Photocopy and Use Authorization
In presenting this thesis in partial fulfillment of the requirements for an advanced
degree at Idaho State University, I agree that the Library shall make it freely available for
inspection. I further state that permission for extensive copying of my thesis for scholarly
purposes may be granted by the Dean of the Graduate School, Dean of my academic
division, or by the University Librarian. It is understood that any copying or publication
of this thesis for financial gain shall not be allowed without my written permission.
Signature ___________________________________
Date _______________________________________
MAPPING SIXTEENTH CENTURY SPANISH TRANSPORTATION ROUTES:
A GIS APPROACH
By
Robert E. Hibberd
A thesis
submitted in partial fulfillment
of the requirements for the degree of
Master of Arts in Historical Resource Management
in the Department of History
Idaho State University
July 2010
Committee Approval
To the Graduate Faculty:
The members of the committee appointed to examine the thesis of ROBERT E.
HIBBERD find it satisfactory and recommend that it be accepted.
_________________________________
J. B. Owens,
Major Advisor
_________________________________
Sarah E. Hinman,
Committee Member
_________________________________
Mark McBeth,
Graduate Faculty Representative
i
Contents
Acknowledgements……………………………………………………………………....iii
List of Figures…………………………………………………………………………….iv
List of Tables…………………………………………………………………………..….v
Abstract…………………………………………………………………………………...vi
Dedication………………………………………………………………………………..vii
Chapter 1: Introduction…………..………………………………………………..………1
Chapter 2: Reconstructing Iberian Transportation Networks across Five Centuries
of Place-Name Changes………………………………...……….……....…...18
Chapter 3: Building a Digital Historical Gazetteer, Locating Missing Places,
and Visualizing Population Centers ……..................................................22
Chapter 4: Methods of Building a Historical Gazetteer, Analyzing Missing Places,
and Adding Further Historical Contextual Data……………………………..51
Chapter 5: Discussion of Results ……….……………………………………………….77
Chapter 6: Conclusion………………………………………………………………..…121
References………………………………………………………………………………129
Appendix A: Glossary…………………………………………………………………..137
Appendix B: Place-name Tables for Toledo and Cuenca Provinces…………………...140
ii
Acknowledgements:
I wish to thank God and those who helped me write this thesis; in particular, Dr.
J.B. “Jack” Owens, the ‘active voice’ for this work, and the inspiration for many of its
ideas, as well as the needed funding. His vision and guidance were essential to the
success of this project, as was his ready knowledge of important sources of data and
scholarship. I greatly enjoyed working with him, and I appreciate what he has done for
my family and me. I also wish to thank Dr. Sarah Hinman, who provided critical
guidance on many issues related to GIS and statistical analysis, the structure of the paper,
and the literature review; Dr. Barbara Stephenson, or the “Purple Peril” (known as such
for her purple editing pen), who provided excellent guidance on the project’s copy editing
and structure, helping me create a more readable text from my conglomeration of loosely
connected bits of scribbling; Trista Hibberd, our honorary GIS Specialist and my patient
wife, who provided needed inspiration on the introduction; Robert Beazer, whose
knowledge of information systems provided needed assistance; and Keith Weber, among
others, with all of whom I have immensely enjoyed working.
iii
List of Figures:
Figure 2.1 Title Page and Beginning Page of ‘Cuenca to Burgos’
Route, Juan P. Villuga……………………………………………………….19
Figure 3.1 Ottawa Map: Cram’s……………………………………………….………...23
Figure 3.2 Villuga Gazetteer Database Model……………………………….…………..25
Figure 5.1 Map of Juan Villuga’s Routes, from Modern Sources……………….....……79
Figure 5.2 Map of Confidence Levels…………………………………….……………..80
Figure 5.3 Map of League Distance Ranges……………………………….…………….89
Figure 5.4 Interpolation Analysis of an Unknown Place’s Location...............................91
Figure 5.6 Castilian Population Centers in 1591, along Villuga’s Routes……………..100
Figure 5.7 Kernel Density of Castilian Population Centers of 1591, along Villuga’s
Routes………………………………………………………………………101
iv
List of Tables:
Table 3.1 Journeys of Barreiros and Villuga Compared…………………………………30
Table 3.3 Mauro and Parker’s Study of Spain’s Population…………………….……….43
Table 4.1 Kilometers per League…………………………………………….…………..62
Table 5.1 Modern and Historic Names from Villuga Gazetteer………………….….…..81
Table 5.2 Density of a Sample of Place-names for Cuenca Province…………………...83
Table 5.3 Average Nearest Neighbor Analysis of League Distance Ranges….…………87
v
Abstract:
Historical GIS provides cognition-enhancing tools for exploring human data. A
traveler’s handbook from 1546 supplies a transportation network data infrastructure for
Spain, producing a spatial context for its political, economic, and religious history. This
handbook lists 139 routes, with distances from one place to another. A gazetteer of
modern counterparts of the historical place-names forms the basis of a GIS of Iberia’s
sixteenth-century routes. Various other narrative sources list variants of the historical
place-names, allowing a comparison across time and the various references. Employing
GIS statistical analyses determines spatial variations in the measurement of leagues. A
reconstruction of the population along the routes in the Castilian provinces produces a
picture of urban life which gradually became increasingly centralized from 1561.
Historical narratives highlight polycentric aspects of the routes and the places contained
in them, and the economic, political, and religious polycentrism that arose partly because
of this urban structure.
vi
Dedication:
To Trista, Emily, and Nathan: There is beauty all around.
vii
CHAPTER 1: INTRODUCTION: HISTORICAL GIS
The complexity of the historical narrative at times dictates the use of tools that
increase a person’s ability to process data. Maps have historically provided one form
of augmentation to the historian’s data-processing capabilities. The onset of digital
technologies has provided further increases in the capability of humans to organize
and process large amounts of data. In the Geographically-Integrated History
Laboratory at Idaho State University, 1 Geographic information systems (GIS) and
other data management and visualization tools provide other essential augmentations
to human cognitive capacities for analyzing history.2 GIS provides a mapping tool that
visualizes digital copies of historical maps. Real-world geographic coordinates, such
as longitude and latitude, are assigned in GIS to each location on a map. Gregory and
Healey pointed out that beyond the ability of GIS to map geographical features, it is a
database that stores and allows queries on both spatial data, which is data based on a
location, and attribute data, or descriptive data assigned to an object such as a person.
It is also useful as a platform for spatial, statistical, and cartographic analysis. They
explained the other advantages of GIS in historical studies, such as the capacity to
combine by spatial location sources of historical information that vary widely in
1
The U.S. National Science Foundation (NSF) funded this work through a grant awarded to Dr. J.B.
Owens (Award Number SES-0740345; $394,000; 2007-2010) for the project entitled D a i
Complexity of Self-Organizing Cooperation-Based Co
e ial Net o ks i the Fi st Glo al Age
(acronym: DynCoopNet). This p oje t fo s pa t of the Eu opea “ ie e Fou datio ’s EUROCORE“
Eu opea Colla o ati e Resea h “ he e p og a The E olutio of Coope atio a d T adi g
(TECT).
2
For a discussion of geographically-i teg ated histo , see J.B. O e s, To a d a geographicallyi teg ated, o e ted o ld histo : E plo i g geog aphi i fo atio s ste s GI“ . History
Compass, 5, 6 (October 2007): 2014-2040.
1
format and purpose.3 Some examples of these varying data sources consist of narrative
stories and tabular population data. However, as Knowles pointed out, GIS in
historical research, known generally as Historical GIS, does not yet have a wide
application, although the number of scholars using it has grown quickly over the past
decade.4 Bodenhamer explained, historians usually prefer manuscripts over databases
and other computer technologies. The computer, he argued, with its inability to work
with imprecise or ill-defined data does not fit well into the historian’s worldview.
Historical sources do not often contain precise data. They often contain ambiguity,
while computers rely on precisely-defined data. However, he argued, historians can
embrace GIS because 1) the historian is familiar with mapping information, which GIS
does well, and 2) it provides a visual integration of complex data, allowing the
historian, in some cases, to see the complex historical narrative more clearly. 5 Gregory
and Ell explained further benefits of using GIS in history. The ability to integrate
space into an analysis comes in two formats, a ‘horizontal’ flow of people and ideas
across space, and the more detailed ‘vertical’ interactions of people in one place. They
pointed out that a place is made up of many smaller places that all fit together in
critical interrelationships. GIS provides the historian with tools for studying the more
3
Ia N. G ego a d Ri ha d G. Heale , Historical GIS: structuring, mapping and analysing geographies
of the past. Progress in Human Geography 31(5) (2007), 638.
4
Anne Kelly Knowles, ed. Placing History: How Maps, Spatial Data, and GIS are Changing Historical
Scholarship.(Redlands, CA: ESRI Press, 2008), xiii.
5
Da id J. Bode ha e , Histo a d GI“: I pli atio s fo the Dis ipli e, I A e Kell K o les, ed.
Placing History: How Maps, Spatial Data, and GIS are Changing Historical Scholarship.(Redlands, CA:
ESRI Press, 2008), 220-222.
2
detailed, finer scales of human action and interaction (the vertical places) along with
the geographically wider scales (the horizontal spaces). 6 GIS provides the ability to
address some data scaling issue by moving between the large and small scales of a
study.7
Gregory and Ell explained the role of GIS in historical research. They stated
that some risks exist in combining history and historical geography with GIS, because
GIS originated in scientific disciplines that have focused on quantitative approaches to
scholarship. Historical geographers, they argued, have rightly avoided too great an
emphasis on quantitative methodologies. Historians and historical geographers can
benefit from using GIS, however, if they focus on answering the question, “what are
the geographical aspects of my research question?” Methods exist for appropriately
using GIS in historical studies.8 Authors such as Gregory, Ell, and Knowles, in their
works discussed above, provide guidance on these methods.
Knowles has written an excellent coverage of some of the research historians
have done in geography and historical GIS. Among the scholarship she listed, she
included historian Fernand Braudel’s concept of geohistory, which worked to
contextualize humans in their geographic locations. Some recent projects have covered
a wide variety of topics. One project, the Salem Witch Trials Archive, combined the
6
Ian N. Gregory and Paul S. Ell, Historical GIS: Technologies, Methodologies and Scholarship. (New
York: Cambridge University Press, 2007), 5.
7
A
8
Ibid, 1-2.
e Kell K o les, GI“ a d Histo ,
.
