Brauer Boundaries Frontiers Medieval

Boundaries and Frontiers in Medieval Muslim Geography
Author(s): Ralph W. Brauer

Source: Transactions of the American Philosophical Society , 1995, New Series, Vol. 85,
No. 6 (1995), pp. 1-73
Published by: American Philosophical Society
Stable URL: https://www.jstor.org/stable/1006658
REFERENCES
Linked references are available on JSTOR for this article:
https://www.jstor.org/stable/1006658?seq=1&cid=pdf-
reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms
American Philosophical Society is collaborating with JSTOR to digitize, preserve and extend
access to Transactions of the American Philosophical Society
This content downloaded from

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
All use subject to https://about.jstor.org/terms
Boundaries and Frontiers in
Medieval Muslim Geography

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
TRANSAC l'IONS
of the
American Philosophical Society

Held at Philadelphia for Promoting Useful Knowledge
VOLUME 85, Part 6, 1995
Boundaries and Frontiers in

Medieval Muslim Geography
RALPH W. BRAUER
Institute for Research on the

Interaction of Science and Culture
Wilmington, NC 28401
THE AMERICAN PHILOSOPHICAL SOCIETY
Independence Square, Philadelphia

1995

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
Copyright ? 1995 by The American Philosophical Society
All rights reserved. Reproduction in any media is restricted.
This work was supported in part by grants from

the Max and Victoria Dreyfus Foundation
and the Foundation for the Carolinas
Library of Congress Catalog

Card Number: 94-78513
International Standard Book Number 0-87169-856-0
US ISSN 0065-9746

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
TABLE OF CONTENTS
SECTION 1
The Geographical Concepts .................................. 1

Boundaries in Arabo-Islamic Cartography ................... 1
Boundaries in the Arabo-Islamic Geographic and Historical Texts 7
Internal Frontiers ....................................... 8
External Frontiers ......................... ......... 11
Terminology ........................................... 11
Hadd versus tugur .................................... 12
Examples from Near East and in al-Andalus ............ 16
Tugur in the Texts of the Arabo-Islamic Geographers ....... 25
Summary of First Section ................................ 26
A ppendix ................................................ 30
SECTION 2
Travelers' Experiences at Internal Boundaries, the Area

Arabo-Islamic Geography, and the Relation of Zone-
to Basic Tenets of Arabo-Islamic Culture .................. 33
Boundaries in the Writings of Travelers in the Islamic
The Concept of Area in Muslim Geographic Thought
Land Tax Data as a Basis for Quantifying the Area of a
Political Entity ....................................... 37
Boundary Characteristics as a Consequence of Embedded
Attitudes of the Culture ................... ............. 40
SECTION 3
Genesis of Boundary Zones Involving non-Arab Muslim

The Case of Asia Minor ................................ 45
The Historical Setting ..................................... 47
The Boundary Zones Associated with the Conquest of Asia Min
Frontier Population and Boundary Zones between Seljuk and
Byzantine States-Ghazi, Turcoman and Akritai ...........
A ppendix ................................................ 60
SECTION 4
Summary and Conclusions .................................. 65

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
SECTION 1
The Geographical Concepts
t has been pointed out that the concept of the state is linked
ception of its territoriality.1 Therefore, the character of th
aries that confine a given state reflect important aspects of
taken by the inhabitants of the nature of their polity. While t
al-Idrisi's "Book of Roger"-the "Kitab nuzha 'l-mustaq fi '
afaq" his great 'Opus Geographicum' I failed to encounter a
ence to boundaries between various political or ethnographi
either the text of this work or the maps accompanying it. In vi
importance of boundary concepts, this observation seemed wo
further inquiry to determine whether it might represent a m
syncrasy of this one author, or prove to be a reflection of a mo
characteristic of the geographic concepts current in the Arab
civilization.
This monograph presents the relevant data from Arabo-Islamic
scholars, relating them to the experiences of observers other than the
scholar-geographers, testing them by reference to a second Islamicized
population, and finally defining some of the questions raised by the data.
Boundaries in Arabo-Islamic Cartography
The earliest known map by Muslim scholars, al-Ma'mun's world map

of the beginning of the ninth century, unfortunately is lost. While its gen-
eral outlines can be reconstructed with some confidence on the basis of
al-Kwarizmi's geography treatise,2 only a few small bits of actual cartog-
raphy attributed to the Ma'munian map have been preserved. One
example is a map of Egypt depicting the course of the Nile (Fig. 1.1,
from a Ms. of al-Kwarizmi's Kitab Siratu 'l-Ard3). This does not show
any boundary between the Islamic lands of Muslim Upper and Lower
Egypt, nor between Muslim Upper Egypt and the non-Islamic Nuba
kingdom not far south of Asuan-a boundary well recognized not long
after the conquest as shown by the so-called 'bakt' a curious document,
seemingly the only "peace treaty" concluded and honored for centuries
1 State and territoriality concept-Lambton.

2 al-Kwarizmi, Muhammad Ja'far Ibn Musa, Das Kitab Siiratu 'l-Ard des Abu Jafar ibn
Musa al-Kwarizmi, ed. H. von Mzik (Leipzig, 1926).
3 Map from Ms. of al-Kwarizmi's Kitab Suratu 'I-Ard, Ms. COD 4247, Universitats Biblio-
thek Strassburg as presented in Yusuf Kamal's Monumenta Cartographica, 1930, vol III,
p. 524.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
'YCL=J
P
ZJflCp
fflyiJpJ
u4ct?????
?L-Jb-iFd
,5J?3
o
Ct
r?i
FIGURE 1.1. Map of the course of the Nile of Egypt

is dated before 218 h = 833AD). The straight lines ru
the page, belad nuba is written in the triangular space
boundaries are shown-the line to the right of the Nile

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
THE GEOGRAPHICAL CONCEPTS 3
between Muslim and Infidel - in this case - Christian states.4 These find-
ings suggest-but are far from establishing conclusively-that, unlike
Ptolemy's Cosmographia,5 al-Kwarizmi's treatise did not include informa-
tion on boundaries, and, by further inference, that this feature may also
have been absent from the original Ma'munian map.
With later Arabo-Islamic maps one is on considerably firmer ground.6
One can distinguish two major schools of Muslim cartography: the earlier
Balkhi school that included the cartographers contributing to the various
versions of the so-called Atlas Islamicus,7 and the school of twelfth- and
thirteenth-century geographers including al-Idrisi and those following
his lead. A major representative of the Balkhi school is Ibn Hauqal whose
maps form an integral part of his treatise on geography.8 Out of a total
of twenty-three maps included in his treatise, twelve showed no bound-
aries. The remaining eleven-including among others all of Ibn Hauqal's
maps of the Middle East-did show boundaries labeled hadd, but what
is shown under that designation is clearly devoid of geographical signifi-
cance. In each case the detailed map sketches are encased in caligraphic
and largely rectilinear frames designated as "h ................. add al-firs or
"h.............add al-kuzistan." Evidently, these (Fig. 1.2) cannot well bear
any relation to real borderlines separating one of these states from
another on the ground. This conclusion is made even more convincing
by six of these maps showing in place of parts of the boundaries framing
a given state two parallel caligraphic lines separated by a small distance,
representing (and labeled as) the boundaries of two adjacent countries,
for instance al-fars and al-kuzistan (numbers 1 vs. 2 and 3, and 4a and b
vs. 5, 6, and 7 in Figures 1.2a and b illustrate this for a rectangular and
a circular 'frame'). Clearly, such double lines would have to coincide were
they intended to represent a boundary line between two states.
Small scale circular world maps from the hand of several of the
authors of this school have survived. These push the point further: they
show whole regions subdivided into more or less rectangular boxes, each
inscribed with the name of a country or a province (Fig. 1.3). The arrange-
ment of these boxes relative to each other approaches randomness-
assignment of a name to a given box bears little relation to the actual
4 The document circumvents the Qur'anic interdict against permanent peace with the
infidel by posing as a commercial treaty.
5 Ptolemaeus, Claudius, Cosmographia (Rome 1478 ed.), Amsterdam, 1966. This work
was transmitted to the Arabo-Islamic scholars at an early date; under khalif al-Ma'mun it
was translated into Arabic by a committee of scholars that seems to have included al-
KwarizmI. There is every indication that this work was germinal to the development of geo-
graphic science in the Islamic world. Boundaries are frequently referred to in the text accom-
panying the tables of geographic coordinates of the Cosmographia, typically in the form:
A. comes to an end ('terminat') at......... naming either a natural feature, like a river or a
mountain range, or a line connecting well-defined places.
6 Cf. e.g., maps in Kamal, Yusuf, Monumenta Cartographica Africae et Aegypti, vols. III
and IV (Leiden, 1932-1935).
7 Miller, K., Mappae Arabicae (1926), ed. H. Glaube (Wiesbaden, 1986), 7-20.
8 Ibn hauqal, abu '1-kasim, an-nasibi, Kitab suratu 'Iard, ed. J.H. Kramers (Leiden,
1967).

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
4 MEDIEVAL MUSLIM GEOGRAPHY
a)
.S?
fi
,y <cJ
i1
'??
?Y ;
'"3
t* "
?s
,a _
!,d
' a
4 ;s
MQ
hi_
^E
=1 ,5,,
_ CY
o x?
it-,
S Ti
_ si
lt:
*,i .v
..i
E~
]
1,
_ ,
I-
.kX
i<s
a
1.
V Id
ti2
."
.^
J> V1|
33^
I5
;:::e
iqs.
Mi
-9 ?
=tf
Cv) tiN

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
identity, and a peripheral zone where the force of sovereignty

tity of that state grew progressively feebler as one moved away
capital, to be replaced gradually, as one proceeded on a line be
capitals of any two neighboring states, by the sovereignty an
of the adjoining state.
This concept is born out by the cartography of al-Idrisi and
sors. I have had the opportunity to examine the originals of th
scripts of al-Idrisi's great work9-the oldest one, the twe
thirteenth-century Paris Ms. Bibl. Nat. FA. #2221, and the lat
Library's Mss. Pococke #345, and Greaves #42. In each case the
world and the text describing it are conceived of as divided i
sections, corresponding to seven climates, aqilim, each furthe
vided west to east into 10 sections, ajza';10 each of the resulti
tions is provided with its own map.11 I have been unable to t
indication of boundaries in any of these maps, including s
those representing regions where Ibn Hauqal's maps deplo
linear hadd, i.e., in the Middle East (sections III 5,6,7; IV 4,5,6;
and western Central Asia (IV and V 7,8) for the two first nam
and sections III 5,6,7 in the case of Greaves #42, as well as in all th
in the sections representing Egypt as well as the North Afric
between Egypt and the Atlantic (III 1-4). In place of bound
encounters expanded text such as "country of 'Iraq,/ or "count
Nuba," written, e.g.,
which may be transcribed as

'bala d nuba'
placed across a region corresponding to

named, often surrounded by a more or
neighboring units (Fig. 1.4). Thus, whil
graphic indications of boundaries or bo
the core regions of the political units c
dence has come to light so far concern
such as it is, it confirms the impression
texts to the effect that he followed al-I
omission of boundaries from his cartog
In sum, examination of maps by twen
raphers13 working between 820 and 13
completely omitted from much of the
9 al-Idrisi, al-aarif, Kitab nuzha 'I mustaq ft-'ktir

10 The full text and accompanying maps are p
Pococke 345, but only the first three climates, 3
11 These seventy maps are more familiar in the
were assembled -mistakenly, I am convinced (cf. m
parative Civilizations Review 26 [1992 : 73-109), b
Stuttgart 1926).
12 Ibn sa'id al-Magribi, Kitab bast al- ard f't -t il
13 A list of cartographers or map sources cons

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
* - - L ; A . ?
FIGURE 1.4. A sample of al-Idrisi's cartography, the sixth part of the secon
resenting the region around the northern end of the Persian Gulf (indicate
area). As usual, the South is at the top. The delta of the Euphrates/Tigris
by the network of channels in the lower right quadrant. Note the comp
boundaries in spite of the fact that the map includes parts of Syria, Iraq, K
Fars. The characteristic expanded labeling of the inner lands or the central r
ognized country is here illustrated by the labels balad ku..........zist
(bottom, center), and balad f..........ars o. .. (just above the horizontal
middle, near the left margin).
are represented only by simple geometric lines having symbolic rather

than geographic significance in the rest. The data suggest that each
country was conceived of as being divided into a core including its center
of power and a periphery separating that core from the nearest adja-
cent country.
Boundaries in the Arabo-Islamic Geographic and Historical Texts
Boundaries by their very nature are not neutral but are strongly con-
ditioned by the political realities of the moment. It should be understood
that the present survey of Arabo-Islamic documents bears essentially on
the period of 440 years from about 750 to 1190 AD. This was relatively

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
free from external threat to the Islamic heartlands, and-de

nial minor fluctuations as opponents raided and counterr
the frontiers, the emergence of internal divisions, and the
moil in al-Andalus-can be considered as having allowed a de
of frontier concepts corresponding to the inclinations of t
actors in the politics of the realm.14 It is permissible, I think
this period as that of Arabo-Islamic geography, a term that r
dominant players without attempting to deny that many of
whose works made up the body of science of the period we
Syrians, Greeks, or perhaps Berbers or Indians.
In discussing the frontiers of the Islamic Empire and i
states, one must distinguish two kinds of boundaries with
ent ideological and geographic characteristics-internal a
boundaries.
Internal Frontiers
Internal frontiers involved two adjoining Muslim states that may or

may not have been at peace with one another. However, during its early
years, Dar al-Islam was governed by the Prophet's preaching and the
Quranic command that there all true believers should be brothers inhab-
iting one single 'umma under a single imam, and hence there could not
exist any internal frontiers. This situation prevailed during the early
stages of the conquest under the four first, "rightly guided" khalifs, and,
although with some reservations, throughout the period of the reigns of
the Umayyad khalifs of Damascus: Dar al-Islam was a single political
unit under one ruler, at first the Prophet ruling from Medina, than the
Prophet and subsequently the first khalifs ruling from Mecca, later on
the Umayyad khalifs ruling from Damascus. This situation changed
when the 'Abbasids of 'Iraq displaced the Umayyads in the second half
of the eighth century. There followed almost immediately the beginning
of the breakup of the Empire with the secession of al-Andalus under the
rule of the Umayyads of Cordoba.
From the death of Muhammad, the khalifs responded to the increasing
size of their dominions by delegating authority to umara' (pl. of amir)
who were in charge of the military campaigns as well as of religious
affairs in the various provinces. Presently, when the first ebullience of the
conquest had subsided somewhat and the matter of gathering revenues
assumed greater prominence, 'umala' responsible for fiscal matters were
associated with these religio-military officials. Given the widening
spaces over which the Empire held sway and the limited means of com-
munication available at the time, it should come as no surprise that these
officers had to operate much of the time independently of the central gov-
ernment, and that with the passage of time they should assume consid-
erable though variable degrees of autonomy. Under the Umayyads this
tendency had been controlled in part by close personal relations between
14 Vatikiotis, P.J., Islam and the State (London, 1987), 21.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
the ruler and his delegates, and in part by assuring that the terms
of many of the provincial governors were kept short.15
Under the 'Abbasids these constraints presently showed th
insufficient, and one by one provincial governorships became
ingly autonomous and assumed hereditary character (cf, F
Thus, the secession of al-Andalus was followed in less than 50
North Africa by that of the Idrisids of Marocco, and shortly
by that of the Aghlabids of Ifriqiya and their successors the Fa
a little later by those of the Tulunids and the Ikhshidids of E
northern Syria the Hamdanids ruled throughout most of the
tury, while in the Iranian lands the Tahirids were effectively i
by 820 AD, to be succeeded by the 'Alids and the Saffarids w
than half a century. In the Khurasan and Transoxania the Sam
before the end of the ninth century and spread their rule in
regions. They in turn were succeeded by the Ghaznavids who
of power eventually shifted southeast into Afghanistan
Punjab.17 All of these arose from provincial governors under
zerainty, and what had been formerly provinces of the khal
ently became autonomous or de facto autonomous states with
rulers, and with their own armies and revenues from which th
minimal or no part to the fisc at Baghdad. While some of the
acknowledged the suzerainty of the Khalif of Baghdad at least
others chose to ignore him even as a matter of form, e.g., by
reading his name into the Kutba, the Friday prayer, or by d
name from their coins. By the eleventh century Seljuks, Selju
Fatimids, Khwarezmshahs, Almoravids, and Ghaznavids between them
had assumed the rule of virtually all of the Islamic territories, leaving the
kalifate with little more than the city of Baghdad-and that only under
the watchful tolerance of Buwaihid or Seljuk masters.
The pattern emerging is graphically represented in Figure 1.5b: where
Muhammad had envisioned a single unified realm dominated by a single
theocratic ruler, the kalifa, within 150 years after his death we find two
units. Thereafter the number of successor states continued to increase
rapidly and progressively over the next 300 years so that by the time t
Mongols extinguished the remnants of the kalifate in the mid thirteen
century the realm of Islam had been split up into fifteen separate Muslim
ruled states. The history of these successor states includes numero
instances of warfare between some of these entities. Thus, the breaku
of the Islamic Empire was associated with the formation of numerous su
cessor states that in turn entailed the establishment between them of
numerous "internal boundaries."
15 Short terms of amirs under Umayyads.

16 Lane-Poole, S., The Mohammedan Dynasties (New Delhi, 1986) after pp. XVI and XX
17 Cf. e.g., Bishai, W.B., Islamic History of the Middle East (Boston, 1968). Part IX
Regional Dynasties during the Late 'Abbasid Period.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
e^o
CL)
(C
I
U3
0
d
LL0
"4
a,-
o52
u '
C~
C)
L a
PB.i
X
D
.fU
g x
O
0 a ,
ti,
w t
u c
*Q
x
r
0
2
w
r-
D;

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
12
. 8
0 200 400 600

year after hiMra
FIGURE 1.5b. Number of coexistin

function of time over the sa
External Frontiers
External frontiers occurred at the periphery of the empire wherever a

Muslim polity adjoined an 'Unbeliever' one. They were recognized from
early on as frontiers facing dar al-harb, the land of war of the Quranically
mandated permanent state of Holy War. In the period here under con-
sideration this had come to be interpreted as implying that they were in-
evitably and permanently frontiers of war even when at a given moment
armies were not marching upon one another.
Terminology
Before proceeding further with the discussion of the actual configura-
tion of these two types of boundaries, it is necessary to pause for a
moment to consider the terminology used in the literature to refer to the
geographic concepts representing the transition from one polity to
another.
Of the Arabic terms denoting some kind of boundary (cf. footnote)18
18 The Arabo-Islamic authors made use of a number of Arabic terms to designate a con-
cept sometimes properly translatable as 'boundary':
1. akir = ^ T ...... the extreme part of anything, including provinces or countries or
many kinds of geographically definable pieces of land or sea;
2. takuim = r .. ..... from takama = to fix the limits of anything; the noun is translated
as boundary, limit, border; mutakhim = (LLt- is used to mean adjacent or neighboring.
3. haiia = L.tL ........ margin (of book), border, seam, edge.
4. farj = ".J ...... related to farjah meaning a mountain or other pass, and occasionally
translatable as border (as e.g., between Sind and Multan, i.e., muslim and kafr).

