Testaments 1 (1869), 466‚Ä“477.
82 HOMICIDE IN THE BIBLICAL WORLD
determination has been made on a linguistic basis. Certain Priestly terms
were not used during the postexilic period, others were replaced by synonyms
in the postexilic period, and others experienced a change in meaning in the
Second Temple period that contradicted their preexilic meaning.41 According
to linguistic evidence, the P source dates from the First Temple period. H,
which incorporated P, can be dated by linguistic criteria as well to the exilic
or early Persian period.42
The third main text on homicide is Deut 19:1‚Ä“13. Deuteronomy as a
whole is a striated book, and although scholars have been able to divide
Deuteronomy into strata, they have had difÔ¬Āculty linking them to particular
time periods. Deuteronomy has correctly been linked to the cultic reform of
Josiah in the late seventh century b.c.e.43 However, the differences between
Josiah‚Ä™s reform and the book of Deuteronomy, as we now have it,44 force
us to admit that while the book certainly is related to the Josianic reform,
it is unclear which literary accretions of Deuteronomy are directly tied to it.
C. Steuernagel and J. G. Staerk separated out the layers of the text on the
basis of the Israelites being addressed in the singular or in the plural,45 but
the divided text cannot be dated more precisely. The structure of Deuteron-
omy as a whole has many afÔ¬Ānities to the Vassal Treaties of Esarhaddon
(672 b.c.e.). Deuteronomy has a distinctive literary style, and from the sev-
enth century onward, the historiographic and prophetic texts in the Bible
exhibit many of this style‚Ä™s features. Deuteronomy is deÔ¬Ānitely linked to the
seventh century, but whether any particular stratum of Deuteronomy pre-
dates or postdates the seventh century is sheer guesswork.
The main Deuteronomic text on the adjudication of homicide, Deut 19:1‚Ä“
13, can be divided into layers,46 but these strata can be assigned a date only
on a relative basis. If verses 1 and 9 are Deuteronomic, then verses 5 and
41 Cf. Avi Hurvitz, A Linguistic Study of the Relationship Between the Priestly Source and the
Book of Ezekiel (CahRB 20; Rome: PontiÔ¬Ācal Biblical Institute, 1982); Hurvitz, ‚ÄúDating the
Priestly Source in Light of the Historical Study of Biblical Hebrew a Century After Wellhausen,‚ÄĚ
ZAW 100 (1988), 88‚Ä“99; Jacob Milgrom, Studies in Levitic Terminology I (Berkeley: University
of California Press, 1970), 8‚Ä“16, 60‚Ä“87.
42 There are some indications that H may include additions from the exilic and early Persian
periods. Cf. Israel Knohl, The Sanctuary of Silence (Minneapolis: Fortress, 1995), 201‚Ä“204;
Jacob Milgrom, Leviticus 1‚Ä“16 (AB; New York: Doubleday, 1991), 27. It should be noted that
the priority of P or H relative to one another has no effect on my analysis.
43 W. M. L. de Wette, Dissertatio critica exegetica qua Deuteronomium a prioribus Pentateuchi
libris diversum, alius cujusdam recentioris actoris opus esse monstratur (Halle, 1805).
44 For example, in contrast to 2 Kgs 23:20, there is no indication that the priests of the bamot
must be killed. Deut 18:6 and 2 Kgs 23:9 directly contradict each other. The title of the lawbook
in Josiah‚Ä™s reform, tyrbh rps, does not refer to Deuteronomy but to the book referred to in
45 C. Steuernagel, Der Rahmen des Deuteronomium (2d edition; 1894; reprint, Halle: Max
Niemeyer, 1923); W. Staerk, Das Deuteronomium (Leipzig: J. C. Hinrichs, 1894).
46 A. D. H. Mayes, Deuteronomy (New Century Bible Commentary; Grand Rapids, Michigan:
Eerdmans, 1979), 283‚Ä“285, 297.
DEVELOPMENT OF PLACES OF REFUGE IN THE BIBLE
8 can be considered post-Deuteronomic, because they refer to Israel as a
whole rather than to a speciÔ¬Āc community waiting to enter the land of Israel
and because they use a late expression to refer to the priests, ywl ynb !ynhkh.
There also may be pre-Deuteronomic law in verses 4‚Ä“5, 11‚Ä“12, assuming
that casuistic formulation indicates earlier material, and post-Deuteronomic
supplements in verses 8‚Ä“10, a contradiction to 4:41‚Ä“43.