3
spatial locations of witches’ homes with court documents and biographical
information about each accused witch in order to contest the prevailing understanding
of the trials among historians.9
This thesis will further demonstrate the contributions GIS can make to the
study of history. It has reconstructed a critical portion of the sixteenth-century trade
network of the Spanish Monarchy, which in that century had a global reach. This trade
network, with population data to further contextualize it, will provide a spatial context
for the historical events of the sixteenth century, to answer the question of how the
spatial constraints upon Iberia’s human population helped to shape its history.
Historian John B. Owens has created an international grant project with a
multidisciplinary scope, analyzing the motives for cooperation among actors in the
global trade networks of the First Global Age (1400-1800), who as rivals should have
demonstrated far more competitive interactions. It focuses on how individuals in the
First Global Age made human trading networks function without a central governing
authority or body of governing rules. These human interactions functioned in a
concealed environment to avoid regulation and taxation by governments, and therefore
could not rely upon whatever weak forms of government existed. The grant project,
funded by the European Science Foundation, the United States National Science
Foundation, and several similar foundations in various sponsor countries, is entitled
“Dynamic Complexity of Self-Organizing Cooperation-Based Commercial Networks
in the First Global Age” (acronym: DynCoopNet). It brings together historians,
9
A
e Kell K o les, GI“ a d Histo ,
-7.
4
GIScientists, and mathematicians in an effort to model those activities. The project
relies upon the paradigm of Geographically-Integrated History, which posits that “the
understanding of historical processes requires an integration of the natural, social, and
cultural environments on the basis of place, space, and time and accomplishing this
integration poses a challenge, which can be met with modern information
management, especially geographic information systems (GIS), and visualization
techniques.”10 The paradigm of Geographically-Integrated History proposes that 1)
“the history of any place is shaped by the way the place is connected to other places
and by the changes in these connections over time,” 2) historical periods consist of
complex, dynamic systems that do not follow a linear path, and disruptions in the
system may cause it to evolve into a new system with different organization and rules,
and 3) within the system of a historical period, people organize themselves into social
networks, creating the source of innovation in the system. Nonlinear dynamic systems
and self-organizing commercial networks comprise major parts of the project’s focus.
Historical dynamics, Owens argued, do not rest upon a long-term chain of events.
Rather, the connections between each place constitute a global complex system of
human interaction, and events occur in every place due to changes in the overall
system, which may develop quickly, similarly to the changes in a weather system.11
10
J.B. O e s, Usi g GI“ to E plo e the No li ea D a i s of Histo i “ ste s, Pape to e
presented at a meeting of the Royal Geographic Society, in London, U.K., September 2010), 2.
11
Ibid, 6-9.
5
The DynCoopNet project aims to produce narratives based on the construction of
social networks and their nonlinear, dynamic, changing interactions. 12
Nordlund argued for a global network analysis of trade. He asked why peoples’
consumption of the world’s wealth differed so greatly across space, and why fifteen
percent of the world’s population consumes fifty-six percent of the global income at
the present time. While the standard argument for the results of Europe’s colonialism,
based in classical economic theories, claims that Europe gained its economic power on
the backs of its colonies, other studies have found a negative correlation between the
build-up of Europe’s colonialism and its economic growth. Nordlund argued that
while some transfers of goods occurred between Europe and its colonies, trade has
always provided the greatest percentage of the transfer of goods around the world.
Wallerstein’s world-system perspective, which Nordlund adapted to his study, argues
for a holistic and interdisciplinary analysis of the global economy, acknowledging it as
dynamic and interconnected. Nordlund argued that the structures of a social network
had a significant impact upon trade relationships, and called for studies that integrate
network analysis with traditional economic studies. He analyzed global trade networks
of two commodities, searching for the degree to which the networks’ structures
influence the amount of benefit the trade brought to the various members of the
12
J.B. O e s, A Multi-national, Multi-disciplinary Study of Trade Networks and the Domains of
Iberian Monarchies during the First Global Age, 1400-1800, I SSPHS Bulletin, 33-2 (2008), 23-30.
http://www.ucmo.edu/asphs/fall2008/multi.html; J.B. O e s, D a i Co ple it of Coope ationBased Self-O ga izi g Co
e ial Net o ks i the Fi st Glo al Age D CoopNet : What’s i a a e?
In David Alonso García and Ana Crespo Solana, eds., Self-organizing Networks and Trading
Cooperation: GIS Tools in the Visualization of the Atlantic Economic Expansion (1400-1800),
(forthcoming).
6
network, and how ecological structures contrasted with these economic networks.13
Andre Gunder Frank proposed a globally-contextualized history that followed a
holistic system based on a “three-legged stool,” which consisted of “(1) ecological,
economic, and technological aspects; (2) political/military interactions; and (3) social,
cultural, and ideological issues” within a global analysis of history. 14 He argued that
European history should focus on how the world created Europe, rather than the other
way around.15 Owens argued that GIS provides a tool that can integrate the three legs
of the stool to produce a more accurate picture of the global economy than classical
economics has done.16 DynCoopNet builds its use of GIS upon this perspective.
Norlund explained a new economic research theory, ecological economics,
which combines world-systems theory with a focus on human ecology, thus
integrating natural resources into the discussion of historical processes.17 The
proponents of this approach acknowledge the need for a holistic approach to history,
economics, and other fields. Scholars are acknowledging the need to cross disciplines
in order to better understand the world. The La Pietra Report of the Organization of
13
Carl Nordlund, Social Ecography International trade, network analysis, and an Emmanuelian
conceptualization of ecological unequal exchange,” (Ph.D. diss., Lund University, 2010), 1-14.
14
Andre Gunder Frank, ReORIENT: Global Economy in the Asian Age (Los Angeles: University of
California Press, 1998), 340-1; see also J.B. Owens, "Toward a geographically-integrated, connected
world history: Employing geographic information systems (GIS)." History Compass, 5, 6 (October 2007):
2020-2022.
15
Frank, ReORIENT, 3.
16
J.B. Owens, "Toward a geographically-integ ated, o
17
Nordlund, Social Ecography, 14.
7
e ted o ld histo ,
-2022.
American Historians (OAH) called upon the discipline of history to seek out a more
internationally-contextualized approach to history teaching, also calling for a renewal
of the discipline of geography, which will provide history with essential
interdisciplinary collaborators to the global approach. It claimed that any history of the
United States, for example, without placing it in a global context, would provide far
less than a history of the U.S. It calls for a refocus of the history of a country as but
one of the many scales on which the study of history should center itself.18 More
frequently, then, scholars have been calling for a scalable approach to history and
other social sciences, to take in a deeper look at one place, its people, and its natural
resources, while also expanding the context of that place to include its global
connections. To re-emphasize, GIS provides a valuable tool in this global and
interdisciplinary approach to history, as well as a useful theoretical approach. This
study provides some tools towards this effort. Further collaborative work among
historians with scholars from other disciplines can utilize the spatial data created for
this study of Spanish transportation networks.
Pérez explained that the major GIS-related goals of DynCoopNet include
integrating dispersed databases lacking a standard design, spatial analysis of the trade
networks to determine such questions as “which networks were most efficient?,” and
“which agents were most active?” The overall purpose of the grant project’s use of
GIS rests on providing a greater understanding of history through a more complete
18
Organization of American Historians, LaPietra Report: A Report to the Profession, 2000,
http://www.oah.org/activities/lapietra/final.html (accessed on July 3, 2010).
8
perspective on the dynamic processes at work along these networks. 19 This study
provides an important part of the spatial data needs of DynCoopNet by reconstructing
the Iberian Peninsula’s trade network and part of its population, which rests at the
heart of the project’s goals. A gazetteer provides its main tool.
Gazetteers provide an intermediate connection format between the historical
record and a GIS, by assigning a geographic location, in the form of map coordinates,
to place-names from the historical record, providing a spatial context for an event.
Place-name changes over time present researchers using a historical GIS, or
geographically-integrated, approach with one of its major challenges. Matching a
modern place-name to past place-names that were used to describe the same
geographic location can produce varying levels of confidence in the matches,
particularly when languages or governments for the location change over time. Spain
offers an excellent case study for gazetteer design for the purpose of connecting placenames over such temporal changes. The Spanish historical record provides extensive
documentation of significant place-name changes enacted by central and regional
governments in the nineteenth and twentieth centuries. Juan Pedro Villuga in the year
1546 created a listing of hundreds of Iberian place-names prior to any of these
centralized changes. Villuga’s Reportorio de todos los caminos de España (Medina
del Campo, 1546), of “all the roads of Spain,” as the Iberian Peninsula was then
19
Esther Pé ez Ase sio, et al., I t odu ió : P o e to D CoopNet, p ese ted at the Reunión
Nacional de la IDE de España, Madrid, 24-25 de febrero de 2010.
9
known, provided vast amounts of data about the Iberian Peninsula, 20 including
sixteenth-century travel routes, with the names of the places that formed the routes, but
in a format difficult for a person to process.
In the year 1834, the Spanish Liberals attempted to diminish the problems
caused by the geographical divisions with which the various Iberian languages roughly
coincided. In 1834, a royal decree introduced new partidos judiciales (administrative
areas) which made significant changes to prior place-names.21 In many cases it added
a descriptive word to names that repeated in several other locations, in order to create
unique place-name labels for each place. The local landscape served as the basis of
many changes, encapsulating identity in these markers. The new political boundaries
also influenced place-name changes. For example, in the Kingdom of Toledo, al
monacid became Almonacid de Toledo. Two places known as la calçada became La
Calzada de Béjar and La Calzada de Oropesa, to distinguish them from each other.
The Spanish Constitution of 1978 permitted the formation of autonomous
communities. The State of Autonomous Communities, part of the 1978 Constitution,
reassembled the earlier provinces. Catalonia, the Basque Country, and Galicia, all
became “historic nationalities.” The regions outside this group became “autonomous
communities.” The restructuring of administrative units gave each autonomous
20
Medina del Campo: Pedro de Castro for Juan de Espinosa, 1546.
21
Instituto Nacional de Administración Pública. Subdivisión en Partidos Judiciales de la Nueva
Subdivisión Territorial de la Península e Islas Adyacentes, 1834. (Madrid: Instituto Nacional de
Administración Pública, 2000), xli.
10
community and historic nationality a certain degree of autonomous home rule.22 Some
of the autonomous communities then dictated that place-names in their jurisdiction
should display on maps and in records in the local native language, thus significantly
changing the place-names from their form in Villuga’s record.