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
three (akir, takum, has'a) imply a view of the relationship stri

within-the end or limit of whatever is inside. The term farj im
way flow of traffic like that across a mountain pass.
In practice, only three of these ten terms were in frequent u
geographers to denote political boundaries: Hadd can be taken
a hostile relation-the edge of the sword facing the enemy. On
hand, as employed by the geographers it seems to imply mos
more than the end of something. More unambiguously agg
tugir (original meaning: the front fangs of a dog!); its Mamlu
ment is nib. Finally,'awasim carries a defensive implication, th
off of an enemy from without, and designates a special feature of
boundary systems, to be discussed later in this essay. One ma
finally the Turkish term "u4" encountered below in conne
events in Asia Minor during the Turkish conquest to designat
in compound words like border fighter.
Hadd versus tugfir. The term Hadd is used by the geographers

it is intended that the reader should understand that some ge
entity came to an end. Thus we find it applied not only to cou
also to cities, to the domain of Islam, to land as opposed to th
the end of a mountain range, or to the great desert and othe
In its political sense Hadd was used frequently by Ibn Hauqa
1.2), by al-Istakhri,19 but also by al-Idrisi20 in the descriptio
confines of specific regions within the realm of the Islam
Egypt, some of the countries of the Middle East, the Sind-bu
anywhere else. One may encounter further specification, by d
of the general configuration of the implied boundary both in term
countries or peoples on the other side of various portion
5. hadd = ....... primary meaning: cutting edge of knife of sword; t

boundary, border, limit, terminus (also, in other contexts, punishment f
bidden by the Qur'an).
6. tagr = . ..... primary meaning: front teeth (as shown by an angry do
Lane emphasizes what appears to be the opposite concept, the hind
mount or a beast of burden, the place whence one drives it forwards); port
ture), harbor, as well as (military) frontier town, or frontier district facin
the plural, tugur = ^ designates a frontier zone or march, and so
boundary.
7 nib, pl. antab= ... lit. canine tooth, fang; used in Mamluk times for frontier
defenses in place of tuguir, a term that by that time had gone out of general usage.
8. 'asim = ........ from 'asama-to hold back, hinder. The noun is used to designate
towns less exposed than the tugur but yet lying near enough enemy territory to be
of military significance. The plural, 'awasim = rl- designates a zone defined by
the presence of a number of 'awasim and lying behind the zone of tugur.
Finally one may mention the Persian terms marz = -. ........ a limit, border, boundary
of a country or a field and not used in Arabic texts, but probably the origin of the term
'marca' in Romance languages or march in English and Mark in German and related
languages.
19 al-Istakri, abu ishaq al-farisi, Kitab masalik al-mamalik (de Goeje, M.J. ed.), 3rd ed.
(Leiden, 1967).
20 al-Idrisi, al-sarif, Kitab nuzha 'I mustaq fi-ktiraq al-afaq (Beirut, 1989).

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
boundary, and, albeit merely in very general terms-by referen

direction in which boundary portions lie with respect to t
country that they delimit: "On the East, the hadd of Egypt fa
of Qulzum, then it bends to face the Kingdom of the Baja, th
between the Land of the Nuba and Egypt. .. ." Mountain range
sionally a town, and rivers may be named as defining such a
although more commonly rivers and the valleys they create ar
by a single political unit. As in the case of the Sea of Qulzum
Sea), sea shores provide sharply defined boundaries, and indee
implied by border formalities in ports to be discussed in a su
section. In some cases one finds the transition from one c
another denoted merely by reference to a town: "The first to
west of Ifriqiya is, . . .' presumably implying progress on a roa
kind, usually not further defined. All parts not defined in th
terms are left vague.21 The concept of hudud as described in th
is well represented by the decorative lines or arcs encountere
Hauqal's maps (Fig. 1.2): generalized designations, indicating
political units abut on each other in a given region, but without an
tion of providing a precise definition of an actual line of sep
frontier zone enveloping a central core in the same sense as t
raphers' symbols, rather than a boundary line of demarcation
a realm within which the power of the central government i
formly. This concept also corresponds to the pattern of label
throughout by al-Idrisi as described above (cf. Fig. 1.4). The o
tions to this perception are in relation to sea coasts and seapo
the few places where the boundary follows the course of a rive
natural feature defines the boundary sharply.
All told, therefore, both texts and cartographic representa
concur in implying a concept of boundaries within the broad c
the Islamic Empire that is not that of a sharp transition from o
entity to the next, but rather a gradual interpenetration of th
communities, reflecting a political reality well described by H
"Before the modern age frontiers were not clearly or precisely
. . . one should rather think of the power of a dynasty as radia
a number of urban centers with a force that tended to grow w
distance. . . 22
Elsewhere I have suggested that Muslim geography of the Middle
Ages is a linear geography, conceived in terms of a network of lines of
communication between cities.23 One might expect that in such a repre-
sentation hudid might intersect these lines of communication to define
points where one passed from one sovereignty to the next, marked per-
21 Cf. e.g., al-'umari, ibn fadl allah, Masalik al-absar fi al-mamalik al-amsar, Gaudefroy-
Demombynes trans. (Paris, 1927), 99.
22 Hourani, A., A History of the Arab Peoples (Cambridge, 1991), 145.
23 Brauer, R.W., "Geography in the Muslim World," Compar. Civiliz. Rev. 27 (1992):
73-110.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
haps by a customs station; as will be shown in Section 2 this is

case, and travelers' reports corroborate the sense of absence of
boundaries we gleaned from the cartographers.
The second term, tagr, while at times translated as boundary,
represents a more complex reality. The original sense mention
has persisted in modern Arabic where it still designates a p
harbor. Alexandria was often referred to in the Geniza documents as
'the tagr'24 the ports of the Levant further to the north also were known
collectively as 'the tugir (al-bahra)'-in both cases reflecting the aversion
of the early Muslim rulers to what they perceived as a sea dominated by
hostile forces.25
This idea presently was applied not only to ports on the seashore but
extended to places on land. In both the Near East and al-Andalus the
earliest usage of the term tagr referred to the region just in front of the
armies facing the Christian enemy. Presently, the singular came to be
used primarily to designate specific localities in a more stable frontier
zone, fortified places that served as residence and staging points for war-
riors engaged in carrying the jihad-the Holy War-to adjacent enemy
lands beyond the confines of Dar al-Islim. Such places might shelter ribait,
the quasi-monastic fortified establishments housing those fully dedi-
cated to fulfilling this religious obligation (as e.g., in Ibn Hauqal's Kitab
suratu 'l'ard where, among a number of places, he mentions that the
people of Damascus established a ribat in the neighboring tagr of
Antartus). The regions in which such fortified places were located nec-
essarily lay in the frontier zone between Muslim and enemy (unbeliever)
countries, and in the literature of the times came to be referred to as the
tugur. Indeed, the term was used exclusively for frontiers of this kind.
Table 1.1 shows that, while boundaries were recognized between Arabo-
Islamic, Iranian and Turkish polities and lands dominated by many
ethnic groups, tugur were recorded when and only when the adjoining
territory was dominated by non-Islamized peoples.26
24 Lapidus, I.M, ed. Middle Eastern Cities (Berkeley CA, 1969), 82.
25 Hourani, G.E, Arab Seafaring (New York, 1975), 53-54.
26 One may test the importance of the ethnic as opposed to the religious factor in rela-
tion to boundary formation in the Islamic Empire by asking whether ethnic differences
influenced the formation of the militarized kind of boundary zone designated as tugur in
the course of such armed conflict. Table 1.1 lists contacts between the several comprehen-
sive categories of ethnic encountered in the course of the history of the Islamic Empire over
the first eight centuries following the revelation to the Prophet, and relates these to the pres-
ence between such pairs of states of recognizable political borders as well as of tugur, the
militarized kind of border zone.
The categories established in this fashion allow one to distinguish three non-vacant
classes: 'neither type of boundary observed'; 'political boundaries but no thughfr'; and
'political boundaries with thughfr' (a fourth category, 'thughfr without political boundary'
is physically impossible). The twelve groups of pairs of neighbors identified in this fashion
are distributed between these three classes so that two fall into the first one, four into the
second, and six into the last. When the several polities are further separated into those pro-
fessing Islam, either from its inception or as recent converts -and those that do not do so -
(Christian and Hindu) (Table 4.4) it is found that in the six cases involving two adjoining
Muslim states there were no tugur, even where the states had been engaged in armed

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
TABLE 1.1. Ethnic Factors in Boundary Structure in the Medieval M

Conquerors Residents Tugur Political frontier
Arabs Yemenites 0 0
if Iraqis 0 0
" Iranians 0 +
" c. A. Turks + +
f" Hindus + +
" Rum + +
" Armenians + +
" Egyptians 0 +
" Berbers 0 +
" Franks and Visigoths + +

Turks Rum + +
" Armenians + +
" Iranians 0 +
Muslims vs. No political borders Borders, no Tugur Tugur

Early Converts 2 4 0
Christians 0 0 4
Polytheists 0 0 2
Tugiir as thus int

smaller fortresses
regions of Muslim
for raids into kafir
in the north of Sy
and Anatolia, form
"ma wara' 'n-nahr,
eastern margin of
Andalus tugir al-ad
the central, and th
adjoining portions
struggle, while in all si

were described that we
between northern Syri
tenceof these structure
were the conquerors at
in Anatolia, tugur or a
compelled to retreat. In
been associated with an
These results suggest t
tugur when, and only
Muslim state and hence
enon thus seems unrela
lated, too, to the natur
evidence available does
there was a religious dif
Table 4.4 is a consequenc
necessarily was protract
tive of the paramount
of boundaries involvin

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
zone of the tugir and the interior Muslim realm an additi

zone known as 'awasim was interposed in some places, n
Cilicio-Armenian frontier (Fig. 1.6) and in parts of the Anda
The perception of the tuguir as denoting a zone, rather than
series of points suggested by the writings of the geographe
part the original usage referring to the frontier between
Christian as 'the tagr'; in part, too, it most probably reflected N
political thinking that was strongly city-oriented and tend
combination of an urbanized place and the surrounding
pertaining to a single common unit-a district-several of w
make up the next higher political or administrative entity,
Andalus as the 'kura'.27 The border districts were distingu
kuras in the central portions of Syria/Mesopotamia or of
being given special administrative, and often taxation regi
from the core parts of the Islamic Empire. In particular, th
the control of a military commander, qa'id, and their link
tended to be much more tenuous than that for regular kur
solidated parts of the state. Indeed, the holders of these ap
fulfilled many of the functions of a regular ruler and at tim
in defiance of the wishes of the central government. It is
prising that there was a tendency for these posts to becom
leading to the foundation of dynasties. In the nature of thi
turn were apt to become foci for rebellion.28
The military importance of the tuguir in the Middle East
al-Andalus decreased rapidly after the middle of the eleve
but even after it had all but vanished these places retained
nificant role as centers for the exchange of merchandize bet
and Christian lands.
Examples from Near East and in al-Andalus. Tugur were described in a

regions where Muslim territory abutted upon non-Muslim lands: In al-
Andalus, in southern Egypt, in Syria, in Mesopotamia, in the Khurasan
and in the Makran. The significance of the usages of the terms describe
in the preceding paragraphs can be illustrated by the history of the con
27 Cf. the extensive discussion of the administrative and geographical terminology re

ative to these entities in: Bosch Vila, J., "Consideraciones sobre al-tar en el Andalus y la
division politico-administrativa de la Espania musulmana," in: Etudes d'Orientalisme dedie
a la memoire de Levi-Provencal, vol. 1. (Paris, 1962), 20-33. Note specifically remission of the
jizia for the border population of the Jarajima (EI 2, p. 456) and Glick, T.F, Islamic and Chris
tian Spain in the Early Middle Ages (Princeton Univ. Press, 1979), 104 and 105 reorganization
of frontier zones under the Umayyads.
28 Examples are the Bani Qasi in the Upper March and their near independence of Cor
doba (Glick, cf. above), 103, and Levi-Provencal, vol. IV-1 (cf. above), 101-104, who con
cludes . . . always unstable and the authority of Cordoba was never fully accepted for an
length of time" and mentions, besides the Banu Qasi, Bahlul Marzuq, and Ibn Chilliqi, an
a revolt in Merida that was not fully suppressed until 930 AD; Chejne, A.J. in : Musli
Spain: its History and Culture (U. of Minnesota Press, 1974), 50 ff. lists revolts of Dhu 'l-Nun
of Toledo in 1016-1085, al-Aflas of Badajoz, Banu Hud of Zaragoza in 1040-1042 AD.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
FIGUR
( pl a3M
shows
the Kh

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
cept in the two regions where it would seem to have played

role-in the near East and in al-Andalus.
In the Near East the dynamics of the conquest established a front
facing the Byzantine Empire well before the end of the seventh centu
and, despite wavering of the line with successive raids and counterra
this frontier zone shifted remarkably little for several centuries, until n
the end of the Buyid domination of the Khalifate in the thirteenth
tury and the beginning collapse of Byzantine rule in Anatolia before
Turkish onslaught.29 A defensive and staging zone was recognize
early as the reign of 'Umar ibn 'Abd al-Kattab and referred to as ta
While widespread flight of the local population resulted in exten
depopulation of the zone, deliberately enhanced by forced depopula
under such Byzantine emperors as Heraclius,30 some of the towns o
earlier date survived and a number of fortified new ones were established
to serve as centers housing warriors as well as to as to serve-especially
at a later date-as centers for cross-border exchange of merchandize. It
is these towns that came to be referred to by the geographers as the tugur
and that gave their collective name to the zone. In the spaces between
these cities numerous lesser fortified sites seem to have existed,31
strengthening the character of the zone to provide defense in some depth
(cf. Table 1.2 below). The near eastern tugir were subdivided at the latest
under the Khalif Harun ar-Rashid into Syrian and the Jazirian
tugur (tugfr ash-shamiya and tugir al-jazirna) with capitals in Tarsus and
in Malatiya respectively. The distribution of these tugur in the works of
the geographers suggests a defensive line of cities as illustrated in Figure
1.7 and 1.10. The designation seems to have persisted for some time after
shifts in the military front had abolished the military significance of these
tugur. As mentioned above, during the early centuries of Islamic domi-
nation this zone was backed by a second zone of primarily defensive sig-
nificance, the 'awasim. This is illustrated for the case of the North Syrian
frontier with the capital of this particular 'asima in Antiochia (cf. Fig. 1-6
for a modern reconstruction of the map of the region; sugur and avasim
are the Russian transcripions of tugur and 'awasim).32 With the passage
of time the distinction seems to have become blurred (thus already both
Ibn Kordadba and Qudama speak of both under the single term awasim
when reporting the revenues of 400,000 or 360,000 dinar from the region),
and by the twelfth century the entire frontier zone came to be referred
to as the tugur. The difference between the inner portions of the Islamic
Empire and this insecure frontier zone is well illustrated in the descrip-
tions of Ibn Jubair, the returning pilgrim, who calls attention to the fact
that as one came close to the northern limits of the Jebel Lubnan (on
29 Honigmann, E., Die Ostgrenze des Byzantinischen Reiches von 363 bis 1071 (Bruxelles,
1935).
30 Canard, M., in EI-1, vol. 8, article: 'awasim p. 761.
31 Honigmann, E., EI-1, article: thughir, vol. 8, p. 739, and M. Canard article: 'awiasim
in: EI-2, vol. 1, p. 761; and Table 2 below.
32 Ter-Jervondian, Armenia i Arabskii Khalifat (Erevan, 1977), 151-154.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
TABLE 1.2. Townsb Designated as _tugir in Writings of Medieval
Arabo-Islamic Geographers
no. of author 1 2 3 4 6 7 8 9 10 11 12 13 14 15 16 17 18
8 8 9 9 9 9 9 9 1 1 1 1 1 1 1 1 1 1
5 5 0 0 0 8 8 9 0 1 1 2 2 2 2 3 3 3
0 1 2 3 8 2 8 7 0 5 5 2 2 5 8 0 6 7
0 0 4 5 8 0 6 0 8 5 nosl names of
towns authors:
Asuan A 1.Ibn Khordadhb
Tarsus \\ \\ \ \ \\ \\\ \\ \ 2.1bn ya'qubi
Antaki\a 3. Ibn al faqih
aClskanderia \\\\ \ 4. lbn rostah
Malatia VA VA V 5. aistakhri
al-Masisa \\\\ \\\\ \\\\ 6. hudud a-'alam
Mar'ssh \\ \\ \\ 7. Ibn hauqal
Antartus 8. al-mqaddasl
Adhan \\\ \\\\ 9. al-andalusi
Washiird
vvz4qi u %%%I'l" 11 10 az-tuhrl
al-Harunla \ \ 11. aHdrlsl

Farbar \\\\ 12. ad-dlma shqi
Samat 13. y /aqut
Sariq \\\\ 14. a d-qazuinl
Isfijab A A 15. It bn sa'id

FU \\ 16. a Ebu 1-feda
Bira \\\\ \\\\ 17. a -himyari
Bughras 18. II bn khaldun

Ra's al-Kanals\\\
Dailam \ \\\
al-Karisa
Qazuin \\\ \\\

Tola\\\
Wadi 1-ha"la \ \
Tarun \
Zamorm
Santarin
Sarqasta \\\\ \
Larida
Barbuna
Tolta \\\ \\\\

'aln zurba\\\
Tartusha\\\
Saragossa \\\
Med. as-Salam\\\
Marida
Nafza
T_. al .~a %a\\ \\\

T. al-awslt \\
T. al-adni \\\\ \\\\
Tat-trak \\\\
T al-jazlra \\\ \ \\\ \\

J. ai-bahr \\\\
. an-nOba \\\\
I. aw-sam \ \ \\\ \\\
b in Syria, al-Jazira, Khurasan, al-Andalus, Egypt

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
FIGURE 1.7. Modern map of the same general area as shown in Figure 1.6 with towns identified in the writings of the Arabo-Islamic
marked by black circles.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
the other side of which the first Crusader towns were then
kans, that is, the travelers' inns, became veritable fortresse
iron gates.33
The population of the countryside of these boundary zones, and espe-
cially of the fortified towns therein, was mixed. It included some of the
original population, mainly Syrians and Armenians, Arabs-mainly, it
seems, Qaisites-from the invading troops, as well as Zott, a wandering
people, perhaps ancestors of the later gypsies and originating in
India,34 irregular border warriors, and the jarajima, somewhat lukewarm
Christians who served the Arabs as spies and scouts, especially after the
capture of Antioch, in return for cash emoluments as well as remission
of the jizia, the head tax normally imposed upon non-Muslims residing
in Muslim territory.35
In the case of al-Andalus, a zone of contact between the invading Mus-
lims and the defending Christians was established early, and well before
the turn of the seventh century was referred to at times as the tagr of al-
Andalus, presumably reflecting the original metaphoric concept of the
threatening teeth presented by an angry dog to his opponent.36 In short
order this came to be a sort of no-man's land, avoided by Christian inhab-
itants to whom this region, abutting on the Marca Hispanica, the
Spanish Marches, had become 'terrae desertae' The zone was presently
provided with a network of fortresses, including sites like Saragossa, left
over from an earlier time, as well as newly established strong points like
medtnat as-salim, i.e., Medinaceli, and came to be referred to by the collec-
tive "the tugir" (Fig. 1.8). In the course of the next decades the Muslim-
controlled portion of the Iberian Peninsula became more or less stabi-
lized, and for some time thereafter Muslim authors conceived of the
Peninsula as divided by a largely imaginary range of mountains, the sar-
rat or Sierra (or could it have been 'the Badlands' from the Arabic root
ash-sarrun?) running west-east from the Mediterranean two-thirds of the
way to the Atlantic at about the latitude of Toledo;37 even in the twelfth
century, in the sectional maps of al-Idrisi's Book of Roger, significant
parts of this imaginary range are preserved to separate 'ard castilia' from
'ard al-andalus' (Fig. 1.9).
Shortly after the arrival and accession of the Umayyads to the western
Amirate, around 753 AD, this border zone was partitioned into Upper
and Middle tugur. In the course of the following decades it was further
divided into subzones, the Upper, Middle and Lower (al-adna, al-awsat,
al-a'la) tugur (Fig. 1.8a), a partition between Muslim and Christian lands
33 Ibn Jubair (1144-1218 AD)- Rihla (Dar Beyrut, 1974), 229.

