If so, two possibilities
arise: 1) !˜qt and jA“¿y were two completely different places; 2) one was part
of the other, the jA“¿y being part of the !˜qt.


. 18
( 55 .)