This study provides a connection between sixteenth-century place-names and
their modern versions. It also creates connections across the considerable amount of
place-name changes of the nineteenth and twentieth centuries for modern Cuenca and
Toledo provinces, according to the limitations of the historical sources. Utilizing
Villuga’s handbook will provide the historian an important tool for understanding
sixteenth-century Iberian commercial and pilgrimage transportation routes. Moreover,
no extensive study of Iberian place-names exists to connect the places of the Peninsula
across their temporal changes. GIS provides an excellent tool for filling that gap. This
study begins that effort.
Spatial data has gained great importance in many fields of interest.
Geoscientists and historians alike have begun serious efforts to build up and expand
available spatial data. Historical GIS projects in several countries, such as the United
States and Great Britain, have begun compiling extensive spatial data about their
countries, included such efforts as mapping all of the changes in administrative
boundaries and place-names over the span of their traceable history. They have also
worked on adding population, county, and road infrastructure data, as well as data for
22
José Alvarez Junco & Adrian Shubert, Spanish History Since 1808. (New York: Oxford University Press,
Inc., 2000), 322-324.
11
other themes important to policy makers, historians, and other scholars.23 The
Infrastructure for Spatial Information in Europe (INSPIRE) seeks to unite efforts in
Europe for the creation, standardization, interoperability, and dissemination of spatial
data throughout the continent. Hundreds of researchers will pursue the project’s goals,
and thousands of organizations have registered to participate.24 Spatial data creation
costs extensive funds and time, and efforts in this direction produce valuable
resources. This study will provide a starting point for the creation of Iberian spatial
data from the sixteenth century.
This spatial data will also aid in researching other questions about Iberia.
Historians have debated the cause of the Castilian Crown’s economic difficulties in the
sixteenth century and the collapse of its economic growth during the seventeenth
century. Ringrose argued that the rise of Madrid as the location of the Castilian Royal
Court in 1561, and its subsequent demand for goods, greatly taxed the Castilian
transportation infrastructure, which already suffered from limitations in available
animals and human operators, as well as other areas of needed improvement. This
23
See the websites for these projects. For the United States, go to, http://nhgis.org/. For data files
free for download, such as county boundaries, go to http://www.nhgis.org/mapping. For Great Britain,
go to http://www.port.ac.uk/research/gbhgis/. This data is also a aila le at the site fo A Visio of
B itai Th ough Ti e, at www.visionofbritain.org.uk.
24
Ma C aglia, Buildi g IN“PIRE: The “patial Data I f ast u tu e fo Eu ope I ArcNews Summer
2010. Redlands: ESRI Press, 2010. http://www.esri.com/news/arcnews/spring10articles/buildinginspire.html (accessed June 22, 2010).
12
transportation drain from Madrid precluded the buildup of essential industry and
expansion of market activity in the Iberian Peninsula.25
Other scholars have argued that the Crown’s economic collapse resulted not
from transportation problems, but from the impact of the Crown’s deficit financing on
the availability of investment capital, or other factors. López García argued against
Ringrose’s thesis. He proposed instead that the restrictions on trade imposed by
Castile’s seigneurial lords limited the geographic scope of Castile’s economy,
precluding its growth into one with a more international significance. 26 Larraz argued
that the influx of gold and especially silver from the New World mines in the Castilian
Crown’s possession brought about serious price inflation in Iberia, which compounded
against the lower prices of the rest of Europe, to which the Habsburg monarchs in
Castile had important ties. Europe’s prices remained lower without an influx of silver
similar to that in Castile. Certain Castilian scholars of the time proposed various
remedies, such as limiting Iberian exports while raising imports from Europe, or
investing in industrialization. Larraz argued that the Crown failed to realize these
approaches. Commerce from Sevilla to the New World colonies presented another
issue to the Castilian Crown. The influx of New World trade accompanied the silver,
and in its attempts to protect the trade between Iberia and its overseas colonies, the
25
David Ringrose, Transportation and Stagnation, i-x.
26
J. M. López García (dir.), El impacto de la Corte en Castilla: Madrid y su territorio en la época
moderna. (Madrid, Spain: EUROCIT, 1998), xv-xvii.
13
Crown also negatively influenced its own part of that trade.27 Calabria argued that the
cost of Castilian empire in Naples and elsewhere outstripped even the influx of money
from Castile’s New World colonies. Plagues, population decline, and famine also took
their toll on the economy. The rises and falls in international demand for goods also
hurt those areas that focused their economies on exports. Naples, situated south of the
Crown’s Lombardy frontier, bore its own weight in taxation as well as Lombardy’s.
The Crown could not envisage causing its critical military forces in Lombardy
economic hardship by taxing them, so Naples bore a double burden. The fiscal burden
of the Crown’s ever-growing military commitments outstripped its growing tax base,
and the Crown turned to deficit spending backed by loans from “obliging but exacting
bankers.” These obligations, coupled with the tax burden on the various regions of the
Crown, precluded needed investment in the economy. Thus, Calabria argued, changes
in the nature of military conflict between the fifteenth and sixteenth centuries-and
overemphasis on war-changed the impact of war from a tool of royal expansion to one
of great burden upon the Castilian Crown. State securities with high return rates-as
much as thirteen to twenty percent on the investment, also intended to finance the
growing costs of empire-further burdened the Castilian Crown in Naples and
elsewhere. Calabria argued that with all of this debt forcing it to focus on gathering
27
José Larraz, La época del mercantilismo en Castilla 1500-1700 (Madrid: Asociación Española de
Historia Moderna, 1963), xix-xxiii.
14
more and more funds, the Crown began a struggle against the economies under its
authority.28
Each of these arguments may contain truth, all contributing to the picture of
Castile’s rise and decline. A synthesis of all of these aspects of the story, and others,
such as how the New World trade affected Iberian prices, however, requires a
tremendous cognitive effort using traditional historical methods alone. Phillips pointed
out that many recent economic studies have expanded knowledge on this subject, but
they have tended to a small-scale analysis, while the problem requires a broader
approach.29 Historical GIS and geographically-integrated history provide ready tools
for the needed synthesis and broader scope. This study combines the information
supplied by its mapping of Villuga’s place-names and routes with sixteenth-century
population data to provide a crucial context for the debate over the Castilian economy.
This study represents an attempt to begin the pursuit of a geographic infrastructure,
analyzed through the lens of geography theories such as Central Place Theory and
Polycentric Urban Regions, which can aid the spatial contextualization of historic data
that focuses on the sixteenth-century economic, political, and religious life of Iberia. It
relies upon historic narratives to demonstrate the applicability of these geography
theories to the time and place under study. A study of economic growth and decline in
sixteenth-century and seventeenth-century Castile will benefit from this
28
Antonio Calabria, The Cost of Empire: The Finances of the Kingdom of Naples in the Time of Spanish
Rule (New York: Cambridge University Press, 1991), 1-6.
29
Ca la Rah Phillips, Time and Duration: A Model for the Economy of Early Modern Spain, The
American Historical Review, 92 (1987), 532-533.
15
contextualization. Scholars of the Castilian Crown can build upon and expand the base
this study provides, producing similar contextualization in Naples, Lombardy, and
other regions of the Crown. This contextualization will produce a more
geographically-integrated, connected history of the Castilian Crown, providing a
historical view wider and of greater scale than previously possible, while not
neglecting the finer scales of the historical picture.30
This study also discusses the vagueness of the Spanish league, which Villuga
used to record distances between each stopping place along his routes. The study
addresses the imprecise nature of the league by producing ranges of kilometers-perleague distance. Performing a spatial analysis of these ranges will determine whether
the places along each of Villuga’s routes utilized these distances ranges in patterns that
clustered together on the map. This paper then discusses fuzzy logic, a method of
computer processing of “humanistic data,”31 or data produced by humans. It also
produces a new application of fuzzy logic to GIS, by utilizing fuzzy rule-based
modeling, a process of fuzzy logic, and adapting it to GIS conditional statements.
This thesis then expands upon the usefulness of Villuga’s guide to historians
by adding population data from the late sixteenth century to each place-name for
which data are available. Through visualization of the population patterns along major
roads, historians will gain a greater understanding of why Iberia’s transportation routes
30
Owens, Toward a geographically-integrated, connected world history, 1.
31
Zadeh, Outline of a New Approach, 28.
16
formed their particular structures and interconnections. This analysis demonstrates that
the sixteenth-century Iberian urban structure resembled a polycentric urban region,
which Dobbs defined as “a system of politically independent units which are
functionally interdependent, but which have become something more than the sum of
their individual parts.”32 In other words, a group of equally-dominant towns, in terms
of political and/or economic power, or human population, that works together. This
study discusses some of the major works on the subject of Central Place Theory, a
geographic explanation for the distribution of human population centers, and
Polycentric Urban Regions, and applies them to Villuga’s Castile, based on population
data. It also provides examples of Iberian political, economic, and religious
polycentrism during the sixteenth century, utilizing historic narratives. Thus,
combining spatial data with narrative sources adds light to our view of the human
patterns of life in sixteenth-century Iberia.
32
Dobbs, Indian Trading Path, 6.
17
CHAPTER 2: INTRODUCTION: RECONSTRUCTING IBERIAN
TRANSPORTATION NETWORKS ACROSS FIVE CENTURIES OF PLACENAME CHANGES
An emerging research focus, geographically-integrated history posits first that
to fully understand history, one must integrate space, place, and time; and second, that
modern data management and visualization tools provide an answer to this
challenge.33 To this end, the development of a historical GIS database, essentially a
digital gazetteer, provides an example of ways in which researchers can integrate
historical data with modern data management tools. It provides an example of ways in
which researchers can integrate historical data with modern data management tools.
Figure 2.1 provides a facsimile of Villuga’s book, demonstrating its narrative format
and how he provided the league distances.
Through GIS-based visualization, the digital gazetteer compares the modern
names with their sixteenth-century ancestors and analyzes regional and temporal
patterns of place-names and the possible reasons behind individual place-name
changes. The royal decree of 1834 reorganized the provinces of Spain. This
transformation greatly changed place-names, making it difficult to write a
geographically-integrated history of the earlier period if researchers lack a means by
which to identify the locations of historic places on the basis of the coordinates of their
modern counterparts.
33
J.B. Owens, "Toward a geographically-integrated, connected world history: Employing geographic
information systems (GIS)." History Compass, 5, 6 (October 2007): 2014-2040; doi: 10.1111/j.14780542.2007.00476.x.
18
Figure 2.1. Title Page and Beginning Page of ‘Cuenca to Burgos’ Route, Juan P.