34 Ferrand, G., article "Zott" in EI-1, vol. VIII, p. 1235.
35 Canard, M., in: EI-II, vol. 3, article Djaradjima, 436-438.
36 Bosch Vila, J., "Consideraciones sobre el tagr en el Andalus y la division politico
administrativa de la Espafia Musulmana" in: Etudes d'Orientalisme dediees i la memoire de Levi-
Provenqal, vol. 1 (Paris 1962), pp. 23-33.
37 Gimenez, F.H., "El convencional espinazo montafioso de orientaci6n este-oeste que
los ge6grafos arabes attribuyen a la Peninsula Iberica," al-Andalus 30 (1965): 201-275.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
FIGURE 1.8a. The Iberian Peninsula showing the approximate configura

tugar in the tenth century (from: Fletcher, R., Moorish Spain, New York, NY
that seems to have been recognized almost to the very end in

twelfth century when Muslim holdings in the Iberian Peni
become restricted to the region south of Cordoba.38 Figure 1-8b
that the detailed configuration of the tugur could be very com
resenting the results of successive local engagements. As in th
East, a more vaguely recognized zone south of the tugfur was
'awasim, towns assigned a defensive role, but the distinctio
tugur and 'awasim became increasingly vague as time went on
The population of the tugur in al-Andalus as in the Near East
in some respects from that of the more central kuras. In particu
38 Levi-Provencal, E., Espana Musulmana, vols. V of Historia de Espana, R.M

Madrid, 1957, p. 32 ff. and cf. 'lamina V' in Bosch Vila.
39 Cf. Ibn Hauqal for fine examples of the waxing and waning of the fortu
boundary places as malatia; also cf. ref. 25 above.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
FIGURE 1.8b. Map of a region in the Middle March of al-Andalus at 1086 AD

the interdigitation of nominal districts at the frontier between Leon and C
Muslim dominions (from: Bosch Vila, J., Albarracin Musulman, Vol. I, Terue
the early centuries after the conquest they provided homestea

bers of the conquering armies. In the middle and lower tuguir
made up in large part of Berbers-derived from sedentary ra
nomad tribes prior to the incursion of the Almohads-while in
(i.e., the most easterly) tugur the Arabs formed the dominan
In both cases there are indications that the remnants of the earlier Celto-
Iberian and Visigothic populations may have been scarce even before the
conquest and became more so as a result of the flight of residents in the
course of the wars of the Muslim conquest.
One can recognize many similarities between these tugur zones and
the Marches in the northwestern European world, and especially the
Marca (vulgo: marcha) Hispanica that from the eighth century on con-
stituted a Christian counterpart to the Muslims' border districts. Like the
Muslim tugur, the term 'marca' as used in this sense tended to imply not
only an ethnic but commonly also a religious boundary: 'marca paga-
norum' is not an uncommon phrase.
40 Bosch Vila, J., Albarractn Musulman, vol. I in: Historia del Albarracin y de su Sierra,
M. Almagro ed. (Teruel, 1959).

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
FIGURE 1.9. Copy of the parts of maps IV-1 and V-1 from al-Idrisi's Book
1154 AD, showing the configuration of the Iberian Peninsula. The map shows
the range of imaginary mountains labeled al-sarrat running just south of To
tula) E-W across much of the Peninsula.
Considering this sequence, it is clear that from the seventh well into
the eleventh century in al-Andalus the concept of tugur derived from
military realities, reflecting a long period of dynamic warfare ranging
for several centuries over a relatively restricted depth of territory, and
that, as a designation of geographic reality, the term designates a mil-
itarized zone of variable width, from about fifty to several hundred
miles in depth (Figs. 6 and 8a), appropriate to the kind of mounted
warfare typical of the time and constituting effectively a no-man's
land,
the 'white' (i.e., waste) lands, the 'terrae desertae' of the Castilians, and-
whether sparsely or more densely populated-studded with fortresses,
and at intervals marked by more important fortress towns that served
as sally ports or as refuges for the warriors engaged in the actual
fighting.41
41 Cf. eg. Bosch-Vila, J., "Consideraciones sobre al-taghr en el-Andalus y la divisi6n

politico-administrativa de la Espafia" in: Etudes dorientalisme dedides a la memoire de Levi-
Provencal, vol. 1. (Paris, 1962), 20-33.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
Tugur in the Texts of the Arabo-Islamic Geographers

Geographers recognized the distinction between the singu
and the plural tugur, designations, respectively, for a zone4
town in such a zone and exercising the functions of a tagr. Examp
distribution of localities so designated are shown in Figure 1
between the Byzantine Empire [including Armenia] and S
jazTra and compare with Fig. 1.6), Figure 1.10 (between
Andalus and the Christian kingdoms on the Iberian Peninsul
with Figure 1.8), Figure 1.11 (the eastern border of the Khur
between Upper Egypt and the Nuba kingdom. The distributi
fortress towns frequently suggests a line of defense or offen
tance inside of what constituted the actual lands claimed
rulers, that is, well within the tuguir and, at least in the Syrian
in that of al-Andalus, just outside the 'awasim.
The geographers' usage of these terms can be deduced fro
ination of the texts (Table 1.3): in the works of 18 geograph
a total of 37 towns listed as tagr, and 8 tugur were mentioned
These data, grouped according to author, are summarized in
the towns named as tagr are shown in the first thirty-seven row
tugur in the last eight. The several numbered columns corres
various authors, and a key to those numbers is shown on the
table proper. Examination of the table as a whole suggests t
the terms was by no means universal; on the average each a
nated as tagr fewer than 3.6 + 3.4 localities out of a total of 35; s
the eighteen authors mention towns designated as tagr, but
out of eighteen mention any districts designated as tugir. The
cation that either towns or districts of this kind were mentioned more fre-
quently over the period covered by this survey.
It is clear from these observations that both the singular and the plural
terms marked concepts familiar to the geographers but were hardly
viewed as of overriding importance. Al-Idrisi appeared to be the least con-
cerned with boundaries of this kind. He used the term only in reference
to a single place, in regard to the 'lands beyond the river' at the eastern
end of the Khurasan. His reluctance to underscore the hostile relations
between Islam and Christianity is understandable considering that as a
Muslim by birth, training, and faith, he wrote under the aegis and at th
court of a Norman king who had fought the Muslims in North Africa an
who played a major role in establishing Christian rule, albeit of an extraor-
dinarily tolerant cast, in Sicily.
To round out the geographers' picture of the zone referred to as the
tugur Table 1.2 showed fortresses named as lying in one such defensiv
zone (tugur and 'awasim) by three geographers, one writing near th
hayday of the Khalifate, one right before the Mongol invasions, and one
42 Ibid., p. 39.
43 For example, Ibn Hauqal: tugur at-turuk ... ma wara, 'n-nahr ... on the extreme
frontier (hadd) of Islam.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
FIGURE 1.10. Modern map of part of the Iberian Peninsula showing locati
referred to as tagr in Arabo-Islamic geographical texts. The Mediterranean
in the lower right hand corner, and the Bay of Biscay runs over much of the
to right. Barcelona is just off the map on the right, and Lisbon on the l
before the establishment of the Ottoman Empire, respectively

summary illustrates the concept of the zone of the tugur as a true
one, studded with fortifications of diverse sizes. Whether Tabl
recognition of a change in the defenses with the course of time
be determined from the available data.
Summary of First Section
To recapitulate the principal result of the foregoing examination of th

manner in which the two kinds of boundaries are represented in texts
and cartography of the medieval Arabo-Islamic period, one can hardly
do better than to cite the great fourteenth century Muslim historian Ib
Khaldun. He summed up his discussion of the overall result of the events
associated with the formation of Muslim states in a passage of hi
Muqaddimma:45
.Ji ,1 . sJJF t,t! 5J. J! L.. P. -Jj; J..
.6jLLJl ,b .L- . :_l . J 1
44 Cf. Honigmann, E., EI-1, vol. 8, article tugair, p. 739.

45 oTV. e jJ j., _ - .^l ; *oJ )O . jY !

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
FIGURE 1.11. Modern map showing region of Khorasan and Bokhara, with
indicating towns described as tugir in the geographical texts.
or in F. Rosenthal's translation:46 '. . wherever their (warrior

bers reach, their advance comes to a stop at (what is called) th
region' [F.R's translation of 'thaghr']. This surrounds each dyna
sides like a belt."
Taken together, data here assembled support the view that Arabo-
Islamic geographers, while recognizing the concept of boundaries as
such, did not accept the idea that these were sharply defined boundary
lines separating either, as internal boundaries, one Muslim state from the
46 Ibn Khaldun, The Muqaddimah, F. Rosenthal trans. (New York, NY, 1958), vol. 2, p. 125.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
TABLE 1.3. Fortresses in Tugur as-SamTa Listed by Three

Arabo-Islamic Geographers
al-Istakhri ad-Dimishqi al-Qalqashandi
Region Fortress 908 AD 1225 AD 1387 AD
T. as-Samia Mar'as x x
Tarsus x x x
Adana x x x
al-Masisa x x
al-Harunia x x
Sis x x
Aias x x
al-Kan.suada x
Ain zurba x
Sirfanakar x
ar-Raha x
Qal'a Ja'bar x
T. al-Jazirmta Malatia x x x
Kamakh x
Simat x
al-Birah x
Hisn x
Mansur
Qal'a ar-Rum x
Hadat al-Hamra' x x
Dalwagi x
Daranda x
Abulustan x
next- deemphas
of all believers
ritory, a conse
fluidity of a fr
cases they appe
zone separating
power from on
47 As a hypothesis
Muslim states I sug
medieval Muslim
termed the 'sover
sultana, both in ter
of as being maxima
ritories immediate
of as radiating in
according to some s
tromagnetic field a
as one ascends abov
regions where the
to be overlaid by t
turn would rise gr
peak at it's capital
A boundary, in su
boundary zones of

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
two possible
sovereignty/distance
- - - - contours of A
* .-
Relative distance from capital of A
FIGURE 1.12a. Hypothetical relation of intensity of sovereignty to distance from the capi-
tal cities of two states, A and B, and their relation to the boundary zone between them.
FIGURE 1.12b. Map to illustrate the relations between states conforming to the hypotheti-
cal relations illustrated in Fig. 12a.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
Such a view differed from that held by the Romans. It per

only over the 350 years of existence of an effective khalifate,
this well into the fourteenth century in spite of vast changes i
ical structures of the governments that prevailed over the
expanse of the Islamic realm. Such endurance of an important
ical manifestation would seem necessarily to reflect peculiarit
Islamic civilization of the Middle Ages.48 Validation of the ba
in terms of the experience of individuals other than the scho
have monopolized the discussion to this point, and inquiry
extent to which this characteristic of Arabo-Islamic geograph
understood as reflecting basic tenets of the society will occupy
section of this essay.
Appendix
Constitutional implications of the proliferation of states
during the Late Khalifate
Despite the emergence of the successor states there is no indication that the
primary task of the Islamic state had changed from that envisioned by
Muhammad: the state was the vehicle that assured that the Muslim could live
a life in accordance with the dictates of his religion-all other functions of t
state were considered subordinate to this mandate.49 Early on, the Muslim su
preceding sections of this essay (and, incidentally, as conforming to the general vi

advanced by Hourani). Such zones would be expected to coincide with the portion of th
curves of Figure 1-12a where the intensities of sovereignty of both neighbors would be
and change slowly with displacement on the ground. The width of such boundary zone
would be determined by the actual parameters of the two curves describing the relati
intensities of each of the two sovereignties in the region where their curves intersect, para
eters that may well have varied from one political unit to the next as well as from tim
another (these relations are illustrated in Figure 1.12a in terms of two alternative hypothe
ical sovereignty contours for unit A; the crosshatched area marks a possible bound
zone). In the two-dimensional map comprising a number of political units (Fig. 1-12b) su
a model yields a mosaic of sovereignty fields surrounded by boundary zones between ea
pair of adjoining powers. Incidentally, this figure emphasizes that the fraction of the t
area of a given entity that falls into boundary belt is substantially larger-and its relat
importance greater-than might have been inferred from consideration of the ratio of
linear dimensions of boundary and core zones.
This hypothesis implies that precisely positioned boundary lines were omitted from t
geographers' statements not by arbitrary choice, but on the basis of a knowledge of t
actual conditions prevailing in the medieval Islamic realm. The picture of a central cap
area surrounded by a zone of increasingly faint identification with the central zone's so
ereignty corresponds very well with the pattern of representation of such units in al-Idris
maps (Fig. 1.5): the central core of each polity is indicated, but the frontier is left blan
reflecting the low gradient of the sovereignty-distance relation in the border zone. Sim
larly, the rarity of border experiences of travelers to be discussed in Section 2 now foll
logically from perception of the variability of the intensity of sovereignty over its territor
Scholars holding the sort of concept proposed here would find such a view of the mat
to conform well with their almost universal tendency to describe a country in terms of
capital and major cities.
48 Maurice Lombard in his l'Islam dans sa premiere grandeur (Flammarion, Paris, 197
has made a case for one aspect of this, the astonishing commercial network that arose f
lowing the establishment of the Islamic Empire.
49 Primacy of religious function of imam.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
TABLE 1.4. Terms Used in Medieval Arabo-Islamic Literature Relating to the

State, the Ruler, and the Subjects
The ruler The subject The entity
khalifa sakin (resident) sha'b (first step in genealogies)
sultan hadari ('Anwohner') daula (dynasty)
malik muwatan ("native") wilaya (sovereign power)
'amir --(XIX cent.) wat'an (homeland)
mulk (kingdom)
"... the term 'citizen' with its connotation of the righ

of government was totally outside the Muslim politi
unknown in Islamic political language."
Bernard Lewis-The Political Lang
ject was bound to the state primarily by his obedi

the leader in war and at prayer. There could be o
no conflicts could arise in the loyalties of a subjec
and complexity this situation could not endure: r
proliferation of increasingly independent function
fragmentation of the empire described above. In v
religion and the conduct of secular affairs under Is
a situation would call for extensive reevaluation of
tutional basis of the Islamic polity.
Two problems, in particular, were at the core of
ification: the problem posed by the existence of
problem of the position of the secular leader, the s
One can follow the progressive breakup of the E
ature: early on it was affirmed that there could
of the Prophet-whose task was to assure that the
Law, was observed throughout his domains. Execu
by the incumbent to a hierarchy of functionaries o
(four such are recognized by al-Mawardi: power o
realm; power over specified affairs in all of the r
parts of the realm; and power over specified affai
realm).50 Over the years this position changed pro
ognition of the possibility that there could indeed
given time on the condition, at first, that the tw
rated by the sea, later on "that the seats of the se
distant from one another."
There was also a shift in the theoretical view taken of the holders of secular
power, from an initial position that saw in them mere functionaries of the khalifa,
through an intermediate stage in which the functionaries could perform their
duties independently but only if they acknowledged the suzerainty of the khalifa,
50 al-Mawardi, categories of functionaries.

51 cf. Lambton, A.K.S., State and Government in Medieval Islam (Oxford, 1981), chapters
IV to XI for extensive presentation of the progression: from Abu Yusuf (d. 798 p.e.), through
al-Baqillani (d. 1013) and al-Baghdadi (d. 1038 p.e.), to al-Mawardi (972-1058 p.e.), and from
him through al-Juwayni (1028-1105 p.e.) and al- Ghazali (1058-1111 p.e.) to ar-Razi (d. 1209
p.e.) and Ibn Taymiya (1262-1328 p.e.), to finish the development in the writings of Ibn
Khaldun (1332-1406 p.e.).

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
to a final position-attained when the khalifa had been physically eli

the Mongols-when the sultan, the holder of 'sultan', the secular pow
ognized as the head of an independent state who might or might n
form the functions of the imam for his subjects. This position had b
ever since al-Ghazali proclaimed that the degree of peace and civic
required for observance of an individual's religious duties could be a
by the existence of an effective ruler, and that this role so far predo
all others that legalists could enunciate the maxim: "Obey whoever i
regardless of whether he is pious and lives a good life or whether he
burden of his sins is upon him and not the subjects."52
It is significant that even while these discussions progressed
learned classes, there are indications that as far as the general pub
cerned, they continued to perceive their first loyalty to their reli
leader-or to the leader of a particular one among the more than a
into which Islam was divided by the thirteenth century- and felt little
to whatever dynasty happened to rule the land they lived in at th
Indeed, the political terminology of the period does not include an
conveys an idea analogous to the modern concept of 'citizenship' (c
leading a prominent Islamologist to conclude: "The term citizen with
tation of the right to participate in ... the conduct of government was
side the Muslim political experience ... and therefore unknown in Isl
ical language."53 The ordinary Muslim inhabitant of the realm was
(ra'iya, literally herd) of a ruler, not a citizen of a state, and his loyalty
dynasty was weak or non-existent unless some specific oath or vow
extracted from him in response to some specific benefit or un
member of some of the categories of people who owed loyalty to th
a matter of ascription.54 Clearly this would have been a factor rende
aries between Muslim states with Muslim subjects on both sides of l
tance in the lives of common people.
52 al-Ghazali, obedience to the sinful ruler.

53 Lewis, B., The Political Language of Islam (University of Chicago Press,
54 Cf. Mottahedeh, r.p., Loyalty and Leadership in an Early Islamic Society (P
versity Press, 1980).