The biblical laws on homicide, Num 35:9‚Ä“34 and Deut 19:1‚Ä“13, as
well as Exod 21:12‚Ä“14, are mutually independent sources. They do not
have a common literary origin, and they stem from diverse historical and
ideological/theological settings. Although previous studies of biblical law
have assumed that each of the legal parts of the Pentateuch can be securely
dated to different periods and have devised schemes of historical develop-
ment in biblical law based on that dating, the legal portions of the Pen-
tateuch cannot be dated with such precision and, in fact, all date from
some time in the First Temple period with no clear evidence for historical
The legal sources in the Pentateuch, P and D, that differ on the character-
istics of the places of refuge do so because they conceptualize the sanctuaries
within their ideological/theological program. They disagree on the number,
the sacred status, the rationale, and the existence of a technical term for the
cities of refuge.
The statute in Num 35:9‚Ä“34 calls for the establishment of six cities as
refuges for the slayer from !dh lag, denoting them by the technical term
47 The narrative sources are also difÔ¬Ācult to date. Two examples, the killing of Joab (1 Kgs 2:5‚Ä“
6, 28‚Ä“34) and the execution of those who assassinated Joash (2 Kgs 14:5‚Ä“6), can illustrate the
quandary. The text of 1 Kgs 2:5‚Ä“6, 28‚Ä“34 forms part of the Succession Narrative (2 Sam 9‚Ä“20;
2 Kgs 1‚Ä“2), a product of an author in Solomon‚Ä™s court who collected a number of independent
stories and wove them together in order to justify the new king‚Ä™s accession to the throne. (Cf.
Leonhard Rost, The Succession to the Throne of David [trans. by Michael D. Rutter and David
M. Gunn; intro. by Edward Ball; 1926; reprint, ShefÔ¬Āeld: Almond Press, 1982]). The rest of
Samuel is assigned to the earlier reign of David. However, can the Succession Narrative so
easily be separated from the rest of Samuel? David‚Ä™s admonition to Solomon to kill Joab, 1 Kgs
2:5‚Ä“6, a part of the Succession Narrative, depends directly on the murders of Amasa and Abner
in 2 Sam 2:18‚Ä“23 and 3:28‚Ä“30, which are not identiÔ¬Āed as part of the Succession Narrative.
More critically, the dating of these texts to the reigns of David and Solomon is problematic.
It has simply been assumed that the narratives are contemporary with the events described,
but there are a number of details that would indicate a later date. (Cf. David M. Gunn, The
Story of King David [JSOTSup 6; ShefÔ¬Āeld: JSOT Press, 1978], 32ff.) Indeed, the reigns of
David and Solomon may have been an alluring setting to writers for centuries. Other narrative
texts are also problematic. For example, the account of the execution of those who assassinated
Joash, in 2 Kgs 14:5‚Ä“6, contains a quotation of Deut 24:16. This could be used as a means of
Ô¬Āxing the date of 2 Kgs 14:5‚Ä“6. Unfortunately, I can argue for mutually contradictory dating
with virtually the same reasoning. Does the quotation of Deut 24:16 by 2 Kgs 14:5‚Ä“6 sig-
nify that 2 Kgs 14:5‚Ä“6 must have been composed after the promulgation of Deuteronomy?
Or does this show that Deut 24:16 circulated independently before the public release of
84 HOMICIDE IN THE BIBLICAL WORLD
flqm, ‚Äúrefuge; conÔ¬Ānes.‚ÄĚ48 This term is not found in the other legal sections
of the Pentateuch, and its appearance here reÔ¬‚ects a characteristically Priestly
concern for technical and ritual terminology.
Three cities were to be established on each side of the Jordan River, and
all six were to be appointed simultaneously after the crossing of the Jordan
River, reÔ¬‚ecting P‚Ä™s emphasis that the Land of Canaan was to be conquered in
one fell swoop and that nothing was to be established there until the conquest
was complete. Canaan was not to be distributed to the tribes piecemeal, but
rather was apportioned at a single ofÔ¬Ācial ceremony. The cities of refuge were
part of a scheme of forty-eight cities assigned to the Levites. The question
of whether the Levitical cities ever existed in reality has been the subject of
heated debate.49 Yet little doubt has been cast on the existence of the cities
According to Num 35, six cities of refuge were to be established, and the
number six develops from the Priestly law‚Ä™s endeavor to schematize Israel‚Ä™s
antiquity.50 God is revealed in stages and, therefore, the relationship between
God and human beings is described in a series of four covenants (implicitly
with Adam, Gen 1:28‚Ä“2:4a; Noah, Gen 9:1‚Ä“17; Abraham, Gen 17; Israelites
in the wilderness, Exod 24:1‚Ä“8).51 The Israelites are divided into twelve tribes
ruled by twelve chieftains. The tribes encamp in four groups of three around
four standards on the four sides of the Tabernacle, the wilderness sanctuary.