Villuga, listing beginning and ending termini, total league distance for the route,
place-names, and league distances in between each location.
A gazetteer for this project, the Villuga Gazetteer, adds a geographic location
to text-based labels of places, or place-names, from a 1546 traveler’s guide to 139
major routes in “Spain” (as the Iberian Peninsula was then known), in which the
author, Juan Pedro Villuga, listed each stopping place along each route and indicated
the approximate distance between each of them.
19
Villuga’s guide provides a valuable tool for connecting sixteenth-century
Iberian place-names to their nineteenth- and twentieth-century changes. His book has a
simple layout, providing a list of transportation routes. It also lists all of the places
along each of the routes. He first listed the beginning and ending termini of each route,
and the route’s length in leagues, a general and imprecise distance. Then he followed
by listing the place-name that resided closest to the beginning terminus, and its
distance from that terminus. For example, figure 2.1 lists the beginning of a route, “Ay
de Cuenca a Burgos,” or ‘from Cuenca to Burgos.’ The record lists the league distance
for the route below on the right side of the page, as “lxvii,” the roman numerals
denoting a distance of 67 leagues. Then it notes “a gillaron,” meaning ‘to gillaron.’
This denotes ‘from Cuenca to gillaron,’ also measuring the league distance between
these two municipalities at one league. Then it lists the next stopping place along the
route, ‘from gillaron to la venta,’ at two leagues’ distance. Note that gillaron is
identified by its distance from two other place-names. This level of context provides
the greatest benefit of the record. The user can search for each place-name on a
modern map by locating one or both of the nearest place-names listed in its route.
Moreover, in most cases, a number of other place-names are sufficiently near to
provide further context for the search. For example, Cuenca provides further context to
gillaron than the two places nearest to it on the route. From this book the user can gain
an extensive knowledge of the sixteenth-century Iberian transportation system and its
most important municipalities. For example, Burgos appears in many of the routes,
revealing that many roads began, passed through, or ended in that place. The guide
20
includes the most important coastal towns, which expands the context of the Iberian
transportation networks. Villuga’s record thus provides a basic structure on which
further historical studies can build to internationalize the context of the Iberian
Peninsula’s transportation routes. Alicante and Sevilla, for example, connected to
Mediterranean Sea routes; Lisboa and Porto, Portugal connected to important Atlantic
Ocean routes. Towns in the north along the Pyrenees mountain range that separates the
modern countries of Spain and France connected to transportation routes in France.
This record provides a valuable context for trade and travel in sixteenth-century Iberia,
and its connections to the international transportation network.
The many types of data the historian may add to this basic transportation
structure include population data, and other economic data, thus reconstructing to a
significant degree the economic structures of Iberia during its time as the center of the
first European empire that had a global stretch. Moreover, even this basic structure
provides a critical context for the available historical records on the economic life of
the Spanish empire. Defining the transportation networks and population centers of
Castile, and its connections to the rest of the Peninsula, reveals many important
aspects of the crucial Peninsular economy, and provides a context for further
explanation of the Castilian decline, through additional historical records.
21
CHAPTER 3: INTRODUCTION: BUILDING A DIGITAL HISTORICAL
GAZETTEER, LOCATING MISSING PLACES, AND VISUALIZING
POPULATION CENTERS
This chapter introduces the concept of a gazetteer database, some of its
scholarly uses, and the structure of the gazetteer for this study. The gazetteer database
for this study supplies a platform for analysis of the league in Iberia, a widely-used but
vague measurement of distance, and a method for finding missing places from the
historical source for that gazetteer. This chapter also describes adding population data
to the gazetteer to further contextualize Villuga’s transportation networks. These
networks exist in a polycentric urban region format, where no one city dominates the
rest in population levels, but cities with a similar population level within a region
interconnect and are interdependent in such functions as economy, political power, and
religious practice. The resulting spatial data provides essential contextualization to
many studies of the Iberian Peninsula’s sixteenth-century economy and society.
describe
3.1 The Gazetteer
A gazetteer’s basic elements consist of a place-name and a geographic location
corresponding to it. Road maps and atlases commonly use this format. One looks up a
place he or she wants to visit in the gazetteer of the atlas and the gazetteer provides a
map page, column, and row, allowing the map user to quickly locate the place-name
and all of the spatial elements surrounding it, on that map page (see figure 3.1 for an
22
example).34 Linda Hill, of the Alexandria Digital Library (ADL), the primary authority
on gazetteer content standards, explained that gazetteers have many uses, which
consist of providing a means of translating between formal and informal methods of
georeferencing, the latter consisting of textual labels for geographic locations, and the
former presenting mathematical coordinates for geographic locations.35
Figure 3.1. Ottawa Map: Cra ’s, page 61, B7
For a good example see: George F. Cram. Cra ’s Quick Refere ce Atlas a d Gazetteer of the World.
Containing 105 Newly Engraved Maps and Over 40,000 Index Entries with the Latest Areas and Census
Statistics. Edited by Dr. Eugene Murray-Aaron. (New York: George F. Cram, 1906), figure 3.1.
34
35
Linda Hill, Georeferencing: The Geographic Associations of Information. (Cambridge, MS: The MIT
Press, 2006), 97.
23
Hill argued that gazetteers, as knowledge organization systems (KOS), or
systems that classify data into schemes, provide a useful platform for organizing any
data that contains a spatial footprint, i.e., a geographic location. Such records as
satellite imagery, historical texts, museum catalogs of historic items, and library
catalogs, can optimize their organization by converting their data to a gazetteer format,
which translates informal spatial labels, e.g., place-names, into formal georeferenced
coordinates.36 Goodchild and Hill explained that many fields of research have begun
incorporating gazetteers into the study of public health, natural history data
management, cultural history, and other scholarly pursuits. Studies of geographic
information retrieval, spatial cognition, and social history have also begun using
gazetteers.37
Beyond the traditional model of the gazetteer as a name with a geographic
location, Mostern and Johnson argued for a ‘historical event gazetteer,’ which not only
expands the traditional gazetteer format, but provides a new paradigm beyond just
recording locations by their names. They argued for a human-based approach that
focuses on recording places as events from the human perspective. No event exists
separate from a place, and humans define places. Moreover, they do it in the context of
specific events. Therefore, they argued for a database that provides connections
between events, producing event timelines that incorporate the events’ locations. This
36
Hill, Georeferencing, 91-100.
37
Mi hael F. Good hild & L.L. Hill, I t odu tio to digital gazettee esea h, International Journal of
Geographical Information Science, 22 (2008), 1039–1044.
24
new gazetteer version properly contextualizes each place-name according to how its
human inhabitants understood it.38 This thesis produces a gazetteer based on the data
available for a snapshot in time, specifically the year 1546, when Villuga published his
guide, and cannot follow Mostern’s and Johnson’s new paradigm, which goes beyond
the scope of this paper.
The Villuga Gazetteer provides the modern versions of place-names found in
Villuga’s guide and their formal geographic coordinates, as well as each place-name’s
corresponding route, or routes. The beginning and ending termini of the historic
Iberian routes consist mostly of still well-known and populated locations, such as
Madrid, Barcelona, Sevilla, and Toledo. These well-known beginning termini
provided a good starting point for locating the rest of the place-names, in combination
with the general distances in leagues provided in Villuga’s record. Figure 3.2 provides
a visual model of the tables of the Villuga Gazetteer database. Databases contain
Figure 3.2. Villuga Gazetteer Database Model
38
R. Moste & I. Joh so , From named place to naming event: creating gazetteers for history,
International Journal of Geographical Information Science, 22 (2008): 1091–1108.
25
various types of data, such as text or numbers. Converting the Access-based database
to a GIS database format, known as a geodatabase, provides spatial data types that GIS
can visualize on a map, such as points, lines, and polygons.
The primary table in the database, the gazetteer, connects each modern placename in Villuga’s guide to their sixteenth-century versions (variants) and the routes to
which he assigned each of them. The gazetteer table is called ‘Villuga_Gaz’. It
connects to a lookup table, called ‘primary_historic_muni_names’ (‘MUNI’ stands for
municipality). The lookup table also connects to the ‘primary_routes’ table, which lists
each of Villuga’s routes, to the gazetteer table. This allows one to query the database
to determine to which route each modern name belongs, and to which historic names
that name corresponds. The lookup table provides the capability of assigning more
than one historic name to each modern name, which proved important in the Villuga
Gazetteer. Villuga often listed more than one version of a place-name that belonged to
the same geographic location. The confidence table provides those interested in using
the database a tool to aid them in determining the degree of validity of each match
between a place-name’s historic and modern variants.
In order to create a greater density, or a larger number of examples, of
sixteenth-century place-names to compare with those used in Villuga’s record, the
1834 administrative reorganization, and current times, the Villuga Gazetteer includes
place-names from relaciones topográficas of the provinces of Cuenca and Toledo.
These records supplied responses from self-governing municipalities to a royal
26
questionnaire from the 1570s that described each place in details such as population
and administrative jurisdiction.
3.2 The League in Villuga’s Record
The league provided sixteenth-century European travelers with a general,
imprecise, distance measurement, which varied in length across the Iberian Peninsula.
Villuga used the league as his measurement unit, with vague and variable lengths
throughout his record. In some places, based on GIS analysis, the league equals
approximately (never exactly) 4.2 kilometers; in others, it approximately equals 5.6,
and several other lengths for the league emerge from this study. This study’s efforts to
match all historic place-names from Villuga’s record to their modern variants brought
about the discovery of a number of missing place-names. Finding these missing places
required more understanding of the league’s usage in the Iberian Peninsula. Ranges of
kilometers-per-league distance, derived from a cartographic classification analysis, fit
the data better than one single measurement. In other words, a range of distances (e.g.,
0 – 5.6 kilometers per league) fit Villuga’s data better than one single measurement,
such as 5.6 kilometers per league. An analysis using spatial autocorrelation determined
where ranges of league distances fit best, and described the spatial patterns of the
measurement length of the league. Spatial autocorrelation analysis tools in GIS can
identify spatial patterns, such as those found within the distance data in Villuga’s
guide. Spatial autocorrelation determines to what degree spatial features, locations on
a map, demonstrate similarity in attributes as a function of their spatial proximity to
27
each other. This study provides a brief discussion of spatial autocorrelation, and then
analyzes Villuga’s league distances and determines what spatial patterns his distances
demonstrated. The results of the spatial autocorrelation analysis will aid researchers in
deciding which of the distance ranges pertains to a particular missing place-name from
Villuga’s record. This information provides a helpful tool in locating the missing
place-names, by providing minimum and maximum distances for each place-name
from the other place-names in its route(s), thus narrowing down their possible
locations on the map.