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
SECTION 2
Travelers' Experiences at Internal Boundaries, the A

Concept in Arabo-Islamic Geography, and the Relatio
Zone-Boundaries to Basic Tenets of Arabo-Islamic Culture
he first section of this essay established that medieval Muslim ge

raphers and historians thought of frontiers in terms of bounda
zones rather than of sharply defined line boundaries. Does t
sense of gradual transition from one state within the Muslim Empir
the next correspond to the perceptions of the ordinary people of t
time? To answer this question we shall next examine the reports of t
elers who visited these lands in medieval days.
Boundaries in the Writings of Travelers in the Islamic Empire
Sea frontiers, unlike most other kinds, were recognized as sharpl

defined borders.55 For this reason the experiences of travelers arri
by ship provide an appropriate basis for comparison with experienc
associated with the crossing of zone type boundaries on land. Cargo
passengers carried on ships entering port from stations outside the ju
diction of a given ruler were inevitably assumed to be coming f
abroad.56 All arriving passengers were subjected to sometimes high
invasive examination. Customs charges-mukus, and later on sim
'ushr (1/10), or still later (under the Seljuks57) kumr (1/5)-cou
amount for non-muslims to as much as 20 percent of the value of a
goods carried. In addition, Muslim passengers were expected to carry
tificates of having paid the stipulated religious dues in their plac
departure, while non-Muslim dimmis had to prove previous paymen
jizia, the head tax. Subjects of non-Muslim countries proposing to v
any part of Dar al-Islam were obliged to possess or obtain a certific
of aman from a responsible Muslim subject of the country to which t
proposed to travel.58
Such customs revisions are documented for the port of Alexandria
Ibn Jubair in the eleventh century59 as well as by Ibn Battuta for
55 Cf. e.g., al-Mawardi on sea frontiers.

56 Khadduri, M., The Islamic Law of Nations (Baltimore, MD, 1966), 21; and G
L.A.R., "Al-Mawardi's theory of the Khilafa," Islamic Culture 12 (1937): 291-302.
57 Yusuf, M.D., Economic Survey of Syria during the 10th and 11th Centuries (Berlin, 1
120.
58 Hatschek, J., Der Mustamin (Berlin, 1919); and Wansbrough, J., "The Safe-condu
Muslim Chancery Practice"' Bull. Sch. Oriental and African Studies, 34 (1971): 20-35.
59 Ibn Jubair, abu 'l-husain muhammad, Rihla (Beirut, 1984), 35-36.
33

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
fourteenth,60 for Akka (while in the hands of the crusaders) i

enth;61 for Tripoli/east;62 for Qusair and al-Tur,63 for Aidh
Jidda.864 For Jidda on the Sea of Qulzum we have the additi
ment of the eleventh century traveler Nasir Khusrau th
exempted from the payment of the usual dues because he had
duced to the local governor as a jurist and a scholar, and that
turn wrote a letter to the governor of Mecca recommending
too, this scholar/pilgrim should be excused from the custom
ments.65 These, then, represent the normal and expected expe
travelers passing sharply defined external borders.66
By contrast, I have encountered no reports of this kind of b
trol in connection with overland travel anywhere within
realm. Circumstantial travel accounts that did mention the author's
difficulties in passing the authorities of ports invariably fail to men
any similar experiences at boundaries passed in the course of inl
travel within the Islamic realm. Examples are Ibn Jubair's Rihla (on
short section of which-the stretch between al-Madina and Damascus-
involved travel with the hajj caravan and hence would have been spar
any customs interference at that time),67 Nasir Khusrau's Safar nam-T
Ibn Battuta's Rihla,69 al-Mas'udi's Muruj adh-dhahab70 and Kitab
Tanbih,71 in scattered places in Ibn Hauqal's Kitab suratu 'I'ard,72 and
Benjamin73 of Tudela's Itinerary. Customs charges were reported on oc
sion as having been collected at the gates of cities well removed from
any frontier, or within the administrative district of such cities, rath
than at what might have constituted the border between two politic
entities.74 Thus, customs collecting stations along routes on land
on the Nile are mentioned for places along routes of travel by river
on land within the territory of Egypt-in Ikhmim72 (p. 149), in Asua
60 Ibn Battuta, abfi 'abd allah muhammad, Rihla (Beirut, 1964), 29.
61 Ibn Jubair, abu 'l-husain muhammad, Rihla (Beirut, 1984), 277.
62 Yusuf, M.D., Economic Survey of Syria during the 10th and 11th Centuries, 121.
63 Bjorkman, W., Encyclopedia of Islam I, vol. V, pp. 176-177.
64 Nasir Khusrau, "The Safar Namih," trans. M. Nakai, Ph.D. Thesis, U. of Tennessee
Knoxville, Tenn., 1979, pp. 56, 59, 66, 70, 149, 150, 151.
65 I have not been able to find corresponding travelers' accounts for either Mediter
nean or Atlantic ports of al-Andalus although clearly a considerable volume of commer
traffic must have moved through them; cf. e.g., S.M. Imamuddin, Muslim Spain 711-1
A.D. (E.J. Brill, Leiden, 1981), 130 ff.
66 Rivers and mountain ranges might also constitute natural and sharply defin
boundaries: Bartold, V.V., Preface to: Hudud al-alam, trans. V. Minorsky (London, 1970
p. 42
67 Ibn Jubair, Rihla (Beirut, 1954).
68 The Safar Namih, Travel Journal of the Persian Nasir Khusrau, trans. M. Nakhai (Knox-
ville, Tenn, 1979).
69 Ibn Battuta, Rihla (Beirut, 1964).
70 al-Mas'udi, muruj adh-dhahab wa-ma'adin al-jauhar (Beirut, 1976).
71 al-Mas'udi, Kitab al-tanbih wa-'l-Ishrdf (Leiden, 1967).
72 Ibn Hauqal, Kitab suratu 'I'ard (Beirut, 1992).
73 Goitein, S.D., A Mediterranean Society (Berkeley, Cal., 1988), vol. 1, p. 61; and Yusuf,
M.D., Economic Survey of Syria during the 10th and 11th Centuries, 118.
74 Tudela, Benjamin of, Itinerary, M.N. Adler trans. (London 1907).

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
CULTURE AND THE CONCEPT OF AREA 35
4 farsah from the land of Nubia; and, in Mamluk times, in t

of Qatia deep in Mamluk territory and not far from present
on goods in passage between Egypt to Syria;75 in North
Aleppo-a collecting station for customs in passage between S
Minor and Diar Bakr, Egypt, and 'Iraq76-and Manbij;72 in the
of the Jazira (Akhlat71); in Ifriqiya-Ajdabia near Barqa in pr
Libya where 'taxes on caravans traveling from or to the Coun
Blacks south of the Sahara were collected by the imam who
the Friday prayers';77 and in the Yemen- Khamdan.72 Other i
episodes akin to border experiences but taking place at som
from any actual frontier include mention by Ibn Jubair of fo
on the flanks of the Lebanon mountains;78 revision of bara'a,
cates of payment of the jizia needed by dhimmis crossing cu
tions of Muslim countries;79 the case of a caravan pilfered (ne
way to Malatiya) by Hamdani military authorities;8081 and p
episode of prisoner of war exchange on Lamis island 35 m
Tarsus.82 The one instance of what at first glance might bes
frontier line experience is a passage regarding the mores of r
on the Frankish side between Damascus and Acre in Ibn Jubai
but the distinctly non-official character of the actors involv
rather to underscore the conclusion that such sharply defined
points in general did not correspond to the ideas of the auth
Taken together, this information-or rather the sparseness
regarding customs and custom houses, confirms the concl
while such institutions existed and functioned in many of th
the Levant, of Egypt, and of Ifriqiya, comparable accounts fo
land are remarkably scarce. When they do occur they refer
variably to experiences in cities at some distance from any c
border crossing.
75 Ibn Battuta, abf 'abd allah muhammad, Rihla (Beirut, 1964), 55.
76 Yusuf, M.D., Economic Survey of Syria during the 10th and 11th Centuries
77 Ibn Hauqal, Abu al-Qasim Muhammad, Kitab suratu 'l'ard (Beirut, 1
Lopez, R.S. and I.W. Raymond, Medieval Trade in the Mediterranean World (New
52.
78 Ibn Jubair, abu '1-husain muhammad, Rihla (Beirut, 1984), 228.
79 Goitein, S.D., A Mediterranean Society, vol. 1, p. 62 and II, p. 384.
80 Ibn Hauqal, Kitab suratu 'I'ard (Beirut, 1992), 166.
81 al-Mas'udi, Abu '1-hasan, 'Alt Kitab al-tanbih wa'l-isaf, trans. Carra de Vaux (Paris,
1896), 259.
82 Ibid.
83 Ibn Jubair, abu 'l-husain muhammad, Rihla (Beirut, 1984), 273: ". . then, early in
the morning we resumed our travel (on the way from Damascas) to the town of Banias.
Halfway there we came across a huge oak tree of enormous bulk. We learned that this was
known as the 'tree of equity' (sajarat al-mizan). When we asked about the meaning of this
we were told: "This is the boundary between safety and danger (literally 'confidence and
fear') on this route, because of the Frankish freebooters. They murder and cut throats of
whomever they catch beyond this (tree) on the Muslim side, even by only a fathom or a
span beyond it; whoever is found on the Frankish side of it is made to continue his way.
This is a rule among them, one of the best and the strangest among those of the Franks."

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
Thus, such evidence as we have been able to draw from the

tation of customs institutions in the several traveler's account
the thesis that, apart from sea frontiers, sharply defined bou
within the Islamic Empire were either non-existent or of littl
importance-a conclusion that is well in accord with the ge
texts and cartography. In Spuler's words: "There were no sign
border area that the traveler was approaching or had already
border from one Muslim country to another."84
The Concept of Area in Muslim Geographic Thought
Line boundaries serve to define the extent of the territory of a

hence its area. In view of this, the neglect of the boundary c
the Arabo-Islamic geographic literature finds its reflection in t
of any perception of 'area' as a sharply quantifiable concept i
Islamic geography-both as a descriptive term on which co
might be based, and as an operant concept forming the basis
pound expressions, such as might provide a measure of intensi
other parameter (e.g., population density, or yield of grain p
a measure of the productivity of land).
It is probably no accident that, with the exception of t
example of the azala (a grossly asymmetric unit-100 by 1 dhi
'Abbassid times as a measure of the size of irrigation canals)
of units of area in the medieval Muslim literature and presum
units themselves are of Persian and Egyptian origin: the Egyp
sahm, yarad and the Sassanian jarib and qabiz85 (the 'asr
term derived from arabic 10-as a measure is 1/10 of the Sassan
The Arabo-Islamic geographers themselves made no use of t
cept of area.86 When they wished to indicate the size of a co
province they gave linear measurements of length and breadt
works of more than thirty medieval Muslim geographers
found any reference to what might seem the most natural next
of multiplying those two dimensions to provide an estimate
area for whatever entity had been described. This, of course,
imply that such a concept was unfamiliar to the mathematici
period. It does mean, however, that in describing real lands, or
structures, the people of the time made little or no use of th
of length to the second power. Thus, while area measurem
familiarly used in Sassanid Persia and in Byzantine Egypt, in
century the Persian Nasir Khusrau refers in his travel account
84 Spuler, B., Trade in the Eastern Islamic Countries in the Early Centuries in: D
Islam and the Trade of Asia (Oxford, 1970), p. 11.
85 Hinz, W., Islamische Masse und Gewichte (Leiden 1970), 65-66.
86 Brauer, R.W., "Geography in the Medieval Muslim World," Comp. Civilizati
(1992): 73-110.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
or twice in explicit terms to the area of a city or a buildi

reporting "length and breadth" of such elements many times
Here again, the facts as reflected in the texts are such as co
to the zone-type of boundary encountered in the geographic
real lands are not conceived of as sharply circumscribed areas
out on the surface but rather as domains of sovereignty of a r
monly reported in the fashion that elsewhere I have termed "
concept of geography" that is, as regions transected by in
routes of travel and marked by scattered towns connected by s
Land Tax Data as the Basis for Quantifying the Area of a Politi
As an alternative to recognizing explicitly the concept "area o
cal unit" as an abstract but quantifiable concept deducible fro
measurements on the ground, one might consider deriving an
lent sense from fiscal data. Thinking of a given province as m
a number of small parcels of individually held or worked l
result in the sum of their areas being a measure of the total ar
province. Encouraged by reports of Khalif 'Umar's cadastral sur
Sawad in Southern 'Iraq,88 one might consider the possibility
such a concept on the basis of data concerning revenues deriv
the land tax. In actual practice it is readily shown that such a
for the area concept could not have been feasible.
Because of the manner in which land ownership and the
tax liabilities were conceived, it is the karaj collected from
farmed by dimmis that is of concern here.89 Thus, the matte
87 Nasir Khusrau, "The Safar Namih," trans. M. Nakai, Ph.D. Thesis Univ. of Ten-
nessee, Knoxville, Tenn., 1979.
88 al-Mawardi, Abou '1-Hasan, Les Statuts Gouvernementaux, trans. E. Fagnan (Algiers,
1915), p. 370-372. Note, however, that more recent indications suggest that such surveys
as did take place produced data that seem incompatible with the actual dimensions of the
lands supposedly surveyed.
89 In the desert lands of the Arabian Peninsula as well as among the sedentary oasis
populations within the original heartlands of Islam, the concept of taxing land holdings
(and indeed the concept of individual ownership of land) was but slightly developed. What-
ever land tax was exacted from the oasis dwellers in pre-Islamic times was based on agri-
cultural production, computed as a fraction of the amount or value of each kind of pro-
duce harvested rather than on the basis of the area worked, and it was paid more in kind
than in cash. Early land revenue measures enunciated by the Prophet Muhammad (F. L0k-
kegaard, Islamic Taxation in the Classical Period (Copenhagen 1950), p. 185ff; and Chelhod,
J., Le droit dans la societe bedouine (Paris, 1971, c. 8) were based on the principle that within
the realm of Islam all land belonged to God and, by delegation, to the community of be-
lievers, and hence was not taxable as such (L0kkegaard, F, Islamic Taxation, c. 1). This idea
was firmly established for the sacred lands of the Arabian Peninsula and also accompanied
the victorious Arabic troops on their conquests: Arabs were not to be subject to any taxation
related to their ownership of land, but neither were they to be directly involved at this stage
with any aspect of agricultural use of the newly dominated lands-a matter left to the
increasing numbers of non-Muslims who, with the conquests, came under the rule of the
Khalifs and their successors. These clearly could not lay claim to a share in the bounty of
the communal land ownership of those confessing Islam, and hence were properly the sub-
jects of taxation related to their use of the land. It was probably under the khalifate of
Mu'awia ibn abi sufyan (661-680 AD) that these ideas were said by later writers (such as
al-Mawardi) to have been incorporated into the theory of a more formal tax structure of the

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
relevant to the present discussion only when substantial land

non-Muslim populations came to be included in the Islamic
a result of the early conquests in the Middle East and in Egypt
lifs 'Umar ibn 'abd al-kattab and 'Utman ibn 'affan. At that p
matter of taxation immediately was complicated by casuistry:
of the insistence of the khalifs upon the principle that soil o
should be restricted to Muslims alone, either in the form of p
benefiting the community of the faithful as a whole or as in
while yet Arabs were not supposed to till the newly conquere
non-Muslim non-Arabs had to be allowed one of several forms of usu-
fruct under conditions determined by the terms under which their land
had been incorporated into the Islamic Empire. Furthermore, in th
Middle East, land without an assured water supply is of little value, an
such as there was, was also assigned to one of several classes that affecte
the tax burden upon the lands served, creating further categories in th
overall tax structure. In the course of time, the financial needs of the
state - strained in their resources by numerous conversions that undercu
the tax base (as well as by the increasing shortage of silver for its
coinage)-led to the determination that transfer of land from dimmis to
Muslims did not (or not necessarily) lift the liability to karaj payments,
and here again, special rulings were required to permit what had earlie
been interdicted and to find a place for these payments in an increasing
bewildering tax edifice.91
All of this still did not establish a suitable standard for determining
the basis upon which any karaj to be collected should be calculated. Tw
alternatives are said to have been recognized as early as 639 AD by 'Uma
in connection with the formulation of a taxation policy for the Sawad
the fertile "Black Lands" of the Euphrates-Tigris Delta:92 a proportion
tax based upon the product of the soil, or a tax based upon the area tille
by a given taxpayer and differentiated according to the kind of crop,
i.. L, and L. = qusama and misaha, respectively.93 'Umar is sai
to have chosen to follow initially the patterns that had been applied in
the region while under Byzantine and Sassanian rule prior to the Muslim
conquest, and this approach became widespread in the sequel.
future Islamic Empire (cf. qur'an or hadith re land ownership): like most other facts of
given individual's life, his tax liability would be determined largely by his religious affili
tion: tithe, .' = 'ushr, and alms, -;j = zakat, for the Muslim; head tax, .> = jizia, an
taxes of various kinds, including in particular the land (use) tax, jointly referred to as e:
= karaj, for dimmis, the 'people of the book' living under the protection of Islam in Da
al-Islam.
90 Note that Hitti contends that the divisions mentioned above may be pure invention
introduced post facto by Mawardi and Ibn Yunus, and that the real situation was as
depicted in the next paragraphs, with minimal input of Qur'anic or later religio-political
concepts (cf. ref. 32).
91 Juinboll, Th.W., in Encyclopedia of Islam, vol. 3, p. 901.
92 al-Mawardi, Abou 'l-Hasan 'Ali, Les Statuts Gouvernementaux, trans. E. Fagnan
(Algiers, 1915), p. 370.
93 E.g., al-Mawardi, Abou 'l-Hasan 'Ali, Les Statuts Gouvernementaux, trans. E. Fagnan
(Algiers, 1915), p. 312.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
In essence there were three such patterns: that, already m

above, prevailing in the Sassanid Empire; that prevailing in t
Empire under Byzantine domination; and that evolved in Egyp
earlier post-Alexandrine and the subsequent Byzantine land t
All three can be reduced to a common denominator, harking b
tax reform undertaken under the Roman emperor Diocletian
This produced the compromise pattern of the "capilatio (et) iu
head and acreage) taxation as the universally applicable basis
structure,94 one readily recognizable in the land tax of th
Empire (although distorted during the later days of the regim
tices such as collective assessment of whole villages or even d
well as being the basis on which taxation of farmers was bas
antine Egypt. It was this pattern that was transposed in the e
structure of taxes on dimmis, karaj and jizia,95 and that was
of elaboration and complication in the manner described.
Even this complex pattern did not endure long. Actual distr
lands involved not only the differentiation of public lands and
tilled by dimmis, but to an increasing extent lands given to
a reward or in lieu of payment in the form of iqta'a of seve
holdings that increasingly assumed a quasi-feudal character a
held by Muslims, were subject to 'usr-the tenth (or later, kums
an impost related in effect to product rather than area of th
As a result of all these tendencies, the areas of any one of t
subdivisions of the Empire, and later on of any of the sever
ruled states, came to be made up of a complex mosaic of land
subject to a bewildering plethora of tax regimens only a sma
which corresponded to the original concept of land tilled by d
paying a well defined area-based (misaha) karaj. As the financi
of many of these political entities became increasingly preca
matter was still further complicated by a whole array of other ch
taxes related to the conduct of different forms of agricultu
revenue reports tended to lump these different kinds of imp
fashion that the figures bore little relation to area tilled or used,
volume of different kinds of produce harvested.99
The result of all these considerations is that land tax return
to fluctuate wildly (cf. e.g., Ibn Hauqal) and did not at any t
unique relation to land area in cultivation or pasture. Figures f
94 Ostrogorsky, G., History of the Byzantine State (New Brunswick, NJ, 196
95 Cambridge History of Iran 3[2], p. 746; and Hitti, P.K., A History of Syria
1951), p. 423.
96 Yusuf, M.D., Economic Survey of Syria during the 10th and 11th Centuries
97 Cahen, C., in Encyclopedia of Islam II, vol. 4, pp. 1030-1034, reports that f
diers were at times allowed to substitute military service for 'ushr payment
98 F. Lokkegaard, Islamic Taxation in the Classical Period (Copenhagen 195
99 Examples of tax revenue lists: UJaMCKO9c4Ha, A.C.- Xap,an> n
reorpa4aM Ix-x eB.- in: M6 H an-4>aKHX- Ax6ap an- SynAaH-
EpeeaH, 1979, p. 139-142, and Cahen, C., "Le regime foncier dans le Fayyoum Ay-
youbide," Arabica 3 (1956): 12 ff.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
area of any political or administrative unit were available for o

kind of tax computation only from early times, and then onl
such as those of Egypt and parts of the former Sassanian Em
area-based land tax had been in use before the advent of Islam.100
Summing up the limited data concerning the role of "area" in describ-
ing geographical entities, we conclude that it was given little or no sig-
nificance by the geographers, and that in relation to state imposts, too,
it exerted remarkably little influence. It was significant, at least in theory,
only in those places and in those periods where the influence of earlier
empires, Byzantine or Sassanian, continued to be strongly felt, and even
there its practical importance came to be overlaid by other considerations
and other measures well before the end of the eleventh century. All told,
these attitudes accord well with a geography in which sharp boundaries
played little or no role.
Boundary Characteristics as a Consequence of Embedded