On the march, two standards precede the Tabernacle and two follow. Within
the Tabernacle, four families serve. The priests and the Levites are given forty-
eight cities, of which six are cities of refuge, three on each side of the Jordan.
All these numbers are based upon the numeral 12, its multiple 48, and its
divisors 2,3,4,6. The precise number of cities of refuge is therefore generated
by the Priestly law‚Ä™s theological numerology.
The Priestly law links the cities assigned to the Levites and the places
of refuge. Many scholars have, therefore, extrapolated that the reason for
assigning a certain city as a place of refuge was due to its status as a sacred city
and/or to the existence of an altar, presuming that the institution of asylum
48 The choice of this term for the cities of refuge reÔ¬‚ects the theology/criminology informing
the signiÔ¬Ācance of the cities of refuge for the accidental killer. See the analysis of the accidental
killer‚Ä™s conÔ¬Ānement in Chapter Four.
49 E.g., W. F. Albright, ‚ÄúThe List of Levitic Cities,‚ÄĚ in Louis Ginzberg Jubilee Volume (New
York: American Academy for Jewish Research, 1945), 1. 49‚Ä“73; Benjamin Mazar, ‚ÄúThe Cities
of the Priests and the Levites,‚ÄĚ Congress Volume 1960 (SVT 7; Leiden: Brill, 1960), 193‚Ä“204;
Menahem Haran, ‚ÄúStudies in the Account of the Levitical Cities,‚ÄĚ JBL 80 (1961), 45‚Ä“54, 156‚Ä“
165 (also Temples and Temple-Service in Ancient Israel [Winona Lake, Indiana: Eisenbrauns,
1995], 122‚Ä“131); J. P. Ross, ‚ÄúThe ‚Ä˜Cities of the Levites‚Ä™ in Joshua XXI and I Chronicles VI,‚ÄĚ
Ph.D. diss., University of Edinburgh, 1973; John R. Spencer, ‚ÄúThe Levitical Cities: A Study of
the Role and Function of the Levites in the History of Israel,‚ÄĚ Ph.D. diss., University of Chicago,
50 Rof¬ī , ‚ÄúThe History of the Cities of Refuge in Biblical Law,‚ÄĚ 225.
51 Wellhausen, Prolegomena, 338‚Ä“342.
DEVELOPMENT OF PLACES OF REFUGE IN THE BIBLE
developed historically from altar asylum to sanctuary in a sacred city. This
begs the question as to why the status of city of refuge was not extended to
all the Levitic cities: If they were sacred or possessed an altar, why did they
not qualify as a city of refuge? Furthermore, a distinction must be drawn
between a city for the Levites to dwell in and the location of the sanctuaries
in which they performed cultic functions; there was no direct link between a
Levitic city and a sacred place. A Levite might live in one place but ofÔ¬Āciate
as a Levite in a different location.52 Abiathar, for example, owns a Ô¬Āeld in
Anathoth, but ofÔ¬Āciates in Jerusalem and Nob (1 Kgs 2:26).
There is another aspect to the cities of the Levites that is critical to the
functioning of a refuge. They were to be distributed evenly throughout the
Land of Israel and Transjordan (Num 35:8). In order to provide equal and
easy access for a slayer, the cities of refuge needed to be distributed evenly
throughout the territory: If each tribe had its own place of refuge, slay-
ers from tribes with smaller territory would have easier access than those
from tribes with greater territory. This is reÔ¬‚ected in the speciÔ¬Āc language
of the command, ‚ÄúYou shall make accessible to yourselves cities of refuge‚ÄĚ
(Num 35:11). It is not the sacredness of a Levitic city that determines its
status as a refuge but rather its geographic distribution.
Understanding that the cities of refuge were selected for reasons of geog-
raphy clariÔ¬Āes a difÔ¬Āculty with associating the right of asylum with a function
of cult sites, with an altar and sacred space. It is assumed that while one who
is in the presence of God is under his protection, those who are not supposed
to be in a sacred space are executed. If the Priestly statute avers that the slay-
ing pollutes (Num 35:33‚Ä“34), would a slayer be permitted to encroach upon
a sacred place? Indeed, the point of the Levitic cities is that they are manned
by personnel not limited by tribal geography and who had access to nonsa-
cred aspects of sacred items. The Priestly law stipulates that the priesthood
was stratiÔ¬Āed into priests, strictly deÔ¬Āned, and Levites, service personnel
limited to assigned tasks of the transport, maintenance, and handling of
cultic items (Num 3‚Ä“4; 8:5‚Ä“22). The Levites were the corps of subordinate
servitors, relegated to nonsacral functions of sacred sites and rites.53 They
assisted the priests (Num 18:2,4) and performed acts that do not pertain to
the altar (Num 16:9). The functions of the Levites are outside cultic sanctity,
according to the Priestly traditions (Ezek 44:11; 46:24).54 Being divorced
from the sacred and being geographically distributed, the Levites are there-
fore the appropriate personnel to oversee the cities of refuge. A killer was to
be kept away from all that was sacred because his offense polluted the land.