Peter Enggass raised the issue of the extreme variability of the measurement of
the league in Spanish history. He stated that nearly half of the Iberian Peninsula’s
regions had unique measurements for the league prior to Spain’s implementation of
the metric system. Obviously, Villuga did not provide this information, because he
probably did not have it. The lack of standardization of units of measurement created
imprecision of measurement. Standardization became law and practice only after
Spain adopted the metric system in the late nineteenth century. 39 Researchers can use
Iberia’s extreme topographical variability as a template for the variation of the
measurement of the league. Less-populated areas used a good deal of guesswork in the
league’s measurement (due to extremities of topography), and they used several
measurements for the league.40 This thesis covers the application of GIS and
39
E ggass, Pete M. The “pa ish league: A geog aphi al o spi a , Jou al of Geog aph 70 (1971),
407.
40
Ibid, 410.
28
statistical analysis tools to produce a solution to the difficulties in measuring the
league.
Comparison of Leagues in Two Traveler’s Routes
Gaspar Barreiros was Portuguese, a “man of the church, erudite, and a writer of
classical skill” who was born in Viseu, and died in 1574. He became a member of the
priesthood, and journeyed to Rome to honor the Pope on behalf of the cardinal don
Enrique.41 On his way to Rome in 1542, Barreiros described the journey from Badajoz
to Milan, Italy. His record presents no description of his mode of travel. Barreiros’s
journey took him from the far western edge of Castile to the far north-eastern corner of
Catalonia. His record, which included each stopping place and the distance between
them in leagues, provides a good comparison with the record of Juan Pedro Villuga.
Analysis of a section of the journey common to both records provides a comparison of
the league as recorded by these two travelers. The analysis considers each stopping
place in both records, as well as the place-names each author recorded. This
comparison provides greater context for the league measurements and place-name
conventions of both records, by determining the degree of similarity between the
sources.
Gaspa Ba ei os, Viajes. I Viajes de e t a je os po España Po tugal: desde los
tiempos más remotos hasta comienzos del siglo XX, comp, trans, ed. J. García Mercadal .
Vols. 1-3. (Valladolid: Junta de Castilla y León, 1999), 117.
41
29
Gaspar Barreiro, 1542
PlaceName
League
Alcalá de
Zero
Henares
Dist.
Barreiros, Column 2
PlaceName League
la venta de
2
Peñalva
Villuga, Column 2
PlaceName League
Guadalajara
Tórtola
Torre
Hita
Padilla
La Casa
Miralrío
Bujalaro
Fraga
Alcaraz
Lérida
Belloc
Cidamón
Mollerusa
Golmes
Bellpuig
La Grassa
Tárrega
Talhadel
Cervera
fraga
alcaraz
lerida
beloch
3
2
1
1
molarusa
1
el puyg
2
tarraga
1
cervera
1
los
mesoncillo
s
2
mon
maneu
1
porcarises
golada
la puebla
piera
masquesa
1
2
1
1
2
Sigüenza
Hijosa
Juan Villuga, 1546
PlaceName League
alcala de
Zero
henares
Dist.
guadalajar
4 a
2
2 tortola
2.5
3.5
1.5 hita
1
1 padilla
1
0.5 la casa
1
0.5 miralrio
1
1 burjlaro
1
los molinos
2
vaydes
1
4 ciguença
2
2
Torralba
Fuencalient
e
1
Nodales
1
Arcos
Mirabueno
Huerta
fuencalient
1 e
3 Ostaletes
Momeneo
Porcarizes
to Igualada
2
0.5
1
Monreal
Ariza
arcos
1 monreal
1 ariza
Contamina
1
2
2
1 Collbató
Esparregue
ra
2
2
1
1
0.5
0.5
0.5
2
1.5
0.5
0.5
1
1
1
2
1
1
Table 3.1. Journeys of Barreiros and Villuga compared.
Table 3.1 presents a comparison of a portion of Barreiros’s journey, with
corresponding parts of two of Villuga’s routes. One place, Alcalá de Henares, falls at
the zero distance, since it sits at the beginning point of the portion of Barreiros’s
journey that overlaps with Villuga’s route for the area.
30
Gaspar Barreiro, 1542
PlaceName League
Juan Villuga, 1546
PlaceName League
Alhama
0.5
medina celi
luna
lama
1
1
1
Bubierca
1.5
vbierca
tequa
terrer
calatayud
1
1
1 Barcelona
1 Moncada
Calatayud
la venta de
San
Esteban
Frasno
La Almunia
casa de los
Romeros
La Muela
Zaragoza
Puebla (de
Alfindén)
Alfajárin
4
2
0.5
2.5
Barreiros, Column 2
PlaceName League
Nuestra
Señora de
2
Montserrat
Martorell
1
San Andrés
0.5
Molins de
Rey
1
La Roca
Linás
San Celoní
Astarlid
(Hostalrich)
Gerona
Madinham
Villuga, Column 2
PlaceName League
martorel
molin
derreche
el espital
2 barcelona
2 moncada
2
1
1
2
2 larroca
1.5 linas
2 sancelonij
2
1.5
2
el fresno
almunia
5
3
4
la muela
çaragoça
5
3
2
1
la puebla
a fajari
Osera
la venta de
Santa Lucía
Bujaraloz
1
2.5
3
3
hosfera
la venta de
santa lucia
burjalalos
2 Vascara
1 Figueras
el puente
de los
3 Molinos
Candasnos
3
candasnos
3 La Junquera
3 Perthus
Total
3 Leagues:
1.5 xunqueras
1 el pertus
Total
112 Leagues:
2.5
1
2 astarlid
5 girona
1
2
5
2 bascara
2 figueras
3
2
3
1
111
Table 3.1(continued). Journeys of Barreiros and Villuga compared.
This table demonstrates many of the issues with using the league as a
measurement of distance. First, these travelers recorded many distances between sets
of two places that did not correspond to each other. For example, while Villuga listed
Guadalajara at 2 leagues from Alcalá de Henares, Barreiros recorded it as 4 leagues.
31
Second, the routes each traveler took do not exactly correspond to each other. Only
Villuga recorded the place-name los mesoncillos, Barreiros recorded that he did not
stop at tequa and terrer, which Villuga records as part of the route. Third, the placenames in one record in several instances do not match their counterparts in the other
record. Villuga’s place-name el puyg, without its geographic location, does not clearly
represent the same place as Barreiros’s Bellpuig. The context of the route suggests that
these places do correspond to each other, but taken outside of that context, the traveler
can easily consider these two place-names as separate locations on the map. Finally,
these travelers have recorded some of these place-names in more than one language.
molarusa, which may be the Castellano version of the place-name, likely becomes
Mollerussa in Valenciano.
3.3 Fuzzy Logic and the Location of Missing Place-names
Villuga provided a data for an analysis of distance by listing the leagues
between each of his recorded place-names, but his ill-defined unit of measurement
creates difficulty for GIS users. GIS relies upon well-defined, or ‘crisp,’ data, rather
than approximations. This section presents fuzzy rule-based modeling, a part of fuzzy
set theory, a soft computing technique, that allows the analyst to use vague,
incomplete, and poorly-defined data, such as Villuga recorded in his traveler’s
handbook. Spatial interpolation tools are designed to fill in the gaps of incomplete
data, interpolating them by producing the most likely data values for the gaps between
the known data. However, it still works within the restrictions of a computer’s demand
32
for precisely-defined input data. This section provides a methodology that augments
existing GIS interpolation methodologies. Fuzzy rule-based modeling allows a domain
expert, in this case a historian, to write condition statements, or rules, which apply
these rules to a number of instances of vague data, by entering that data into the rules
as their parameters, or input values. This section discusses traditional GIS methods,
and then demonstrates how fuzzy rule-based modeling augments GIS analysis of
vague distance data. Our data comes from Villuga’s sixteenth-century guide.
The Significance of Unknown Places in Villuga’s Routes
The commercial and pilgrimage routes throughout sixteenth-century Spain and
Portugal list many familiar places, but the modern reader will not find a number of
them on the current map. With modern changes in transportation routes, some of
Spain’s historic ventas (country inns), small villages, way stations, and albergues
(hostels) have drifted into disuse and the dusts of time have obscured them. The old
ventas and small villages still exist in the more urban or politically-central regions,
with most having modified names, and expanding urban centers subsuming others.
Torreblanca de los Caños, for example, started out as a place separate from Sevilla, but
now Sevilla has subsumed it. Some of Villuga’s roads are obsolete, replaced by the
ferrocarril (railway) and the Red de Carreteras del Estado (Spain’s version of the
Interstate Highway). 42 The carters, muleteers, and pilgrims of the day used these routes
most frequently in the time of Villuga. They traveled mainly in regional or local
42
Mapa Oficial de Carreteras 2000. Centro de Publicaciones, Ministerio de Fomento.
33
patterns, but at times these companies traveled great distances. 43 While modern maps
still display the majority of Villuga’s place-names, approximately 260 of the 1,486
(17.5%) places do not exist on the modern map. Certain routes have many more
missing places than others, such as the routes from the Zamora region to Andalucía,
especially those going into Sevilla or Córdoba.
The missing places in Villuga’s routes offered key stops on travelers’ routes,
especially in the many unpopulated areas of Iberia, and finding them may provide
important information about the structure of the sixteenth-century travel system in
Iberia. GIS provides some useful tools for the interpolation of missing data, and can
aid the search for missing places from Villuga’s guide. A GIS analysis cannot prove
the proposition ‘Villuga’s league is here measured at 5.5 km’ to be exactly true.
Villuga did not record the location from which travelers measured the distance
between places. His data lacked precision and exact definition. Yuan explained the
difficulty of historic changes in place-names and their exact locations:
Over time, places may have multiple names, multiple foot prints, and multiple
identifications that question if the place remains to be the same place or should be
considered as a different place. Place-names are often labels synthesizing a coherent
space emerging from environmental, political, historical, and cultural interplays.
Change in place-names or place footprints have profound implications to events
occurring in these places.44
43
Transportation and Economic Stagnation in Spain, 1750-1850. (Durham, N.C.: Duke University Press,
1970), 72-74.