Attitudes of the Culture
The data presented justify the conclusion that neglect of boundaries

was not only a fact of geographical science as it was written in the Muslim
empire from the eighth to the twelfth century. It also corresponded to
the experience of contemporary travelers, and was at least not in conflict
with the views of administrators of the period. These observations sug-
gest that this peculiarity of Muslim geographical perception, rather than
being fortuitous or determined by lack of technical resources, must
reflect certain fundamental traits of the intellectual or religious culture of
the Islamic Empire. In fact, at least six aspects of the medieval Arabo-
Islamic culture jointly or separately could have contributed to this result:
1. The 'umma, the community of believers, was defined originally in
purely religious terms as coherent and united by its creed-in this sense
all Muslims were brothers, not separated by any internal borders within
Dar al-Islam.l01 With the eclipse of the 'Abbasid khalifate, this percep-
tion gave way in the tenth and eleventh centuries to a period of decentral
ization and formation of separate states within the former boundaries of
the Islamic Empire. The most widely adopted position of the jurists prob-
ably was that of al-Mawardi (974-1058)102 who proclaimed that "only
where the Islamic lands are divided by a sea (and in later views perhaps
also by other natural barriers) the territory of Dar al Islam can the realm
be conceived of as divided into two (or more) political communities the
rulers of which are independent of each other though they owe ultimately
subservience to the Imam." Thus, internal boundaries, at the outset were
viewed as conflicting with the basic tenets of Islam; from the tenth
100 For example, Cahen, C., "Le regime foncier dans le Fayyoum Ayyoubide,/ Arabica
3 (1956): 8-30.
101 Khadduri, M., The Islamic Law of Nations (Baltimore, MD, 1966), 15-17.
102 Ibid, p. 21; and Gibb, L.A.R., "Al-Mawardi's theory of the Khilafa," Islamic Culture
12 (1937): 291-302.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
century on, such boundaries were recognized implicitly, but

to be viewed as in a sense non-canonical (cf. Appendix to S
2. Primary function of the state, in a view universally acc
Islamic jurists, was to assure that all Muslims could lead a life
with the mandates of the Qur'an and discharge their obligatio
Neither the unitary 'Umma of the seventh and eighth centuri
subsequently emerging political communities were territoria
although, to assure the basic functioning of such entities, cer
that have territorial consequences were prescribed for the Im
later times for the multiple rulers succeeding him in his sec
tions.104 In the resulting relation between the people at larg
rulers loyalties remained focused on religious aspects rather t
dynastic or territorial claims of the state.
3. In the relations between the territories forming part of
realm and adjoining non-Islamic states, the overtly dominant
played by the weight given to the command to engage in
the Holy War against the Unbeliever.105 In the Qur'an (IX, 29)
opinions of the early jurists-including most notably Abu
this command mandated Holy War against the Unbeliever only
latter threatened to interfere with the functioning of the Islamic
the middle of the tenth century the position had changed, a
from Abu Yusuf on pronounced the duty of all Muslims to wag
mittingly against the Unbeliever until all mankind should be
Islam.107 While the making of temporary peace-if it was adv
to the Muslim cause-was allowed, such peace must never l
than ten years at a time. For the centuries before the sixteenth
concept of coexistence and the principle of reciprocity in int
affairs came to be more widely accepted and the 'hot war' man
to be replaced increasingly by what Prince Juan Manuel in the
century called 'la guerra fra', 'the cold war'108) this mandate
dered the frontier between Dar al-Islam and Dar al-harb, the l
of the Unbeliever, permanently unstable and subject to the f
a continuing war of offense. Later on this changed to defens
largely as a war of movement (although siege warfare often
important role in the eventual outcome of these campaignsl0
4. Landed property, within this system, was owned excl
103 Vaticotis, P.J., Islam and the State (London 1987), 30.
104 al-Mawardi, Abou 'l-Hasan, Les Statuts Gouvernementaux, trans. E. Fag
1915), 30-32.
105 Ibid., 31; and Khadduri, M., War and Peace in the Law of Islam (Baltimore
that this mandate by itself defined the Islamic state as a 'transcendental gr
sense of Watt rather than as a state in the modern sense of the term.
106 Abu Hanifa, an-nu'man ibn thabit, kitab al- musnad, ed. Safwad al-saqqa (Aleppo,
1962).
107 Cf. Lambton, A.K.S., State and Government in Medieval Islam (Oxford, 1980).
108 Arias, L.G., El concepto de la guerra y la denominada 'guerra fria' (Zaragoza, 1956).
109 Lot, E, l'art militaire et les armees au Moyen Age en Europe et dans le Proche Orient (Paris,
1946), vol. 1, p. 17.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
Allah or, by delegation, by the Prophet and his successors, th

Man may have been granted the right of making use of the l
is never wholly clear whether such right of usufruct is conf
the community of believers as such or whether it can be held b
uals without violating some of the concepts of a strict orthodo
situation is further complicated by the fact that both the land
occupied territories and the rights of individuals to benefit from t
not uniform but were perceived by the 'rightly guided' khalifs
sequel by the jurists as falling into a number of categories dete
such factors as the mode of acquisition (e.g., by force or by pea
the role played by the possessor in the actual warfare leading to
sition of a particular stretch of land, his belonging to one or anoth
grouping, and of course his religious status as a Muslim or as
of the protected religions, a dimmi.111 As mentioned above, all
tors affected the tax liability upon the land, a matter further
by the fact that land in the Middle East for the most part can
tivated without water rights. These, too, were subject to furthe
zation, special provision being made for how the land was use
those lands not effectively utilized. Thus, the character of land
but precluded any effective conception of 'territory of a stat
essence of the identity of its citizens. Later writers who procl
duty of the subject to obey whoever held power over him112 fu
tributed to rendering all frontiers vague and temporary in th
minds.
5. The medieval Islamic Empire shared with most of its predecessors
in the Middle East a strongly urban orientation.113 While its soldiery
was often derived from the open lands-the camel-rearing bedouin at the
outset, and other non-city dwelling populations like the Turks later on-
its organs of government and administration, its intelligentsia, and the
repositories of its wealth were concentrated in its towns. The sedentary
farmer was held in little more respect among the urban populations of
the capitals than among the nomad bedouin of the heartlands: to all of
these the role of the rural population was primarily that of raising the
food needed to supply the cities' wants or to supplement the nomads'
diet,114 of supplying soldiers, of breeding beasts of burden, and perhaps
of providing some of the caravaneers for the transport of goods between
different cities. It seems inevitable that under these circumstances the
projection of power over the territory of a given ruler was not uniform
but rather radiated from his urban center(s) in ever diminishing intensity
110 F. L0kkegaard, Islamic Taxation in the Classical Period, c. II and III.

111 Cf. preceding paragraph and: F. L0kkegaard, Islamic Taxation in the Classical Period
72-91.
112 Ya'qoub, abou yousouf, kitabu 'l-karaj, trans. E. Fagnan (Paris, 1921), 15; and Vati-
cotis, P.J., Islam and the State (London, 1987), 36.
113 Vaticotis, P.J., Islam and the State, 20.
114 Cf. Brauer, R.W., "The camel and its role in shaping mideastern nomad societies,"
Comparative Civilizations Review 28 (1993): 106-151, esp. 117-118.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
as the distance from the capital towns increased. Thus, as develo

the preceding section, frontiers between Islamic states, rather tha
sharp dividing lines separating adjoining political units, came to b
where, as the power exerted by one ruler diminished to the van
point, the power an adjoining ruler was able to exert increased gr
from nil toward the level it might attain near the urban center of hi
rule (cf. 'Hypothesis' as discussed in footnote 47 and Fig. 12).
6. In what might seem to constitute a contradiction with wha
said in the preceding paragraph one must not overlook the p
influence exerted early upon Islamic political thought by the way
bedouin nomads of the Arabian Peninsula,115 an influence that
vasive especially in the field of law, and that left an indelible im
upon the sari'a land-law that has never been wholly dissipate
the present context the relevant factor is the matter of territori
as they are claimed by all components of the segmental society
great bedouin tribes. In a desert environment large stretches of l
worthless as such; property rights are in a sense punctate-they
to a small piece of home land, the hima (often based merely on t
of first occupation),117 to wells and to pasture areas. Rights ma
attach to the passage of others over any part of the tribal terri
defining lines of travel that the group may close or leave open at
cretion, but such rights are largely a measure of a tribe's ability to en
such rights militarily. None of these perceptions encourage the
nomads to think in terms of definite boundary lines surroundi
larger territory, although tribal memory may enshrine landma
help to define where a given group will claim rights on the basis
memory.118119 ('. . . points that marked the limits beyond whi
might reasonably expect to become the subject of hostile acts by m
of a neighboring tribe'.) The nomads' point of view came with th
warriors of the futuh to the newly occupied territories and was r
in the insistence by 'Umar ibn 'abd al-kattab and his successors u
importance of communally owned lands as a significant reso
the Islamic community, worthy of being defended against e
encroachment by lands allocated to individuals either in the
iqta'a or some of the more permanent forms of usufruct analogou
nomad hima. Here again, the overall effect of the several strictu
render the concept of a sharply defined boundary inconsistent w
'set' of the culture as a whole.
115 F L0kkegaard, Islamic Taxation in the Classical Period, 20.

116 Ibid., 20.
117 Ibid., 14-37.
118 Raswan, C., "Tribal Areas and Migration Routes of North Arabian Bedouins," Geogr.
Rev. 20: 494-502, 1951; cf. also: Stewart, EH., Bedouin boundaries in Central Sinai and the
Southern Negev (Brill, 1963).
119 Cf. Gulliver, P.H., Nomadic Movements in: Monod, T., Pastoralism in Tropical Africa
(London, 1975), 382 regarding the absence of operating concepts of territoriality among a
wide range of African and Near Eastern nomads.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
All these considerations, thus, would have c

scholars as well as the general public to think
Dar al-Islam and between Dar al-Islam and states of Unbelievers, as
vague and blurred rather than as sharp lines of demarcation. We must
conclude that this state of affairs-that corresponds to what was found
in the works of the geographers-is an expression of the very tenets that
defined the Muslim world as one subject to the religious precepts of
Islam and to all their implications for the management of political affairs.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
SECTION 3
Genesis of Boundary Zones Involving non-Arab Muslim

The Case of Asia Minor
id Muslim imperial societies other than the Arabo-Iranian

develop against lands of Unbelievers "external boundaries" t
show characteristics similar to those of the tugur of the Nea
East and the Iberian Peninsula?120 The one example of the formati
an enduring empire at the gates of the Arabo-Islamic state, and that
laps the later phases of that empire's waning, is represented by the
tracted series of struggles, lasting from the tenth to the fourteenth
tury, at the interface between Turkish forces and the Byzantine Em
The result was to bring about the Turkification of and the establish
of Ottoman predominance in Asia Minor. It may be significant that
the final firm establishment of Ottoman hegemony, the open and z
type frontiers of interest here seem to have closed to become repl
increasingly by firm borders more or less on the style of the late R
empire. This observation may well be of significance to a clearer u
standing of the political meaning of such frontiers in the late med
world.
Because of its protracted nature this situation affords a valuable study
of the forces giving rise to such frontiers and the peculiar characteristics
that they displayed. The Arab-led conquests of the seventh and early
eighth centuries that led to the establishment of the Arabo-Islamic
Empire progressed at such a rapid pace that it is difficult to unravel there
the processes underlying the formation of the tugur, the marches. By con-
trast, the Turkification of Byzantine Asia Minor leading up to the estab-
lishment of the Ottoman Empire lasted more than 300 years from the late
tenth to the fourteenth century and involved recurrent struggles between
the Byzantine Empire and entities of the Turko-Muslim world. As will
be shown presently, these events, like the earlier conquests, were asso-
ciated with the formation of boundary zones. Because of the extended
time span occupied by the events in Asia Minor, the repetitive nature of
the action, and the substantial physical expanse affected, one may hope
to find here a greater opportunity to analyze the interplay of the several
factors underlying the development of these zones than in connection
with the events of the Arabo-Islamic conquests.
120 The obvious question that remains is whether similar structures were to be encoun-
tered in connection with non-Muslim societies. A subsequent essay assesses the degree
of diversity of sequences leading to the formation of such geographical structures, and the
social and conceptual significance that may be attached to them.
45

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
TABLE 3.1. A Brief Chronology of Events in Asia Minor Pertinent to Emergence of Boun
Date p.e. Seljuk Turks Other Turks Date p
900 begin Oghuz migration 800-1050 Byzantium
begin formative stage/Seljuks

Tughril named Sultan/Baghdad begin Turcoman incursions in east Asia
Minor (1027 p.e.)
Battle of Manzikert-penetration of Danishmend ghazi in Asia Minor (1060)
Turcoman raiders to shores of
Aegean sea
Seljuks of Rum rulers in Konia Danishmendits eliminated (1180 p.e.)
all Asia Minor under Seljuks of Rum
1219-1236 Kaiqobad sultan of Seljuks at peak

of their power
1230 Baba Ishaq rebellion
1243 Kose Dagh-defeat of Seljuks by renewal of Turcoman raids on Byzantine 1
Mongols Asia Minor (about 1245)
1250 begin formation of ghazi beyliks- 1
Qaraman, Germian, Ottoman as well as
(after 1280) of pirate beyliks on South
and West coasts of Asia Minor
1300 disappearance of Seljuks of Rum strengthening of Ottoman State
1326 take Bursa
1331 take Nicea
1340 in control of all of Asia Minor 1
1356 victory over European knights at
Nicopolis
1400 all of ghazi beyliks in Asia Minor under
Ottoman rule
1402 Ottomans defeated at Ankara by Timur-I
Lang
1453 14
conquest of Constantinople by Mehmed II

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
GENESIS OF BOUNDARY ZONES 47
The Historical Setting*
We know little about the Oghuzz, the people from whom

took their origin in their pre-expansion domain north of
Their realm does not seem to have included any major tow
with other Turkish tribes of the period it is reasonable t
their tribal elements, too, were characterized by conical tri
terns and a pronounced chief-oriented organization,121 an
up of a mixture of sedentarized farmers and pastoral nom
sheep, cattle, perhaps some two-humped camels, and especi
In the course of little more than fifty years one group of th
known to history as the Seljuks,123 had become the m
largest of the Central Asian nomad empires, a realm e
great portion of what had been the Islamic Empire at
extending from Transoxania to the border of Syria, and fr
of Northern India to that of Georgia on the Black Sea (cf.
course of the centuries here under consideration the now
ized tribal grouping that gave rise to the dynasty of the G
entered Baghdad. Its leader, Toghril ibn Muhammad, accr
Khalif, had taken the title of Sultan, and assumed de fact
of the Islamic heartlands. In doing so he had deliberately
style of rule to the forms of the urbanized and sedent
Iranian bureaucratic government he had found in place in t
basid territories and adopted the manner expected of an I
(cf. Table 3.1 and Appendix to Section 3 for more complet
of the ensuing complex sequence of events).
By 1060 AD the Ghaznavids had been displaced from
remaining holdings in the Khorasan, the Qipchaqs had
the territories between the Caspian and the Aral Seas a
some distance to the east of there, and the Great Seljuk Em
empire of the Seljuk dynasty ruling in Baghdad), nearing i
impinging upon Byzantine territories in Georgia, Arm
Minor124 (Fig. 3.2b).
In the years following the Battle of Manzikert on the t
Armenia in 1071 AD, the battle in which the Byzantine em
* Cf. the Appendix to this section.