52 Menahem Haran, Temples and Temple-Service in Ancient Israel (Winona Lake, Indiana:
Eisembrauns, 1995), 119‚Ä“120.
53 The distinction between priests and Levites may be a distinction only in Numbers. Cf. Levine,
Numbers 1‚Ä“20, 65, 81, 104‚Ä“105.
54 Haran, Temples and Temple-Service in Ancient Israel, 61.
86 HOMICIDE IN THE BIBLICAL WORLD
Just as the theological and social program of Numbers informs the
statutes establishing the cities of refuge in that book, so too does the social
and theological program of Deuteronomy shape its statutes on homicide.
Deuteronomy manifests anxiety over the possibility that !dh lag might com-
mit an accidental homicide because he could kill any slayer with impunity
outside the city of refuge ‚Ä“ ‚ÄúWhoever came with his fellow into the forest
to cut wood: as his hand swings the ax to cut down the tree, the ax-head
falls off the handle and hits the other so that he dies ‚Ä“ that man shall Ô¬‚ee
to one of these cities and live, lest the blood avenger pursuing him in his
hot anger, overtakes him and slays him because the distance is too great, yet
he was not liable to the death penalty because he was not hostile to him in
the past‚ÄĚ (Deut 19:5‚Ä“6). Deuteronomy is concerned with the slaying of the
accidental killer and the effect it would have upon the Israelite people. In
contrast to the Priestly law, where the slaying of the accidental killer does
not incur culpability at all, the Deuteronomic statute evinces the belief that
the killing of a fugitive who has not yet reached a refuge does. In his case,
the fugitive‚Ä™s status as an accidental or intentional killer is not yet clear: He
may potentially be an accidental killer, whose death is unwarranted. In an
ironic transformation, the same term for culpability for the victim, yqn !d,
‚Äúinnocent blood‚ÄĚ (Deut 19:13), is used to refer to culpability for the killer
slain before he reaches the city of refuge (Deut 19:10).
If the blood avenger manages to overtake the fugitive and kill him, the
Israelite people as a whole are responsible, according to Deut 19:10. The
Priestly law, by contrast, avers that the land, not the people, will be polluted
by the presence of the unpunished slayer, not the death of the accidental
slayer. The Priestly law is concerned with the purity and the pollution of
space, the Deuteronomic with that of the Israelite people. The Priestly law
is concerned with the pervading presence of God in the midst of Israel,
while Deuteronomy focuses on the conduct and fate of the Israelite people.55
Indeed, Deuteronomy is completely unconcerned with the immanence of
God: for Deuteronomy, the Temple in Jerusalem is not the dwelling place of
God; it is the place where God causes his name to dwell. The Priestly law
is concerned with the polluting effects of a slaying, whereas D is concerned
with the social aspects of the law.
Like the statutes in Numbers, Deuteronomy‚Ä™s places of refuge are di-
vorced from any link with the sacred or the priesthood, but this is informed by
the Deuteronomic trend toward secularization and the separation of sacred
and secular. This tendency is part of the larger program of Deuteronomy.56
Warfare, for example, is stripped of its sacred ritual in Deuteronomy. There
55 Moshe Weinfeld, Deuteronomy 1‚Ä“11 (AB; New York: Doubleday, 1991), 25.
56 Weinfeld, Deuteronomy and the Deuteronomic School, 233‚Ä“243. The term secularization
may be too strong a term for this aspect of Deuteronomy‚Ä™s program, as Weinfeld him-
self notes (‚ÄúOn ‚Ä˜Demythologization and Secularization‚Ä™ in Deuteronomy,‚ÄĚ IEJ 23 ,
DEVELOPMENT OF PLACES OF REFUGE IN THE BIBLE
is no mention of the sounding of the priestly horns or of the plunder that
is to be dedicated to the sanctuary.57 The use of the ark and the holy ves-
sels is missing in Deuteronomic warfare. Even the function of the priests is
secularized. In P, the priests are to sound their horns so that the warriors
are remembered by YHWH (Num 10:9). In D, by contrast, the priest ad-
dresses the people to inspire their courage (Deut 20:1‚Ä“4). This secularization
is carried through in the Deuteronomic tradition of the cities of refuge. The
Levites, who are considered the Priestly class in Deuteronomy, are not con-
nected with the cities of refuge. The high priest is not a factor in the stay of
the accidental homicide in a city of refuge.
Like Numbers, the Deuteronomic refuges are established on the basis of