44
Ma Yua , Mappi g Te t, i The Spatial Humanities: GIS and the Future of Humanities Scholarship,
ed. David J. Bodenhamer, John Corrigan and Trevor M. Harris (Bloomington and Indianapolis: Indiana
University Press), 9.
34
Finally, the road system of the sixteenth century no longer exists as it did for Villuga’s
travelers; therefore, the distances between locations likely did adjust as a result of the
road network changing. If a new road evolved between two places, it may have taken
either the same course, a more direct course, or a longer course than it did in 1546.
The two final options would have changed the distance between two places, and the
analysis cannot determine conclusively the exact route of the sixteenth century.
Therefore the proposition ‘Villuga’s league is here measured at 5.5 km’ cannot be
conclusively proven. This demonstrates the need for a methodology which answers the
problem of data imprecision when using computers to process historical data.
GIS requires the use of exactly defined data, which the historical record did not
provide (e.g., the phrase “about 2 miles from here” does not give a crisp definition of
distance).45 Traditional GIS tools have attempted to provide methods of dealing with
this problem, but they require the use of exactly-defined input parameters. The Map
Algebra tool in ArcGIS 9.x, which runs conditional statements that help the analyst
locate spatial patterns, requires Boolean operators, which assign only a true or a false
state to data. Map Algebra assigns a piece of data to one of a number of available sets,
which are groupings of identical or similar data. Map Algebra assigns each piece of
data a state of either full membership in a set or no membership; a true or false state.
This requires pushing data into an exact definition, with no room for ambiguity or
vague meaning. This lack of room for error is the main characteristic of “mechanistic”
45
Lotfi Zadeh , Outli e of a Ne App oa h to the A al sis of Co ple “ ste s a d De isio
P o esses. IEEE Transactions on Systems, Man, and Cybernetics SMC-3 (1973), 29.
35
data, or machine-based data. In “humanistic” data, on the other hand, membership in
categories comes in degrees.46 While the computer or physical sciences, engineering or
mathematics, crisply define data, “humanistic systems” vaguely define data. 47 Coppola
et al., paraphrasing Zadeh, stated that “because of their complexity, ‘humanistic
systems’ (human-centered ones) cannot be modeled in the way one models
‘mechanistic systems [machine-based systems].’”48 The kilometer distances for each
league Villuga recorded fall into general ranges, rather than consistently using one
precise kilometer measurement for each league. Fuzzy rule-based modeling expands
on Boolean operators to allow degrees of membership in a set, and may assign data to
more than one set at a time. It expands upon existent GIS interpolation methods by
generalizing them, to fit them to the needs of fuzzy data analysis. 49 The analysis
focuses on adaptations of fuzzy rule-based modeling to GIS interpolation methods.
One interpolation tool, kriging, provides a method for the adaptation. It will be
discussed in chapter 4.
46
Emery Coppola, J.B. Owens, and Szidarovszky, Fuzz Rule-Based Modeling of Degrees of Trust in
Cooperation-Based Networks: Close Research Collaboration among Domain Experts (Historians) and
Mathe ati al Modele s. (presented at the TECT Strategic Workshop, Technical University of Madrid,
25-26 Sept. 2008).
http://mapas.topografia.upm.es/Dyncoopnet/presentations/Madrid_fuzzy_owens.pdf. (Accessed 27
February 2010). 8.
47
Ibid, 28.
48
Coppola, Owens, and Szidarovszky Fuzz Rule-Based Modeling of Degrees of Trust in CooperationBased Networks, .
49
Michael Smithson, Fuzz “et I lusio : Li ki g Fuzz “et Methods ith Mai st ea
Sociological Methods & Research 33 (2005): 431.
36
Te h i ues
Bolstad explained kriging and inverse distance weighting (IDW), two
algorithms of spatial interpolation. IDW relies upon a simple algorithm that divides
the inverse of the distance from a location by a given z (attribute) value, whether that
value represents elevation, or any other data the researcher desires to use. IDW assigns
the various locations on a surface an interpolated z value based on their distance from
the z value of the nearest control points. This method produces one possible problem:
the occasional spikes or pits in the interpolated surface. Kriging functions similar to
IDW, but eliminates the spikes and pits by not fitting the continuous surface to the
exact z value recorded at each control point. Kriging’s complexity is its drawback.50
This study’s interpolation model demonstrates the use of an interpolation surface,
which supplies one of the best methods available for creating continuous surfaces of
membership.51 Kriging provides one option from the GIS toolkit for locating the
missing places in the sixteenth-century Iberian transportation network.
Fuzzy Rule-Based Modeling and the Location of Unknown Place-Names
Zadeh, in his theory of fuzzy information granulation (TFIG), a part of fuzzy
set theory, explained that fuzzy granules consist of imprecisely-defined subsets of a
larger set. Zadeh addressed the imprecise nature of data from human sources by
creating a method of Computing with Words (CW), which fuzzy rule-based modeling
utilizes. Fuzzy rule-based modeling applies mathematical rules to textual descriptions
50
Paul Bolstad, GIS Fundamentals: A First Text on Geographic Information Systems. (White Bear Lake,
MN: Eider Press, 2008), 459.
51
Ibid, 457.
37
of phenomena that utilize natural language variables. Fuzzy if-then textual statements
describe rules of membership, and relationships between sets of data values. Fuzzy
rule-based modeling then combines each textual statement, previously converted to
numeric percentages, providing degrees of likelihood of membership in a given set, to
produce the most likely output value. Unlike set theory, which utilizes absolutes in the
measurement of a data value’s membership in a set, fuzzy rule-based modeling, by its
use of fuzzy if-then statements, offers the capability to describe the degree to which a
variable has membership in a set or number of sets.52 In set theory, even when an
object has partial membership in a set, it resides fully in that set.53 Set theory contains
no method of describing a variable of an intersection as having only partial
membership in one set in the intersection and relating more fully to the other set in the
intersection. The complexity of the human situation does not lend itself to such welldefined relationships among its parts.
Zadeh argued that TFIG can apply to any “concept, method or theory,” which
opens up the application of mathematical modeling of real-world processes to a
universal audience. The historian’s application of TFIG, through fuzzy rule-based
modeling, provides a multi-disciplinary focus that facilitates collaboration with a
mathematician trained in fuzzy rule-based modeling. Fuzzy if-then statements provide
a basis for CW. In CW, words provide labels for fuzzy granules. For example, the
word ‘orange’ refers to a whole variety of possible variations of the color that one may
52
Kuratowski, Introduction to Set Theory and Topology. (New York: Pergamon Press, 1972), 27-36.
53
Ibid, 27-36.
38
describe with that word.54 ‘Orange’ is a fuzzy subset of a variety of possible
combinations of ‘red’ and ‘yellow.’ The word ‘red’ identifies a fuzzy set, a grouping
of variations of the color, as does the word ‘yellow.’ The linguistic labels provide
fuzzy constraints upon variables. CW becomes a crucial tool when 1) the available
information is too sparse or imprecise to be described by numbers or 2) when the data
demonstrate “a tolerance for imprecision, uncertainty, and partial truth” that a human
can utilize in order to achieve a quick and sufficiently accurate answer or conclusion.55
Zadeh gave several examples of humans using TFIG to make decisions about what
route to take to make a round trip to and from the airport; how long a commute usually
takes, and so on. “In everyday decision-making, humans use only that information
which is decision-relevant. For example, in playing golf, parking a car, picking up an
object, etc., humans use fuzzy estimates of distance, velocity, angles, sizes, etc.” 56
Zadeh’s purpose for TFIG is to “exploit the tolerance for imprecision, uncertainty and
partial truth to achieve tractability, robustness, low solution cost and better rapport
with reality.”57
Hill explained that “a bounding box (minimum bounding box (MBB) or
minimum bounding rectangle (MBR)) is the smallest box that completely encloses a
54
Lotfi Zadeh, To a d a Theo of Fuzz I fo atio G a ulatio a d its Ce t alit i Hu a
Reaso i g a d Fuzz Logi . Fuzzy Sets and Systems 90 (1997): 115.
55
Ibid, 115.
56
Ibid, 116.
57
Ibid, 123.
39
spatial footprint.”58 Setting a preliminary general location for one of the missing places
from Villuga’s guide, in the form of a bounding box, supplies a key element for
utilizing fuzzy rule-based modeling to constrain the possible location of the unknown
place. The analysis of Villuga’s place-name, Luna, created a bounding box for the
fuzzy locations of the unknown places, and it marked the extreme outer coordinates of
the box as endpoints in the model of the location. The raster surface in figure 5.4
implies a bounding box that corresponds spatially with the raster surface. Smithson, in
his article on fuzzy set inclusion discussed how fuzzy set theory can augment or
generalize existing techniques, to handle fuzzy data. 59 Verkuilen argued for the need
for endpoints which constrain fuzzy sets. 60 Bounding boxes, within which fuzzy rules
may operate, provide an example. Their boundaries provide the endpoints.
When applied to this study, fuzzy if-then statements offer a key to finding the
locations of the unknown place-names listed in Villuga’s handbook. Villuga’s
distances in leagues facilitate a general location for the unknown place. Villuga’s
imprecise distance measurements create a general bounding box around the unknown
place’s location, providing room for error and generalization. The historian will
identify the various physical attributes of that general location, dependent upon its
historical context, as variables of if-then statements that will narrow down the possible
location of an unknown place. If, for example, a river or a mountain resides inside of
58
Hill, Georeferencing, 69.
59
“ ithso , Fuzz “et I lusio ,
60
Ibid, 468.
.
40
the bounding box, or if travelers required pasture at a stopping place, then these
features will constrain the possible location within the bounding box where people
could have built their venta.
3.4 Population Data Contextualizing the Transportation Structure
The user of the gazetteer for the study of Villuga’s guide can benefit further
from seeing the population levels of the network structures. Population data will aid
the user in determining the causes behind the structure of the transportation network.
The network in turn reveals something about the interactions among the
interconnected population centers, such as how well they connected to each other, and
where geographical or administrative barriers separated them. Visualizing these data
can determine where a region had strong network ties to one region, but not another. It
can aid in determining what population centers formed the strongest base for a region,
how the region connected with its neighbors, and what products went to which regions
via network connections between major population areas. It can aid in verifying the
level of population strength the well-connected areas had. This visualization can
identify courses of longer-distance travel, and consider their population levels, which
can reveal further information about the degree of strength the long-distance trade of
the Peninsula enjoyed at the time. It can aid in identifying which towns or groups of
towns influenced the transportation networks more than others.