121 Barfield, T.J., "Tribe and State relations/' in P. S. Khoury and J. K
Statehood in the Middle East (University of California Press, Berkeley, 1990
122 Cf. Roux, J.-P., "Le chameau en Asie Centrale," Central Asiatic J.
123 It should be noted that, while all of the Oghuzz tribes seem to hav
selves to share a common blood line, 'tribes' of Turks in Central and W
the period here under consideration were often merely synthetic ones, the
ering of warriors around some particularly astute military leader with
invented ex post facto. This was certainly the case for the Osmanlis, an
Danish mendites as well as for the Seljuks. The leadership role of the ch
in the migrations and struggles that led to the formation of the Seljuk e
documented (cf. Cahen, C., trans. J. Jones-Williams, Pre-Ottoman Turke
124 Bosworth, C.E., "The political and dynastic structure of the Iranian
bridge History of Iran, A. Boyd ed., vol. 5. (Cambridge 1968), 298-305.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
FIGURE 3.1. The Seljuk empire near the end of the eleventh century (from
Pre-Ottoman Turkey, New York, 1968, p. 18).
IV Diogenes was defeated by Toghril's successor Alp Arslan

opened the floodgates of Turcoman raiding into Anatolia125 (cf
the representatives the Seljuk dynasty had despatched to Anat
themselves increasingly independent of the government at Ba
that from about 1110 on until about 1300 AD, while the power
Great Seljuks was waning,126 one can speak of the Seljuks o
rulers of an independent state, waxing and waning in extent as
tunes of war and politics changed in Asia Minor. Throughout it
this state resembled that of the Great Seljuks of the Near East i
rulers aspired to be recognized as Islamic rulers in the full sen
word, building mosques, making gifts, and above all ruling th
bureaucracy of Iranian style, supported by a professional arm
composition resembled the slave professional armies of Bag
Cairo.127
The almost constant warfare of the three centuries following
juks' entry into the Anatolian theater-from the time of the battle o
zikert until the rise of the Ottomans-saw the boundaries between the
Byzantine and the Turkish domains swept no less than five times over
large portions of Asia Minor:
125 Cahen, C., trans. J. Jones-Williams, Pre-Ottoman Turkey (New York, NY, 1968), p. 58.
126 Kopriilii, M. Fuad, trans. G. Leiser, The Origins of the Ottoman Empire (Albany, NY,
1992) (orig. 1935), 43; and Cahen, C., trans. J. Jones-Williams, Pre-Ottoman Turkey, 103.
127 Kopriilii, M. Fuad, trans. G. Leiser, The Origins of the Ottoman Empire, 77.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
0;
U o
( _
(W
df
O54
ao
0.0
St
-4 2.
'U
*_
JV
(i
u,

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
0
U)
4a
EU
1q
U,
a,
U)
a,
a,
U)
a,

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
1. in the initial advance of the Turks during the eleventh

twelfth century,
2. during the recoil of the western and southern boundari
Turks following the first crusade,
3. in the course of the recovery and completion of the ren
quest of western Asia Minor by the Turks in the late twelfth
4. as a result of the rebound of the Empire under the Lasca
first half of the thirteenth century and recovery by them o
and parts of central Asia Minor,
5. during the rebound-this time for good-of the Turk
Mongol pressure from the east in the second half of the
century.
In addition there were internal fights among the Turkish beyliks (Fig.
3.4), most important in the present context those that in the course of the
fourteenth century led up to complete domination of the peninsula by
the Osmanlis.
The Boundary Zones Associated with the Conquest of Asia Minor
The broad outlines of the geography of the successive borders

between the Turkish and Byzantine-held territories are reasonably well
defined and have been summarized, for instance, in the scholarly analysis
of Honigmann.128 The actual configuration of the interface between
Turkish and Christian lands in Asia Minor, however, and especially in
Anatolia-like that between Moorish and Christian lands in al-Andalus
(cf. Fig. 1.8b)-was almost certainly much more complex. Especia
during the first two centuries of the conquest, the Seljuk advances we
associated with Turcoman raids that often reached deeply into Byzanti
territory, leaving a mosaic of ravaged lands in which conditions appro
imated those characteristic of the major frontiers. No detailed data see
to exist to describe the actual distribution of the resulting fragment
boundary patches. There is, however, a consensus among the historia
dealing with these events that all of the resulting boundary represent
a no-man's land of some depth, rather than a front or a borderline in t
modern sense. This conclusion is enunciated over and over in Wittek's
work129 as in that of Kopruilii,130 in Cahen's131 as in Lindner's,132 Vry
onis'133 or Kwanten's,134 each based on numerous original sources.
These descriptions recall many aspects also characteristic of the Syria
and the Jazirian tuguir, especially in regard to the wasteland aspect o
128 Honigmann, E., Die Ostgrenze des Byzantinischen Reiches (Bruxelles, 1935).
129 Wittek, P., The Rise of the Ottoman Empire (New York, NY, 1971 [original 1938]).
130 Kopriilii, M. Fuad, trans. G. Leiser, The Origins of the Otttoman Empire.
131 Cahen, C., trans. J. Jones-Williams, Pre-Ottoman Turkey.
132 Lindner, R.P., Nomads and Ottomans in Medieval Anatolia (Bloomington, Ind., 1983
133 Vryonis, The Decline of Medieval Hellenism in Asia Minor and the Process of Islamizati
from the Eleventh through the Fifteenth Century (Berkeley, CA, 1971).
134 Kwanten, L., Imperial Nomads (Philadelphia, PA, 1979).

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
FIGURE 3.4. The Beyliks of Anatolia and dates of their establishment and disappearance, from: K
Ottomani Empire (Albany, NY, 1992 lorig. 19351), p. 3

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
the territory and the composition of the border population

resented boundary zones between Arab-ruled and Byzantine
As the invasion of Asia Minor progressed, the process can b
of as a gradual displacement of these old boundaries toward
even while the Muslim forces changed from those of the 'A
ifate to those of the Seljuk Empire. Under these conditions
surprising that the boundary elements should preserve muc
character in the process. What is new in the case of Asia Mi
here such boundary zones, rather than coalescing into a cohe
increasingly came to consist of numerous dissociated element
in the course of time, a substantial part of the area of the ent
and that this situation endured for several centuries, engulfi
regions in frontier conditions.135
Frontier Population and Boundary Zones between Selju

Byzantine States-Ghazi, Turcoman and Akritai
The most impressive similarity between the tugur of
Byzantine boundary in the Middle East and the Turko-Byzan
tiers in Asia Minor is revealed in the character of the popula
iting the two kinds of boundary zone. Their actual composit
135 Clearly, these boundaries arose in the course of warfare. All the evid
that to a large extent the conquest of Asia Minor by the Turks was the result o
out war of motion. In this kind of warfare the extreme mobility of the Tu
contrasted with the more laborious movements of the heavily armed Impe
frequently allowed the former to raid around and behind the backs of the
of Turcomans or Ghazi warriors and their raiding activities recurs like a pe
throughout the segment of history summarized in Table 3.1. While regu
troops undoubtedly played a primary role in consolidating the actual conques
tive to surmise that to a significant extent it was the role played by the irreg
ghazis and their tribal allies, that laid the groundwork for the enduring oc
land. Indeed, at some of the critical moments these forces acted, to say the
dently if not in direct opposition to the central government that controll
troops. To cite only two instances: In the second half of the eleventh centur
and after Manzikert, there can be little doubt that the Seljuk rulers, Alp A
Malikshah, would have preferred to concentrate on their campaigns or cam
Egypt and Ghazna rather than becoming embroiled with the Byzantines. Th
treatment accorded by sultan Alp Arslan to the captive emperor Romulu
the generous conditions offered the Byzantines testify eloquently to this pref
the raiding activity of the Turcomans both before and after Manzikert that fo
of the Seljuk sultan in spite of his wishes and caused him to dispatch his kin
with a modest force to lend some air of Seljuk interest in the resulting con
there can be no doubt that even under the Mongol protectorate over the S
after their defeat at Kose Dagh, and after the return of the Basileus to Con
Seljuks would have preferred to preserve the peaceful relations they had e
the Empire of Nicea under the Lascarid emperors. Renewed Turkish attack
by Turcoman raiding activity were resumed after 30 years of peace on thi
shortly after the defeat of the Seljuks in 1243 BC by 1300 BC had resulted
of Byzantine rule in western Asia Minor. These events laid bare the weakne
ernment and rendered its attempts at restraining the Turcomans ineffectiv
picture of the struggles between the Byzantine Empire and their Turkish
the coexistence of irregular ghazi warriors with more disciplined profession
appears as a recurrent theme. Raid and counter-raid produced a special kin
the borderland zone, that appears over and over as the true home of the b
and that as such put its stamp upon the Turko-Byzantine boundary.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
circumstances involved in shaping this society are most clearly

ible in the case of the protracted events in Asia Minor and mu
sidered next.
As rulers the Seljuks-both the Great Seljuks and those of Rum-had
successfully assimilated their government to the urbanized and sedentar-
ized government of the former 'Abbasid territories. Yet, it was inevitable
that large numbers of those who had come from an undisguisedly tribal
and nomad background should persist in their ways. Seljuk society
remained a mixture of sedentarized, urbanized subjects and of nomads
every bit as dependent as their forefathers upon-and as enthusiastic
for-the economic benefits and the excitement of ghazua-raids upon any
suitably rich and preferably sedentary peoples they could reach.
In the new setting of the Seljuk Empire such activities were sure to
be perceived of as troublesome by the rulers of the state unless they could
be directed away from their own subjects, preferably towards actual or
potential enemies. Under the circumstances it is hardly surprising that
the Seljuk rulers encouraged their tribal nomad subjects-termed Turco-
mans in much of the literature136-to raid across the borders, and espe-
cially into the relatively rich territories of the eastern portions of the
Byzantine Empire.
This movement was facilitated by the pattern encountered in the tran-
sitional Turkish empires 500 years before the time of the Seljuks where
there already was a spatial separation of the sedentarized (and at times
urbanized) elements at the geographic core of the state from the nomad
elements relegated to, or that sought out, the frontiers.137 In view of
what has been said about the makeup of the Seljuk state, it is not strange
that in much the same way the tribal nomad (and raid oriented) element
of its population again should have tended to congregate at the new fron-
tiers, giving a distinct character to an awakening frontier society, different
from that of the more peaceable kind of the more secure inner regions
of the realm.138 Such warriors came to be referred to as 'uy-turks' turks
136 Turcomans were raiding Turks said to have appeared in the Muslim literature for
the first time in al-Muqaddasi's 'Description of the Muslim Empire' (al-Muqaddasi, "ahsan
al-taqasim fi ma'rufa 'l-aqalim," BGA III, p. 160, referring to two frontier forts near of Isbijab
in ash-shash as "Frontier posts (sadd) against the Turks and the Ghuzz." In relation to
events in medieval Asia Minor it has come to be used to describe nomad elements retaining
tribal ties and economy, including a strong liking for the pleasures and the economic advan-
tages of raiding in the age-old tradition. (cf. HM3aM yn-MynK- Lt4iceT-HaM3-
nepeeoASt .H.3axoAepa- MocKBa 1949), p. 109 regarding the role of Turco-
mans in the Seljuk state). These elements were primarily, but not exclusively, Turkish; they
included among others some Kurdish groups. It is sometimes assumed that frontier war-
riors falling into this category, can be distinguished from, though at times associated with,
ghazi, warriors (cf. Cahen, C., trans. J. Jones-Williams, Pre-Ottoman Turkey [New York, NY,
1968], p. 58 comparing Turcomans in the army with Turcomans 'riding as ghazi')-
supposed to have been driven by religious motives-a view that I, in accord with Lindner
(Lindner, R.P., Nomads and Ottomans in Medieval Anatolia (Bloomington, Ind., 1983) find
implausible (see below for discussion of this point).
137 McGovern, W.M., The Early Empires of Central Asia (Chapel Hill, 1939).
138 Wittek, P., The Rise of the Ottoman Empire (New York, NY, 1971) (original 1938) and
Wittek, P., Das Fiirstentum Mentesche (Istanbul, 1934).

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
of the frontiers (the Turkish 'us' means either 'frontier' or 't

a weapon' and thus is closely analogous to the etymology of
'thughur', cf. footnote 18).139 These made up the core of the b
lation but were joined by other warlike elements, includin
Armenians and Slavs as well as Islamized members of the pre
Byzantine provincial society, all combining to form what W
the "mixed border population." Because the Turks swept over
parts of the peninsula, the frontier between Byzantine and T
more complex and disjointed in the course of the ensuing wa
Turks were introduced into additional partially devastated dis
which by no means all the Greek inhabitants had fled,14
further regions that to a significant degree acquired the chara
a Greco-Turkic frontier society (cf. Table 3.2 that calls attention t
ilarity in the general makeup of the border population in An
in the Syrian and Jazirian tugur).
Adding weight to the resulting movement was that by this
of the Turkish tribes had been Islamicized for several decades
icized for the most part by somewhat heterodox but all the m
dervish preachers. Among the duties that these promulgated
ticular vigor was that of pursuing the Holy War against the U
and especially the unreformed Christian. Men who took u
selves this obligation were popularly referred to as ghazi-
unintentional indication that what was aimed at was the 'ghaz
in its original meaning as used to this day among the Bedoui
for the cattle and slave raid in the ancient tradition. Thus, for
condite marauders the economic attractions of raiding into By
ritory were reinforced by the religious motif that made folk
what might otherwise have been no more than successful in
The poems singing of the exploits of Dede Qorqut,142 of
Danishmend143 and of Sayyid Battal144 bear witness to this
enon. Historians have debated the relative weight given am
ghazis to economic motivation and to the primarily religious
of devout Muslim warriors engaging in Holy War against infi
in accordance with Qur'anic mandate. Some have focused the
upon the self-description of these warriors as 'ghazis' to infer
religious motive76 while others have noted that these 'ghazis'
any means restrict their raiding and war-making to infidel o
to heretics, but gleefully and in more than one location a
orthodox believers.145 It seems to me that the argument is s
139 Kopriilii, M. Fuad, trans. G. Leiser, The Origins of the Ottoman Empire,
W., Turkestan down to the Mongol Invasion, 2nd edition (London, 1958), 61.
140 Vryonis, The Decline of Medieval Hellenism in Asia Minor and the Process
from the Eleventh through the Fifteenth Century (Berkeley, CA, 1971), p. 25.
141 Wittek, P., The Rise of the Ottoman Empire (New York, NY, 1971) (origin
142 Rossi, E., trans., II Kitab-i Dede Qorqut (Citta del Vaticano, 1952).
143 Melikoff, P., trans., La geste de Melik (Danishmend, Paris, 1960).
144 Ethe, H., trans., Die Fahrten des Sayyid Battal (Leipzig, 1871).
145 Lindner, R.P., Nomads and Ottomans in Medieval Anatolia, 78.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
TABLE 3.2. Composition of the Mixed Border Population

In Syrian and Jazirian In Anatolian
In al-Andalus tugur border zones
Berbers (from sedentary tribes) Arab tribe
[In middle and lower turcoman)
tugur]
Arabs [in upper tugir] Iranian tribesmen (Arab tribesmen)
Celtiberians and Visigoths Slavs (Slavs)
[largely displaced]
Gypsies (Zott, Sayabiga) Armenians
Jarajima Islamized Greeks
Original Population
(Armenians, Syrians)
[Bosch Vila-Albarracin [Wellhausen-Das Arabische [Wittek-Das Konigreich
Musulman] Reich] [Canard-Djaradjima] Mentesche]
futile. Clearly, the logics of the case render this not an 'eith
rather an 'and' kind of situation.
The numbers of warriors of this kind and of associated nomadic
groups converging upon Asia Minor were swelled, especially during
thirteenth century, by nomads displaced from their pasture lands by
advances of the Mongols across Khwarezm and much of Persia.
The existence of such a warlike border society did not remain wit
influence upon the conduct of the hostilities between Byzantines a
their Muslim opponents. Curbing the movement of these frontier warrio
and especially adjusting the timing and the intensity of their raidi
activities-to conform to some sophisticated political motives could
be easy for any central government. Indeed, throughout the 300 y
long campaigns that eventually made Hellenistic Asia Minor over in
Turkey one finds that events are apt to be driven by two distinct w
not infrequently in conflict (cf. note 129): On one side is a central
ernment with aspirations of restoring the Islamic Empire as it had b
during the heyday of the 'Abbasids by the carefully timed and dire
use of organized troops made up increasingly of slave levies as the tr
element was thinned out. On the other hand we find the unruly T
coman and ghazi forces, more concerned with booty, pasture, and
Holy War than with the strengthening of a state that they saw as li
more than a relentless collector of taxes.146'147 Ghazi raids thrust
beyond or around the advances of the regular armies are, in co
quence, a frequent feature of these wars and could not fail to have a
found influence upon the configuration of the actual boundary.
146 Cf. e.g., Cahen, C., trans. J. Jones-Williams, Pre-Ottoman Turkey, 103; and Kopr
M. Fuad, trans. G. Leiser, The Origins of the Ottoman Empire, 78.
147 Cahen, C., trans. J. Jones-Williams, Pre-Ottoman Turkey, 88; and Pertusi, A.,
storia e leggenda: Akritai e ghazi sulla frontiera orientale di Bisanzo" in: Rapports d
Congres International des Etudes Bizantines, IIe theme: Frontieres et Regions Frontieres du V
XIIe siecle (les Frontieres Asiatiques) (Bucarest, 1971), 68.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
A final factor influencing the character of the society eme

border districts was the situation on the Byzantine side of
This was somewhat similar to that on the Turkish side but, i
more complex. Here, too, one finds border warriors dedicat
ghazis, to unremitting frontier war of raid and-especially,
out- defense against raids. By choice, Byzantine frontiers we
ever possible in regions where natural defenses made defense
cially mountains and mountain passes. From the Greek term
passes these warriors came to be known as Akritai (at a late
of the frontier fighters were known as 'apelatoi' a name wit
derogatory implications and not entirely clear meaning, wh
Akritas' came to be restricted to the leaders of such frontie
Unlike the ghazi one does not encounter mention in the lit
these Greek frontier warriors engaged in extensive raiding-t
the centuries here under consideration seems to have been
sure defensive. One may wonder to what extent this differ
from the historical situation in which they were cast, and to
it may reflect that the Akritai arose in a sedentary rather th
inantly nomad society. Like the ghazi they formed a populat
position of which was distinct from that of the peacefu
Among other warlike elements they, too, included not a
largely, it seems, brought in from the north, with Bulgha
elements.149 Thus, the march population on the Byzantine
altogether different from that on the Muslim side. However
to the ghazis, these Byzantine frontier warriors counted on
direct support from the state, in terms not only of remission
assignment of lands-these applied as well to the uy-Turk
terms of direct subventions. When these sources of suppor
pered with, as they were under the Paleologi, Byzantine fron
were prone to desert to the enemy and the subsequent colla
defenses in that case testifies to the importance of the Akrit
of the borders.150
The border populations on the two sides of the boundary
148 Roux, J.-P., "Le cheval et le chameau en Asie Centrale," Central Asi
(1959): 35-76, p. 76; Ahrweiler, H., "La Frontiere et les Frontieres de Byza
in: Rapports du XIV Congres International des Etudes Bizantines, IIe theme, Fro
Frontieres du Vile au XIIe siecle (les Frontieres Asiatiques) (Bucarest, 1971),
"Tra storia e leggenda: Akritai e ghazi sulla frontiera orientale di Bisanzo"
XIV Congres International des Etudes Bizantines, IIe thtme, Frontieres et Regions F
au XIIe siecle (les Frontieres Asiatiques) (Bucarest, 1971), p. 37; Lindner, R.P., N
mans in Medieval Anatolia, 12.
149 Vasiliev, A.A., History of the Byzantine Empire, vol. 1 (Madison, Wisc.,
Honigmann, E., Die Ostgrenze des Byzantinischen Reiches (Bruxelles, 1935),
150 Lindner, R.P., Nomads and Ottomans in Medieval Anatolia, 17, 38-40;
storia e leggenda: Akritai e ghazi sulla frontiera orientale di Bisanzo" in: R
Congres International des Etudes Bizantines, IIe theme, Frontieres et Rdgions Fron
XIIe siecle (Bucarest, 1971), p. 37; Koprulii, M. Fuad, trans. G. Leiser, Th
Ottoman Empire, 80 and note that Lindner, p. 17ff. also comments on th
encroachment of big land owners on landholdings of small farmers at th
tributing to their loss of interest in frontier defense.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
many basic attitudes and experiences. Similar occupations a