When coupled with further sources of historical data, the population-enabled
gazetteer provides further contextualization of historic questions. Studies using it may
41
consider the political strength of the elite classes, and compare this to how wellconnected to other regions these elites were. Others can consider the economic habits
of the inhabitants of a region and its localities to determine what impact the
geographic structures of their population had on the type of economic activity in which
they engaged. Religious traditions and the strength of a centralized church in the farflung locations in a region can rest upon the clergy’s ability to transport religious
traditions from the central church, and how well those people who wished for isolation
accomplished their task. Moreover, scholars may use the gazetteer database to test the
impact of the various causes historians have proposed for the economic crisis of the
sixteenth and seventeenth century Castilian Crown.
Polycentric Urban Regions
Ringrose, in Transportation and Economic Stagnation in Spain, 1750-1850,
explained that in the eighteenth and nineteenth centuries, Spain’s underdeveloped
roads and transport technology, lack of navigable rivers, land ownership policies and
patterns, and the dispersed nature of settlement, stretched Spain’s transportation
infrastructure to its limit, causing a bottleneck in the peninsular economy.61 The
economy and transportation of Spain in the sixteenth century must have been much
like that described by Ringrose, with less traffic and different international economic
foci. Mauro and Parker explained that Spain had a population level between 6 million
61
Ringrose, Transportation and Stagnation, i.
42
and 7.5 million in the sixteenth century, and about 10.5 million in the early nineteenth
century; “the population of Spain in the 1590s was not matched again until the 1730s,”
and then grew during the late eighteenth and early nineteenth century. 62 In other
words, the population declined from the late sixteenth to the seventeenth century, and
then began to grow in the mid-seventeenth century until the end of the eighteenth
century.
Table 3.3. Mau o’s a d Pa ke ’s stud of “pai ’s populatio .
Mauro and Parker’s study explained that Spain’s population during the time
period Ringrose studied, in the eighteenth and nineteenth centuries, grew significantly;
yet the population remained in dispersed settlement in the period Ringrose analyzed.
The Spanish population, and in connection, transportation, in the sixteenth century
would have thus existed in a more dispersed pattern, barring some revolution in
settlement patterns between the sixteenth and eighteenth centuries, involving
urbanization and concentration of the population. Transportation between the
62
F édé i Mau o a d Geoff e Pa ke . “pai . I An Introduction to the Sources of European
Economic History 1500-1800. eds. Charles Wilson and Geoffrey Parker, (London: Methuen & Co. Ltd.,
1980), 37, table 3.3.
43
dispersed population centers in Spain rested on animal transport, roadside pastures,
and spontaneously-created roads, and therefore moved slowly. 63 Each route contained
stopping places at an average of about 4 to 8 kilometers apart.
Christaller, in his central place theory, argued that settlements crop up out of a
necessity for central locations of certain demanded commodities; the more urgent
commodities create a larger central place, and the less necessary, more expendable
commodities create smaller central places. Goods have a minimum and maximum
distance of sale; just over the outside edge of the minimum distance around a central
place is where sufficient portions of the population purchase the central good (i.e., the
good produced and sold in the central place) to make it profitable enough for the
producer/seller to justify its production and sale. The maximum distance represents the
distance at which people may no longer efficiently travel to the central place for the
good, and must find another central place that exists nearer-by to travel to for the
purchase of the central good.64
Christaller explained the effects of having various forms of social and spatial
organization in the peripheral areas of a central place. If nucleated villages, sufficiently
large and populous, fill the countryside, their geographic patterns support the growth
of central places; if it consists of dispersed farmhouses, this precludes or diminishes
63
Ringrose, Transportation and Stagnation, xxi-xxii.
64
Ibid, 49-58.
44
the strength of central places.65 If a central place rests near enough to a village, the
villagers may purchase its bread more cheaply than they can make it at home, due to
the small distance and the limited difficulty in traveling to the central place, versus the
cost in time and man power of home-made bread. If the consumer lives in the
countryside farmhouse, at a larger distance from the central place, he or she will spend
less to produce bread at home than by traveling to the central place to buy the bread;
this leads to the diminishing of importance of the central place.66 Spain’s settlement
pattern often followed the latter pattern, which taxed the transportation infrastructure
more than a highly nucleated system.67
Spain’s transportation challenges, harsh environmental setting, and dispersed
settlement have partially caused the regionalism and inter-provincial rivalries and
cultural divisions that mark Spain’s identity throughout its history. Without sufficient
modes of transportation, among other reasons, the population remained fractured into
local spheres-of-influence that the unifying royal authority of the Castilian Crown
could not fully unite.
A study of polycentric urban regions, which are a topic in urban planning,
augments the contribution of central place theory to our understanding of sixteenthcentury Spain’s urban context. Central place theory, according to G. William Skinner,
65
Christaller, Central places in Southern Germany. trans. Carlisle W. Baskin. (New Jersey: Prentice Hall,
Inc., 1966), 31-33.
66
Ibid, 33.
67
Ibid, 47-52.
45
focuses mainly on retail activities: what products a central place makes available (i.e.,
dominant urban center connected to a hinterland) and how far from that central place a
customer will travel to buy a retail product. Variables such as topography and the
transportation system and vehicle technology, as well as price competition from other
urban centers, factor into the geographic reach of the central place. That reach varies
by good; one good offered may have only fifty percent of the distance reach held by a
more popular and scarce good offered at the same urban center.68
Polycentric urban regions, on the other hand, comprise urban systems in which
no one urban center holds full dominance; each center works with the others in a
complementary and synergistic relationship.69 This system provides the benefit of
shorter travel distances within the region, and the heightened labor pool from which all
of the urban centers in the region can draw for their needs. Some examples of this
kind of region consist of the Randstad in the Netherlands, consisting mainly of
Amsterdam, Rotterdam, and The Hague; and the Rhine-Ruhr region in Germany.70
Whebell contended that most urban regions connect to their neighboring
regions via a central place at the periphery of the region (a “regional entrepôt”);71 that
68
G. William Skinner, Cities a d the Hie a h of Lo al “ ste s, i The City in Late Imperial China.
ed. G. William Skinner. (Stanford: Stanford University Press, 1977), 277.
69
Evert Meijers, Pol e t i U a Regio s a d the Quest fo “ e g : Is a Net o k of Cities Mo e
tha the “u of the Pa ts? Urban Studies, 42(2005), 765-781.
70
David Batten, Net o k Cities: C eati e U a Agglo e atio s fo the
32 (1995):313-327.
71
Dobbs, Indian Trading Path, 21.
46
st Ce tu . U a “tudies
the transportation infrastructure provides a major determinant in the spatial
configuration of an urban system, and that this factor tends to draw urban systems into
a more or less linear pattern; and, some dominant urban centers develop at the edge of
regional markets, and act as connectors for one region to neighboring regions, or to a
wider network that may reach nationally or globally. 72 He said that Christaller
discovered more than he may have realized when he postulated the “traffic principle,”
which stated that the most favorable spatial pattern of a group of central places follows
important traffic routes.73 Dobbs modeled the North Carolina Piedmont Urban
Crescent using a blend of central place theory and polycentric urban region theory.
She modeled how in the region several urban centers developed together which
possessed similar levels of regional dominance. Their geographical isolation from
major regions to the south and north constrained the movements of merchants in and
out of the region. Topography on the edges of the North Carolina Piedmont region
constricted the development of a dominant central place which would connect the
region to its neighbors; instead, several equally dominant places on the periphery of
the region developed, with each connecting to various routes which accessed the
Piedmont.74 Dobbs defined polycentric urban regions: “a system of politically
independent units which are functionally interdependent, but which have become
something more than the sum of their individual parts; not merely a conurbation but a
72
C.J.F. Whebell, Co ido s: A theo
Geographers, 59 (1966): 1-26.
73
Ibid, 2.
74
Dobbs, Indian Trading Path, 6.
of u a s ste s. Annals of the Association of American
47
locus of heightened creative energy.” 75 These systems produce centers of synergy and
complementary relationships between the various centers. 76 Dobbs described the
critical variables of these systems as consisting of trade, banking, credit, and
commercial systems.77
Skinner’s work on China’s system of central places gives a good source of
comparison for Spain’s configuration of its system of central places. “Urban
development,” he argued, “in the core areas caused urban underdevelopment in the
peripheries.”78 He stated that “major transport routes of all types foster linearity by
attracting (as it were) central places that would otherwise [according to the central
places theory’s ideal model] be sited on a triangular lattice”.79 Skinner argued that
populated places will often grow near a river (this would apply to road systems, as
well), when in the model they would develop away from the river on a triangular
lattice. This alignment distortion of the model towards the nearest transportation
network resembles a tenant of Whebell’s linear theory which states that transportation
networks such as rivers provide a central aspect of any settlement system. 80 Dobbs
asserted that the Piedmont urban system developed in such a way that it demonstrated
75
Ibid, 6.
76
Meije s, “ e g ,
77
Dobbs, Indian Trading Path, 18, 21-22, 26.
78
“ki
79
Ibid, 293.
80
Whe ell, Co ido s: A theo
-768.
e , Cities a d the Hie a h of Lo al “ ste s,
of u a s ste s,
48
-3.
.
this phenomenon. Whebell’s linear theory locates some dominant sites on the
periphery, and central place theory often locates them in the geographic center of their
region.
Skinner also pointed out that within regional economies the actual pattern of
central places more frequently follows the central place theory model in the core areas
than in the periphery; distortion away from the model occurs at the periphery because
the regular pattern of the model often gets interrupted by the geographic features, such
as rivers or mountains, that often inform the boundaries of a formal or informal region.
This interruption creates modifications in the pattern that limit the regular relationships
between places up and down the regular hierarchy of places, i.e., a lower-level place
on the periphery of a region will have fewer higher-level places from which to choose
for supply of its needed goods.81 This interruption of the model evidences itself in
many regions in Spain, due to the constraints on the model forced by Spain’s
mountainous terrain. Spain’s lack of navigable rivers also has an effect on the central
places model. Together, Spain’s abundance of mountainous terrain and its lack of
navigable rivers constrained the connections between regions, so that similar to the
polycentric focus of the Piedmont region in North Carolina, there existed multiples of
equally dominant urban centers for each locally-oriented regional network. This
pattern demonstrated itself throughout Spain prior to the development of Madrid as the
Household and Court of the Castilian Crown in the late sixteenth century. The Court
81
Ibid, 297.
49
gradually influenced the realignment of the transport and economic networks of the
whole of the Iberian Peninsula.