to the groups on either side of an ill-defined frontier's per
they had more than a little in common-in spite of the allege
differences created by adherence to different religions, an
loudly proclaimed themselves 'warriors of the(ir respect
Intermarriage provided a subject for some of the most famo
on both sides of the frontier.151 Travelers' reports and the r
epics of the period confirm that out of them arose a mixed bord
more or less homogeneous within itself but profoundly dif
the population of the undisturbed hinterland on either
course of time it comprised a mixture of Turkish frontier f
antine frontier fighters (Akritai and Apelatoi) including not
and Slavs, and, as the frontier conditions endured, a substan
tion of the old residents of the conquered territories.152 T
that these borderlands represented a region largely indiffere
ment control, they were also places of refuge for martial ele
out of favor with the central government.153
Thus, this border population constituted a distinct elem
Seljuk state and its successors. Distinct in many ways from
and sedentarized elements of the core areas, it also was in m
unlike the nomadic and tribal elements subsisting at a distan
frontier. Unlike what seems to have been true in the Near East where the
tugiur accounted for a small fraction only of both space and population,
the history of Asia Minor in the first three centuries of the present mil-
lennium caused these frontier zones to engulf a substantial part of the
peninsula and hence to have a major effect on the characteristics of the
emerging society of the newly formed Ottoman Turkey. In this sense it
seems clear that the boundary zones that developed in the course of the
Turkish conquest of Asia Minor did indeed perform at least one key fron-
tier function in the sense of Turner and his critics.154
The Turks' conquest of Asia Minor is part of their advance toward the
West-one arm of a movement of these central Asiatic peoples away from
a region that was becoming both more crowded and climatically more
inhospitable to the activities of nomadic pastoralists.155 In the thirteenth
century the impetus of this movement was increased under the impact of
151 Cf. e.g., Digenis Akritas, The two-blood Border Lord, D.B. Hull trans. (Akron, Ohio,
1972).
152 Pertusi, a., "Tra storia e leggenda: Akritai e ghazi sulla frontiera orientale di
Bisanzo" in: Rapports du XIV Congres International des Etudes Bizantines, IIe theme, Frontieres
et Rdgions Frontieres du Vile au XIIe siecle (les Frontieres Asiatiques) (Bucarest, 1971), p. 38; and
Canard, M., Les relations politiques et siciales entre Byzance et les Arabes (Dumbarton Oaks
Papers 18, 1964), p. 45; and note that Cahen, C., trans. J. Jones-Williams, Pre-Ottoman
Turkey, 63, presents evidence for not infrequent fraternization between border warriors
from the two sides fraternizing.
153 Cf. e.g., Wittek, P., Das Fiirstentum Mentesche, 6.
154 Turner, F.J., The Frontier in American History (New York, NY, 1920); and Walsh, M.,
The American Frontier Revisited (Atlantic Highlands, NJ, 1967); also cf. Discussion in Section
3 below.
155 Climate in central Asia in twelfth to fourteenth century.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
the Mongol conquests.156 The unique development of the fr

have seen in Asia Minor in the long course of the confl
Byzantium and the Turks, however, remains unexplaine
takes into account the presence of the frontier warriors at
between the two. In particular the ghazi on the Turkish sid
warriors wholly dedicated to aggressive frontier war, p
extraordinary mobility, retaining links with the nomadic s
which a great part of them sprang, and are responsible for t
dinary resilience of the "mixed frontier population" in shifting
rior to pastoral means of gaining a livelihood. The historica
show them everywhere associated with the Turko-Byzantin
and with the recurrent formation of a boundary zone cons
man's land that yet was not void of inhabitants. Compariso
counterparts on the Byzantine side has revealed the many po
ilarity between the two bodies of frontier warriors, and the
association renders plausible their melding to form a mi
society that provided a boundary zone population distinct in
from that of the peaceful hinterland-a society inclined t
rather than to exclude other population elements.157 Only su
existing over centuries under the shifting conditions of th
between Byzantium and the Turkish invaders can account fo
that converted conquered hellenized Asia Minor into Ottom
We have here the conditions that gave rise to the boundar
were so prominent in Asia Minor. As was shown in the firs
this essay, similar structures were shown to be present
theaters that saw the development of boundary zones
boundary lines between Muslim and Infidel societies: the tu
northern extremity of the Arabo-Islamic Empire, and thos
vened between largely Berberized forces in the Iberian Penin
Christian kingdoms that persisted in the northern part of th
The common denominator among these events is the associa
zone type of external boundary with Muslim nomads (cf. T
least three different nomad populations were involved in t
Turkish ghazis and Turcoman tribesmen, Arab Bedouin,
thus, within the narrow limits of the available evidence the t
of these nomads would appear to have been of secondary im
In each case, too, the nomad population involved was ass
other elements that formed part of a fully organized state, soldi
as administrators. In the case of the Turks considered in th
156 Kwanten, L., Imperial Nomads.

1573.B. YVanoosa, A.n. Ka<AKaH, P.M.,EOpT"KHIFH- COLHianHaA
CTpyKTypa BOCTOM4HIX rpaHMU BH3aHTHM CKO HMFnepHM- in: Rapport
du XIV Congres International des Etudes Bizantines, IIe theme, Frontieres et Regions Frontieres d
Vile au XIIe siecle (les Frontieres Asiatiques) (Bucarest, 1971), pp. 21, 26; Ahrweiler, H.,
Frontiere et les Frontieres de Byzance en Orient in: Rapports du XIV Congres International
Etudes Bizantines, IIe theme, Frontieres et Regions Frontieres du Vile au XIIe siecle (les Frontiere
Asiatiques) (Bucarest, 1971), p. 16.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
pages these elements were made up of the several states dir

the Great Seljuks, the Seljuks of Rum, the Ottomans, the
(though the latter proved unstable and do not seem to hav
to any great degree to the eventual course of affairs). In th
Arabs of the Futuh, the contrast between Bedouin warrior
the denizens of the cities was unmistakable from the time
Muhammad on and persisted into the time of the 'Abas
regular army had come to be increasingly staffed by slave
the case of the conquests on the Iberian Peninsula the
fighting was carried largely by Berber troops while the govern
Emirate of al-Andalus was largely dominated by the more
Arabs who saw to it that the frontier regions, especially
more inhospitable mountain regions would become the
former soldiers.159
It is tempting to see in this common pattern a special v
antinomy between 'the desert and the sown' perceived by
as the main driving force of history in North Africa.160
Appendix
The Historical Sequence of the Events Leading to the Conversion of
Byzantine Asia Minor to Ottoman Turkey
Predecessors of the Turkish tribes that eventually impinged upon the Byz-
antine Empire are recognizable at least as far back as the second century before
the present era: the Hsiung-Nu that appear in Han Chinese historiography were
horse nomads probably akin to the later Turks, occupied the eastern portions of
the Central Asiatic steppe region, northwest of the borders of the Han Empire
(Fig. 1). They showed considerable organizational and military aptitude as mounted
archers, and may well have been the ancestors of the Huns that overran much of
Europe in the first century of our era. The literature suggests that they gave little
weight to the individual ownership of land and to frontiers.161 Yet, they had a
strong sense of the importance of holding on to tribal lands as witness the state-
ment of the Hsiung Nu ruler Mao Tun (ca. -190 A.D.): "Land is the basis of a state;
disintegration of a tribe is inevitable when its land is lost."162 Turks appear first
under that name as iron workers to the Yuan Yuan in the sixth century. From the
earliest mention of them on they appear as a people that was not purely nomadic;
pastoral nomads among them coexisted with farmers and blacksmiths163-a pat-
tern that in later times allowed the coexistence of sedentary and nomadic turkish
elements within the confines of the highly urbanized Seljuk state.
158 Bedouin vs. urban, Hodgson, Venture of Islam.

159 Bosh Vila, J., Albarracin Musulman- El reino de taifas de los Beni Razin, hasta la constit-
ucion del seiorio Cristiano (Teruel, 1959), 65, 66; Fletcher, R., Moorish Spain (New York, NY,
1992) 19-20.
160 Ibn Khaldun, The Muqaddimah, F Rosenthal trans. (New York, NY, 1958).
161 Cahen, C., trans. J. Jones-Williams, Pre-Ottoman Turkey, 3.
162 Watson, B., Records of the Great Historian of China (New York, 1961), p. 2:162; also
Kwanten, L., Imperial Nomads (Philadelphia, PA, 1979), p. 43.
163 Kwanten, L., Imperial Nomads, 32 ff.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
A Turkic empire arose in the steppes of eastern Central Asia by th

tury but broke up sometime before the end of the eighth under attack
nese and Tibetan forces164 to give way to the formation of a numbe
states such as the Uighur, the Qaraqanid (according to Cahen the firs
Turkish states-in contrast to others such as the Ghaznavid state, tha
fact Islamic states ruled by Turkish ghulam [ex-slave soldier] prince
and, a bit later, the Khwarezmian and Seljuk empires.166 All o
reported, were sedentarized, separated from one another by wide f
inhabited by Turkic nomad tribes. As will be seen presently, this
frontier-zone pattern, too, was carried by the Turks into their occup
Ottoman Asia Minor.
At the beginning of the tenth century the region south of the Caspian and Ara
Seas immediately east of the Persian desert (the Khorasan) and extending into
the southern reaches of the Land beyond the River (Oxus)-Ma wara' n-nahr o
Transoxania-had fallen to Arabo-Iranian advances and then, under two of the
Khalif's governors, had become largely independent of Baghdad to form the
Samanid and Ghaznavid states. For a time, the Samanid state constituted the
border between Islamicized lands and at best very incompletely Islamicized
Turkic Qaraqanids. Beyond the northern border of the Samanid state, between
the Caspian and the Aral Sea and for some distance to the east of the regions
dominated by the Qaraqanids were the lands of the Oghuz tribes to whom the
Seljuks trace their origin167 (cf. Fig. 2a).
In a kaleidoscopic series of changes the next sixty years saw the disintegration
of the Samanid state and its conquest by the Ghaznavids by 1030 AD, the expan-
sion of the Oghuzz holdings to occupy most of the Khorasan by 1050, and the
projection of Seljuk power through central Persia into Baghdad, overthrowing
the Buyid usurpers of the 'Abbasid power168 and forming a firm alliance with
the Khalif.
The eleventh century169 witnessed the entry of the Seljuks into Baghdad and
the recognition of their chief as the Sultan by the Khalif. Sultan Toghrul as well
as his successor Alp Arslan adopted as their grand strategy the concept of a
restored Sunni orthodox Islamic Empire including Egypt as well as the marches
of India; the conquests this implied required that the Seljuks should keep their
western flank safe by ensuring peaceful relations with their Byzantine neighbors.
By contrast, autonomous nomadic groups of Islamized Turcomans entering the
region saw a rich opportunity for plunder in the wealthy western provinces of
Byzantium and began a series of raids beyond the frontiers of the Seljuk terri-
tories from about 1020 AD on, even while the Seljuk armies studiously avoided
any confrontations with the Greek Empire. The battle of Manzikert changed that
situation insofar as here Alp Arslan effectively destroyed the ability of Byzantium
to defend its eastern frontier. This in turn allowed what had been a pattern of
isolated Turcoman raids to swell to a flood that presently engulfed all of Asia
Minor from the Armenian marches to the Aegean Sea. In this advance Seljuk
164 Ibid., 27-48.

165 Cahen, C., trans. J. Jones-Williams, Pre-Ottoman Turkey, 10-11.
166 Kwanten, L., Imperial Nomads, 49ff.
167 Houtsma, M.Th., "Die Ghuzenstamme," Wiener Zschr. f/d/ Kunde des Morgenlandes 2
(1888): 219-233.
168 Hourani, A., A History of the Arab Peoples, 296.
169 For references upon which this and the succeeding paragraphs of the Appendix are
based see text of the section above.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
troops and Seljuk leadership became involved only haltingly, enter

only after the floodgates of what amounted to a Turkish mass mi
opened. By the end of the century all of Asia Minor from the Armenia
gian marches to the shores of the Aegean Sea was in Seljuk hands.
Turcoman raiders now no longer invariably returned with their boot
ever they had started from but in many cases came to stay among the e
itants in what had been Byzantine lands. Some of these newly resi
mans now remained as apparently peaceful transhumant herdsmen y
their potentially aggressive inclinations, while others showed thems
guisedly as frontier warriors entirely dependent for their livelihood
and the sale of prisoners in eastern slave markets. Contacts with the
antine population were inevitable and by no means always hostile,
this early time would seem to have included some cross marriages.
process resulted in the formation of a mixed border society much
familiar from the Syrian and Mesopotamian tuguir, a society that wo
become characteristic of much of Asia Minor. This stage also saw t
ment of the state of the Seljuks of Rum, its capital first in Nicea and la
as an entity increasingly independent of the Near Eastern Great Sel
inclined to copy their aspiration to be recognized as a bona fide Islam
contrast to this, Turcoman warriors carved out for themselves what
called a 'ghazi state' identified with the Turcoman ghazi warrior ch
mend. This entity came into being not long after the battle of Manzi
about a century dominated the northern central parts of Anatolia.
Byzantium had seen a great deal of internal unrest during the cen
had weakened its military stance as well as it border defenses and
greatly to the defeat at Manzikert. Near the end of the eleventh centur
the passage of the army of the First Crusade had brought it military
allowed it to recoup some of its defensive strength and to recover
western Asia Minor and of the southern coastal districts, forcing the
ments back onto the Anatolian Plateau. The advance of the Byzanti
an end with their crushing defeat at the hands of the Seljuks at My
a defeat again in part attributable to the harassing activities of auton
coman raiders. During this period the Danishmend state broke up in
of internal quarrels, leaving the Seljuks as the dominating force in
Minor.
By the end of the twelfth century they had recovered the territory lost to the
Byzantine counterattack and once again were masters of all of Asia Minor. The
thirteenth century opened with the conquest of Constantinople by the armies of
the Fourth Crusade and the consequent displacement of the Byzantine Basileus
from the European continent to exile in Nicea, resulting in the formation of the
Nicean Empire out of the remaining Byzantine holdings in Asia Minor under the
government of the capable Lascarids. The attention of the Empire now came to
be concentrated on Asia Minor, strengthening its military position there to the
point where western Asia Minor was once more recovered from the Turks, so that
the Byzantine-Turkish frontier was pushed eastward well onto the Anatolian pla-
teau. The military strengthening of the Byzantines was by force recognized by
the Turks and provided the basis for nearly a half century of peaceful and even
friendly relations between Seljuks and Lascarids, a period marked by thirty
years' suspension of Turcoman raiding into the Empire.
The next stage was predicted beginning early in the thirteenth century by the
appearance in eastern Asia Minor of numerous Central Asian Turks displaced

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
as a result of Mongol advances. There is evidence of unrest among T

ments within the confines of the Seljuk state in Asia Minor: the rev
with the name of Baba Ishaq was at base a religious disturbance bu
threatening proportions because it was taken up by disaffected elem
frontiers. On the Turkish side this situation was rendered more th
when at Kose Dagh Seljuk forces suffered a disastrous defeat at th
relatively small detachment of Mongol troops, and when, to the d
its ghazi warriors, the Seljuk government showed itself increasingly
to Mongol dictates in the sequel. By the end of the century, the Selju
increasingly deprived of the support of the warrior elements that h
backbone of its strength, faded into insignificance. Simultaneously
with the Byzantines was increasingly compromised by Turcoman g
activities across the western frontier when these tribes no longer re
check on their activities by the fading Seljuk government. Indeed, b
1300 AD the Seljuk state had in effect vanished.
The final stage in the process of Turkification of Asia Minor had b
around the middle of the thirteenth century by the formation of a
of Ghazi beyliks- states dominated by warlike elements of the type
the frontiers- Germian, Mentesche, Karaman, Osmanli, numerous an
conflict with one another at their several frontiers. Simultaneously the
a number of essentially pirate beyliks along the western and southe
Asia Minor-Sarukhan, Aydin, Karasi, Tekke, Sinope. It is interestin
late that these states may have transferred the concept of a no-man
the terrestrial borders to the seas adjacent to the peninsula of Asia
In any event, these developments were ominous from the point of
Byzantine holdings in Asia Minor, compromised once more when i
Latin rulers were expelled from Constantinople and the seat of
returned from Nicea in Asia Minor to Constantinople on the Europe
the Hellespont. With this move, the focus of interest of the Empire
to its European domain, accompanied by relative neglect of its aff
Minor. Simultaneously, the Lascarids were replaced by the Paleolog
that, in confronting its increasing fiscal difficulties, chose to do so b
on the old established privileges of its border troops in Asia Minor
what loyalty to the state these border warriors had retained- as wel
lishing the powerful Byzantine war fleet in 1284 AD. The result was
ritory gained by the Lascarids in the mid-thirteenth century was los
of that century all of Asia Minor once more came into Turkish hands
most part has remained there since.
The fourteenth century is marked by the rapid progress of the Ot
as the dominant factor in Asia Minor at the expense of the smaller
had arisen in the second half of the thirteenth century. The check
development at the hands of Timur-i Lang in the battle of Ankara i
the subsequent development of Ottoman domination in eastern Euro
fall of Constantinople to Sultan Mehmed II, all fall outside the perio
in the present context. From the middle of the thirteenth century
activity in the region was increasingly carried out by organized ar
close control by the rulers of the several states, and with this devel
kind of border no-man's land that concerns us receded into insign

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
SECTION 4
Summary and Conclusions
e have now concluded the presentation of what seemed

the essential factual material bearing on boundary conce
the medieval Islamic world. It will be useful at this poi
recapitulate briefly the course of the inquiry to this point:
The observation giving rise to this study was that indications of
boundaries are altogether absent from the greatest of the Muslim g
treatises, the twelfth century "Book of Roger" of the geographer
working at the court of the Norman ruler of Sicily. It was demonstr
in fact medieval Arabo-Islamic geographers, from the earliest scholar
at the beginning of the ninth century right through to the fourteenth
days of Ibn Khaldun, if they admitted the existence of political bou
all, did not conceive of the margins of adjoining individual states
borderlines. Regardless of whether they addressed boundaries form
states falling within Dar al-Islam or between Dar al-Islam as a whole
of Unbelievers, geographers described all such borders in terms im
boundary zones of significant depth surrounding a core area for a
political entity within which its capital was located. Transition zon
ated with external frontiers were shown to be occupied by a mixe
population differing in its composition from that of the core areas
states. A hypothesis was formulated to describe the conception of the t
of a state implied by these characteristics.
Along with this perception of boundaries as transition zones, the
of 'area' as a descriptor of geographic facts was shown to be absent
medieval Muslim literature dealing with geographical concepts.
The transition zone concept of political boundaries was not confine
scholars but rather corresponded to the experience of the people of
as reflected in travel accounts that imply that there was nothing to
traveler that he was approaching or crossing a political boundary;
and document control in general were experienced in towns, by no
always near any frontier.
Boundaries involving Muslim territory on one side and 'unbeliever
tory on the other were shown to have been characterized by the dev
of a system of defense in depth, consisting of primary frontier zon
and awasim. The tugir were shown to have encompassed fortress t
lesser fortresses in a wider defense zone under military administrat
to have been occupied by a distinct boundary population, differing f
the regular armies and the rural as well as urban population of the
states involved. The origins of the terms were traced and it was sho
reference to such structures is widespread in the geographical literat
time, but that in the works of any given author the concept seems to
given but slight emphasis.
65

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
All of these observations were based on data from Arabo-Iranian societies.

For comparison, a study was reported concerning the displacement of antique
hellenistic Byzantine communities in Asia Minor by Seljuk and Ottoman
Turkish ones during the conquest of Byzantine Asia Minor by the Seljuks and
their successors during the period from the eleventh to the fourteenth cen-
tury. Development of boundary zones even more extensive than in the Near
East or the Iberian Peninsula was encountered in this case, and was shown
to have been associated with the formation of a distinct "mixed border popu-
lation" made up of border fighters from both sides together with political or
religious refugees and a remnant of the original inhabitants. In the course of
the protracted warfare associated with these events the actual boundary swept
over a large part of the Anatolian peninsula, and thus by the time of the
Ottoman predominance in the peninsula a considerable fraction of the popu-
lation remaining in the region had been subjected to border conditions and
to the processes of assimilation to which they gave rise.
Closer examination of the actual progression of events showed that the
fighting involved, in addition to the organized armies of essentially sedentary
societies, irregular frontier warriors-ghazis on the Muslim side-still closely
linked to a nomadic way of life and more often than not fighting as autono-
mous units in raids uncorrelated with the campaigns of the Turkish princes
ruling the states or beyliks involved.
In retrospect, these characteristics of the conquest of Asia Minor by the Turks
were seen to find close parallels in earlier Muslim conquests, dominated by Bed-
ouin Arabs in the Middle East and by Berbers and Arabs on the Iberian Penin-
sula, and leading to the formation of tugur and awasim in those settings.
Thus, the available data suggest that zone boundaries surrounding the core
of a given state may be a phenomenon common to Muslim conquest societies.
It was shown finally that the concept of zone frontiers reflects basic reli-
gious as well as administrative and ethnic tenets of the culture of the medieval
Muslim world, summarized in Table 4.1.
It is now possible to recast these data to examine the principal ques-

tions we have answered as well as those that remain to be answered and
that may call for additional factual data.
The inquiry confirms the division of boundaries in the minds of the
Muslim scholars into two types: Internal (between two Muslim states)
and External (between a Muslim and a non-Muslim state) boundaries.
Examination of the dynastic history of the Islamic Empire reveals a pro-
gression from a single unitary Dar al-Islam to a system of at least 15
states, confirming the existence of numerous Internal boundaries within
the realm of Islam from the ninth century on despite the original reli-
gious emphasis on a single Umma.
The geographers' concept of zone type boundaries at these interfaces
was shown to correspond to the experiences of travelers. This concept
and the insignificant importance attached to political boundaries within
Dar al-Islam it implies would seem to have been of fundamental impor-
tance to the preservation of what may be called the medieval Islamic
"common trade zone."170
170 Richards, D.S., ed., Islam and the Trade of Asia, Oxford 1970, especially articles by
B. Spuler and A.L. Udovitch.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
SUMMARY AND CONCLUSIONS 67
TABLE 4.1. Factors and Attitudes Tending to Favor Vagueness of G

Boundaries in Medieval Islamic Society
Qur'an and Hadith
Command of brotherhood of all members of 'Umma
Mandate of permanent jihad against unbelievers
Concept that all land belongs to Allah
Geographers' and Administrators' Attitudes

Non-use of Area concept
Complexity of land tax structure
Concept of the state as a sum of its cities
Structure of society
Persistence of Bedouin concepts of rights flowing from land ownership
Predominance of cities (and city/kura complexes) in political and commercial life
Coexistence of sedentary and nomad elements in the population
Individuals
Absence of concept of citizenship
Acquired as well as category-determined loyalties to individuals only-not corporate
entities
This material confirms the impression that these scholars' writings

conformed closely to popularly held concepts of the characteristics of
medieval Muslim geography. In addition, we made the significant point
that geographers do not seem to have made any use of the concept of
area in their writings. Since that concept was thoroughly familiar to early
Arabo-Islamic mathematicians, this attitude would seem to represent
another example of poor communication between Muslim specialists in
the theoretical and applied sciences (elsewhere we have called attention
to a similar disparity between the understanding of Muslim geographers
and, in that case, navigators with regard to determination and signifi-
cance of latitude south of the equator).171
External frontiers of the type designated in the Arabo-Islamic litera-
ture tugur were shown to develop at all interfaces between a Muslim and
a non-Muslim dominated country and only there. They have been char-
acterized as zone type boundaries, and their geographic distribution as
well as the distribution of fortress towns and of lesser fortresses within
them have been described.
With these data we concluded that, in accordance with Ibn Khaldun's
dictum, medieval Muslim states were indeed conceived of as being sur-
rounded on all sides by boundary zones and hence lacked the sharpl
defined territory that would require border lines. Clearly, one could con
clude that such states cannot have been conceived of as territorial states
by the people of the time.
171 Brauer, R.W., "The Dynamics of Change in the Magnitude of Geographical Coordi-
nates in Three Civilizations," American Neptune 53 [4]: (1993).

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
TABLE 4.2. Comparison of Boundary Characteristics in Asia M

Near East
Arab conquests
Subject in Near East Turkification of Asia Minor
The warfare
time course of warfare Quick advance then sustained struggle, frontier
relatively stable shifting back and forth
frontier
non-Muslim opponent Byzantine, Sassanid Byzantine Empire, crusaders
Empires
nature of invading forces Arab tribesmen Seljuk professional army plus
under (more or turcoman and ghazi
less unified) warriors operating
command independently
tribal structure acephalous, conical, chief-oriented
segmental
numbers of conquerors small initially, modest initially, growing
never large continuously
The results
progress of islamization in rapid, but initially gradual, partly reflecting

zone limited exodus of Christians
character of boundary
continuous
zonestrip broken up and mosaic
no-man's
appearance of landscape land no-man's land
cities and fortifications present in thughur, inconsistent
defensive system
population of opposing side scarce; deliberately Akritai and associated border
of border depopulated fighters
blending of opposing border limited information, literature suggests consider-
populations some border able intercourse
elements crossed
back and forth
This configuration of the boundaries was shown to follow in large mea-

sure from certain religious mandates. It was suggested that these reflect
the essentially transcendental character of the Islamic state. Other evi-
dence, for example, the concept that such states reserved to themselves
a monopoly (or a near monopoly) of the use of force, modifies this view
and leads to the perception that in the medieval Muslim world the idea
of the state was a hybrid one in the sense of Watt.172
Comparison of events leading to the formation of boundary zones in
connection with the Arab conquests of the seventh and eighth century
and the Seljuk conquests in Asia Minor in the tenth to fourteenth
showed (Table 4.2) that, while boundary zones were formed in both
cases, the events leading to their formation as well as the configuration
of the structures actually developed showed certain differences as well
as similarities and the significance of these requires further analysis. On
172 Watt, W.M., Islam and the Integration of Society (Edinburgh, 1960), c.V.

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
SUMMARY AND CONCLUSIONS 69
the one hand the differences may reflect the different nature
fare giving rise to the boundaries in the two situations-the p
initial conquests of the Arabs as compared to the seesaw chara
Turko-Byzantine conflict. On the other hand, it may be wor
consider that the character of the conquests reflected underly
ences in the social structure of the tribal societies from which the t
quering forces were ultimately derived, the essentially aceph
mentally structured society of the Bedouins as contraste
stratified, chief-oriented society of the Central Asiatic Turks.
One can, I think, now give a succinct answer to the ques
which this inquiry began: while the almost total disregard of
that characterized al-Idrisi's work is extreme, examination of
of more than twenty Muslim geographers' writing over the p
820 to 1320 AD shows that boundaries were indeed given littl
or were ignored altogether in the literature. The core and bou
hypothesis for describing any state appears to represent the u
attitude adequately. Somewhat similar boundary concepts rec
larly in other civilizations but the factors giving rise to the conce
to differ markedly from one civilization to the next. As an ex
might note that the boundary zone concept characteristic of p
frontiers of Republican Rome was replaced by the idea of firm
defensive boundaries after Augustus' reign, and after D
reforms, boundaries reflecting in-depth defense replaced thes
interpretation of the social significance of the facts develope
Islamic world will have to be deferred until more extensive co
data are on hand.
173 Cf. Dyson, S.L., The Creation of the Roman Frontier (Princeton, 1985); Parker, S.
Romans and Saracens (Los Angeles, 1985); and Williams, S., Diocletian and the Roman Recove
(New York, 1985).

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
INDEX
[The transcription used to romanize Arabic terms throughout the tex

Standard International Alphabet as employed, for instance, in H. Wehr's
of Modern Written Arabic. In the following Index the first recognizable
defines the letter under which the word is classified. Words beginnin
' are shown at the end of the corresponding letter category-e.g. 'ala woul
at the end of the entries beginning with 'a'; kums after kura. ]
Abbasids, 8, 9, 56, 61 Baba Ishaq, 63

Abu Hanifa, 41 Baghdad, 9
Abu Yusuf, 41 bakt, 1
Aglabids, 9 Balkhi school, 3, 5
Aidhab, 34 bara'a, 35
Ajdabia near Barqa, 34 bedouin, 42
Aklat, 34 Benjamin of Tudela, 34
Akka, 34 Berbers, 23
Akritai, 57, 58 beyliks, 51
akfm, 12 Black Lands', 38
al-adna, al-awsat, al-a'la, 21 Book of Roger, 1, 21, 65
al-Andalus, 8, 9, 15, 16, 21, 22, 24, 25 border population, 53
al-Idrisi, 6, 12, 13, 21, 25, 65 boundary zones, 45
al-Idrisi', 1 Bulghar, 57
al-Istakhri, 12 Buwaihid, 9
al-Khwarizmi, 1, 3 Buyid, 18, 61
al-Ma'mfn's world map, 1, 3 Byzantine Empire, 15, 18, 25
al-Mas'udi, 34 Byzantine war fleet, 63
al-Mawardi, 31, 37, 40
al-Tur, 34 cadastral survey, 37
Aleppo, 34 capitatio (et) iugatio, 39
Almoravids, 9 certificate of aman, 33
Alp Arslan, 48, 61 citizenship, 32
Anatolia, 15, 48 climates, 6
Ankara, 63 collective assessment, 39
Antartus, 14 common trade zone, 66
Antiochia, 15, 18 communally owned lands, 43
apelatoi, 57, 58 concept of coexistence, 41
Arabo-Iranian bureaucracy, 47 conical tribal descent patterns, 47
Arabo-Islamic geography, 8 conquest of Constantinople, 62
Arabs of the Futuh, 60 Constantinople, 63
Area, 36 Constitutional implications, 30
area-based (misaha), 39 core-and-frontier-zone pattern, 61
Armenia, 47 Cosmographia, 3
ard al-andalus, 21 Countries of the Blacks, 34
ard castilia, 21 customs, 33, 35
Asia Minor, 45 collecting stations, 34, 35
Asuan, 34
Atlas Islamicus, 3 Danishmend, 55, 62
Aydin, 63 dar al-harb, 11
azala, 36 Dede Qorqut, 55
akir, 12 defense in depth, 18
'alids, 9 dervish preachers, 55
'Abbasids, 8, 9, 56, 61 desert and the sown, 60
'awasim, 12, 16, 18, 22, 25. See also tugur Diar Bakr, 34
71

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
72 INDEX
domains of sovereignty, 37 Karaman, 63

dynamics of the conquest, 18 Karasi, 63
dynasty of the Great Seljuks, 47 Khalifa, 9, 31, 32
Dar al-Islam, 40 Khalifate, 9, 38, 42
dimmis, 33, 39, 42 Konia, 62
Kose Dagh, 63
Egypt, 16 kura, 16
estimate of surface area, 36 Kwarezmian Empire, 61
External frontiers, 11 kwarezmshahs, 9
kamdan, 35
fadan, 36 kans, 21
kums, 39
farj, 12
kurasan, 16
Fatimids, 9
kutba, 9
forced depopulation, 18
frontier function, 58
frontier zone, 14 Lamis island, 35
land tax, 37, 39, 42
Landed property, 41
Georgia, 47 landmarks, 43
Germian, 63
Lascarids, 51, 62
Ghazi beyliks, 63 Line boundaries, 36
gulam (ex-slave soldier), 61
Greek inhabitants, 55 linear concept of geography, 37
linear geography, 13
guerra fria', 'the cold war' 41
gazi, 55 Makran, 16
gazi state, 62 Mamluk times, 34
Ghaznavids, 9, 47, 61
Manbij, 34
gazua, 54 Manzikert, 47, 53, 61
Mao Tun, 60
hajj caravan, 34 Marca Hispanica, 21
Hamdani military authorities, 35 marca paganorum, 23
Hamdanids, 9
march population, 57
hellenized Asia Minor, 59 marches, 23
Heraclius, 18 Mecca, 34
hima, 43 Medinaceli, 21
Holy War, 11 medinat as-salam, 21
horses, 47 Mehmed II, 63
Hsiung-Nu, 60 Melik Danishmend, 55
hadd, 3, 6, 12, 13 Mentesche, 63
hasia, 12
Mesopotamia, 16
misaha, 38
Ibn Battuta, 33, 34 mixed border population, 55
Ibn Hauqal, 3, 12, 13, 14, 34, 39 mixed border society, 62
Ibn Jubair, 18, 33, 34, 35 Mongol advances, 63
Ibn Khaldun, 26, 60, 65 Mongol pressure, 51
Ibn Khordadba, 18 Mongols, 56
Ibn Sa'id, 6 mukus, 33
Idrisids, 9 Muruj ad-dahab, 34
Imam, 31 Myriokephalon, 62
Intermarriage, 58
Internal frontiers, 8 Nasir Khusrau, 34
iqta'a, 43 Nicea, 62, 63
Ikmim, 34 no-man's land, 51, 59
Ikshidids, 9 nomads of the Arabian Peninsula, 43
Nuba, 1
jarib, 36
jarajima, 21 Oghuzz, 47
Jebel Lubnan, 18 Oghuzz tribes, 61
jihad, 14 Osmanli, 63
Jidda, 34 Ottoman state, 63
jihad, 41
jizia, 21, 33, 35, 38, 39 Paleologi, 57, 63

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00
INDEX 73
pirate beyliks, 63 Tahirids, 9

places of refuge, 58 tax, 37(especially n. 89), 38, 39, 40. See also jizia
power, radiation of, 42 tax reform / Diocletian, 39
primary function of the state, 41 Tekke, 63
Prince Juan Manuel, 41 terrae desertae, 21, 24
principle of reciprocity, 41 territorial states, 41
property rights, 43 Timur-i Lang, 63
Ptolemy, 3 Toghril ibn Mohammad, 47
Toghrul, 61
transfer of land from dimmis to Muslims, 38
qabiz, 36 transoxania, 61
Qaisites, 21 tribal lands, 60
Qaraqanid, 61 Tripoli/east, 34
Qarluq, 61 Tulunids, 9
qa'id, 16 Turcoman raiding, 48
Qipchaqs, 47 Turcomans, 54, 61
Qudama, 18
Turkic empire, 61
qusama, 38 Turks, 42, 60
Qusair, 34 Turner, 58
two-humped camels, 47
ribat, 14 tagr, 14, 21, 25
right of first occupation, 43 tugur, 12; defined, 14, 15, 25; as zone, 16;
right of usufruct, 42 configuration, 22, 23; geographical term,
'rightly guided' khalifs, 8 25, 26; military importance, 16, 24; popu-
Rihla, 34, 35 lation, 22, 23; towns, 19
river Oxus, 15 _tugr as-samiah, 15
robber bands, 35 tugur al-adn', 15
Romanus, 47 tugur al-awsat, 15
_tugur al-a'la, 15
tugur al-jaziria, 15
Safar nam-e, 34 tugur at-turuk, 15
Saffarids, 9
sahm, 36 uC, 12
Samanids, 9 uc-turks, turks, 54
Saragossa, 21 Uighur, 61
Sarukhan, 63 Umayyads, 8, 21
Sawad, 37 urban center(s), 42
Sayyid Battal, 55 'umala', 8
Sea frontiers, 33 umma, 8, 40
segmental society, 43 urban emphasis, 42
Seljuk, 9, 47, 48, 51, 54, 60, 61, 62 'usr, 33, 39
Seljuk Empire, 47, 53, 54, 61 'Utman ibn 'affan, 38
Seljuks of Rum, 9, 48, 53, 54, 60, 62, 63
Sinope, 63 village of Qatia, 34
size of irrigation canals, 36 Visigothic populations, 23
slave professional armies, 48
Slavic elements, 57 water supply, 38
soil ownership, 38 white (i.e., waste) lands, the 'terrae desertae',
sovereignty, 28-30 24
Spuler, 36
state, defined, 30; function, 41 yarad, 36
sultan, 31, 47; conflict with imam, 31, 32 Yuan Yuan, 60
sultan, 32
Syria, 16 zone, 16, 18, 21, 25, 28, 30. See also 'awasim,
sarrat, 21 tuigir
sari'a, 31 zone type of external boundary, 59
sarl'a land law, 43 Zott, 21

156.35.225.71 on Wed, 24 Apr 2024 09:32:06 +00:00

Brauer Boundaries Frontiers Medieval

Uploaded by

Copyright:

Available Formats

Brauer Boundaries Frontiers Medieval

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Brauer Boundaries Frontiers Medieval

Uploaded by

Copyright:

Available Formats

Boundaries and Frontiers in Medieval Muslim Geography

Author(s): Ralph W. Brauer

Stable URL: https://www.jstor.org/stable/1006658

This content downloaded from

Medieval Muslim Geography

This content downloaded from

American Philosophical Society

VOLUME 85, Part 6, 1995

Boundaries and Frontiers in

Institute for Research on the

THE AMERICAN PHILOSOPHICAL SOCIETY

Independence Square, Philadelphia

This content downloaded from

This work was supported in part by grants from

Library of Congress Catalog

This content downloaded from

The Geographical Concepts .................................. 1

Travelers' Experiences at Internal Boundaries, the Area

Genesis of Boundary Zones Involving non-Arab Muslim

Summary and Conclusions .................................. 65

This content downloaded from

The Geographical Concepts

Boundaries in Arabo-Islamic Cartography

The earliest known map by Muslim scholars, al-Ma'mun's world map

1 State and territoriality concept-Lambton.

This content downloaded from

FIGURE 1.1. Map of the course of the Nile of Egypt

This content downloaded from

This content downloaded from

This content downloaded from

This content downloaded from

identity, and a peripheral zone where the force of sovereignty

which may be transcribed as

placed across a region corresponding to

9 al-Idrisi, al-aarif, Kitab nuzha 'I mustaq ft-'ktir

This content downloaded from

are represented only by simple geometric lines having symbolic rather

Boundaries in the Arabo-Islamic Geographic and Historical Texts

This content downloaded from

free from external threat to the Islamic heartlands, and-de

Internal frontiers involved two adjoining Muslim states that may or

14 Vatikiotis, P.J., Islam and the State (London, 1987), 21.

This content downloaded from

15 Short terms of amirs under Umayyads.

This content downloaded from

This content downloaded from

0 200 400 600

FIGURE 1.5b. Number of coexistin

External frontiers occurred at the periphery of the empire wherever a

This content downloaded from

three (akir, takum, has'a) imply a view of the relationship stri

Hadd versus tugfir. The term Hadd is used by the geographers

5. hadd = ....... primary meaning: cutting edge of knife of sword; t

This content downloaded from

boundary, and, albeit merely in very general terms-by referen

This content downloaded from

haps by a customs station; as will be shown in Section 2 this is

This content downloaded from

TABLE 1.1. Ethnic Factors in Boundary Structure in the Medieval M

" Franks and Visigoths + +