50
CHAPTER 4: METHODS OF BUILDING A HISTORICAL GAZETTEER,
ANALYZING MISSING PLACES, AND ADDING FURTHER HISTORICAL
CONTEXTUAL DATA
Methods for each aspect of this study provide solutions to challenges
frequently faced by scholars of historical GIS and geographically-integrated history.
Gazetteer best practices discovered by major authorities contribute to the study’s
design. Confidence levels for each modern place-name provide a measure of validity
to the place-name matches provided by the gazetteer for this study. A comparison of
the same place-name across multiple time frames reveals place-name changes listed in
records across several centuries, to discover the most commonly-used variant for a
given location in Villuga’s guide. An analysis of league measurements provides the
researcher with data ranges, thus making the data more compatible with GIS tools, and
utilizes spatial autocorrelation analysis to determine what patterns of use existed for
each distance measurement for the league. Traditional GIS methods of locating
unknown data and fuzzy rule-based modeling are combined to provide GIS users with
a method of utilizing the vague data so often found in historical sources. At times, data
needed for a study will already exist in digital form, and this chapter describes a
method of integrating two gazetteers to integrate existing digital data inside a GIS.
4.1 Gazetteer Methods
In the creation of this gazetteer, standards of design offered important guidance
on what elements the gazetteer should include. Many authorities on gazetteer
standards list the following common core elements for gazetteers: geographic
51
coordinates (or footprint), place-name, date (time stamp), feature type, and data
source note.82 Some of the authorities also list other common key elements: nearby
place, and part of (political hierarchy). This work cannot cover the full scope of a
gazetteer’s content; therefore, the discussion only covers core elements.
The Alexandria Digital Library (ADL) project at the University of CaliforniaSanta Barbara provides the most ambitious model of best practices in gazetteer design
and implementation.83 To overcome the difficulty with many different gazetteer
designs, the ADL Content Standard supplies the need for “agreed-on standards of
access and data sharing specifically designed for distributed, independent
gazetteers.”84 Linda Hill, of the ADL project, explained that the essential elements in a
gazetteer consist of: 1) a name, and optionally, its variants; 2) a class/type, chosen
from a classification scheme of feature types; and 3) a geographic location,
represented by some format of coordinates or spatial footprint.85 Goodchild and Hill
refer to three core gazetteer elements that give the minimum required description of a
geographic location: a name, a footprint, and a type. 86
82
For an example see: Merrick Lex Berman, Pla es i the Past: What’s i a Na e? P ese ted at
PACSL GeoHistory Network Conference, Philadelphia, PA, December 2005).
83 Li da Hill’s o li e pape i de . , http://www.alexandria.ucsb.edu/~lhill/paper_drafts/.
84
Hill, Georeferencing: The Geographic Associations of Information, 93.
85
Ibid, 107.
Mi hael Good hild & L.L. Hill, Introduction to digital gazetteer research,” I ter atio al Journal of
Geographical Information Science, 22 (2008): 1041.
86
52
The ADL Content Standard gives the optional elements of “beginning
date,”“ending date,” and “time period note.” The ADL Content Standard does not
mandate these elements. Making these mandatory may not be feasible when historical
narratives do not provide a clear date for their place-names. While the Villuga
Gazetteer resembles the more limited schema for temporal data of place-names used
by the ADL Content Standard, this results from the relative simplicity of the Villuga
Gazetteer. The Federal Geographic Data Committee (FGDC), an important authority
on geospatial data in the United States, which offers no gazetteer content standard of
its own, refers the user to the ADL Content Standard for guidance on gazetteer
design.87
The Open Geospatial Consortium and the International Standards Organization
produced standards of gazetteer design at the same time the ADL produced its
standard. The OGC relies for their gazetteer standard upon the model provided by the
International Standards Organization (ISO) 19118, published in 2006. The OGC’s
webpage provides a description of its gazetteer under “Best Practices Documents,” as
“Gazetteer Service - Application Profile of the Web Feature Service Implementation
Specification.”88 It prescribes four related tables for the gazetteer that provide the user
with all of the core elements of gazetteer design. The Villuga Gazetteer provides all of
87
Fede al Geospatial Data Co
ittee Histo i al Data Wo ki g G oup, Digitisi g Histo i al Maps a d
Cha ts, FGDC We site, http://www.fgdc.gov/participation/working-groupssubcommittees/hdwg/DigitMaps/Alexdig/?searchterm=ADL.
88
Ope Geospatial Co so tiu , Best Practices Documents: Gazetteer Service - Application Profile of
the We Featu e “e i e I ple e tatio “pe ifi atio , Ope Geospatial Co so tiu ,
http://www.opengeospatial.org/standards/bp 2006, figure 2.8.
53
these elements except for a web-based interface, but users may access it via the
internet at: http://giscenter-rd.isu.edu/SpainPortugalRoutes.
The China Historical GIS (CHGIS) project has contributed to the topic of
gazetteer creation by its methodology of disambiguating place-names (i.e., making all
place-names unique) across temporal and spatial extents. It advocates for the inclusion
of a time stamp element that identifies place-names by their temporal extent, thus
avoiding ambiguity among temporally-variable place-names.
Berman, of the CHGIS, argued that the major challenge of place-name
gazetteers lies in disambiguating the historic and modern name variants to their
locations, so that a query can only return one name as its output. Time stamps or
values, he argued, can distinguish place-names from their variants easily if each
variant has its own time stamp.89 For the purposes of the CHGIS, an ontology-based
scheme, i.e., a hierarchy of administrative divisions, such as autonomous regions and
their provinces, and a hierarchy of feature types, such as populated places and their
subtypes, allowed the non-ambiguous matching of place-names from different time
periods to be greatly increased, due to the use of a time stamp, a feature type, and a
parent country.90 The Villuga Gazetteer geodatabase design covers temporal variation
by the use of a modern place-names table and a historic place-names table, which
89 Me i k Le Be a , Geo efe e i g Histo i al Pla e a es a d T a ki g Cha ges o e Ti e.
(presented at the Georeferencing Workshop, Harvard University, Cambridge, Massachussetts, March
2008), China Historical GIS, http://www.fas.harvard.edu/~chgis/ (accessed April 12, 2010).
90 Be
a , Geo efe e i g Histo i Pla e a es,
.
54
relate to each other such that users can make queries to tie the modern and historic
names to their geographic coordinates, and to each other.
The design of the Villuga gazetteer includes the crucial elements called for by
gazetteer design authorities discussed above: a ‘primary key’ field, to aid in its use in a
relational database structure; a ‘modern name’ for each location; ‘longitude’ and
‘latitude’ fields, which provide a modern footprint; a ‘data source’ field which
identifies the source from which the coordinates came (e.g., an online gazetteer); a
field for a ‘unique source ID’ which the source of the coordinates provided; a ‘part of’
field to denote to which political boundary (administrative unit) the place-name
belongs in the present-day; a ‘feature type’ field, which provides a classification
scheme for map features such as mountains, rocks, rivers, or man-made elements like
buildings; and, a ‘near’ field to provide information on a nearby place-name location.
Including time stamp fields, which provide a record of the temporal extent of a placename for a geographic location, would provide further documentation, but go beyond
the needs of the Villuga Gazetteer. Sources from different temporal extents exist in
separate tables. The coordinate fields only provide modern coordinates, which may
cause some confusion if the user desires to track the location of a village’s or town’s
geographic coordinates over a temporal scale, but a record of the temporal changes in
spatial coordinates for Villuga’s place-names would go beyond the scope of this thesis.
A compilation of 139 Excel files lists all of the place-names from Villuga’s
record, one file per route, and lists the distance in leagues for the entire route, and in
between each place-name. These files were compiled into an Access database. ESRI’s
55
ArcGIS software converted the Access database into a geodatabase format, and then
provided relations, or links, to all of the other tables in the geodatabase. This
conversion allows the user to query by each route listed in the historic record. The
Access database (figure 3.2) includes a table for the gazetteer, a table listing each
route, a lookup table, and a table of confidence levels for each place-name. The routes
table lists each route’s starting and ending termini. The lookup table lists each historic
name, its route, and its sequential order in each route. These data allow the user to
query by the route(s) of each place-name, and also to query which place-names reside
in each route. Confidence levels aid the user in determining the merit of the data
available from this study. The confidence levels for each place-name vary in range
from 1 to 4; 1 denoting high confidence and 4 denoting low confidence. Figure 5.2
maps these confidence levels. Confidence levels for most places, 72%, fell between 1
and 2. The confidence levels for each place-name depend upon three criteria, 1) the
level of accuracy, derived by hand, between the actual distance (in kilometers) of a
place-name to the next place-name, and the distance in leagues provided by Villuga
between these two place-names; 2) the degree to which the modern variant names
match their historical variants in Villuga’s record; 3) modern province assignment and
nearby place-names also contribute to the confidence level, by assuring each placename’s uniqueness.
After searching modern atlases, queries of the online gazetteers by modern
place-names determined which modern place-name variants sufficiently matched the
historical place-names. The U.S. Geospatial Intelligence Agency’s GEONet Names
56
Server (GNS) required precise spellings in order to get a match.91 The ADL Gazetteer
Server Client92 would often provide many variants distinguished by their spatial
location. To assure a unique name, the Villuga Gazetteer documents each one in the
context of its province and a nearby place-name. This proved especially important in
Spain, due to the many place-names that repeat multiple times throughout the country.
For example, this process aided disambiguation of a given place-name from nearly
identical names in other modern Spanish provinces. Disambiguating place-names in
this manner provides support while searching online gazetteers that allow the user to
search on large or small spatial scales. The data structure of Villuga’s record provided
a great deal of local context for each of the place-names, by identifying places before
and after each place-name, and the distance between each of them in leagues. Each
route started and ended at generally well-known locations, most of which still exist on
the map. This enabled a starting point of high confidence, which allowed the next
place on the route to be ascertained with some confidence due to its location near the
starting terminus of the route, thus aiding the process of populating the gazetteer with
valid data.
91
U.“. Natio al Geospatial I tellige e Age . GeoNa es “e e GN“ , http://earthinfo.nga.mil/gns/html/index.html (accessed April 12, 2010).
92
Alexiandria Digital Library. Gazetteer Server Client Website.(University of California at Santa Barbara,
2005). http://alexandria.ucsb.edu/gazetteer/ (accessed May 18, 2010; no longer operational).
57
Place-Name Density Across Time
To create a greater density of sixteenth-century place-names with which to
compare the 1834 Reorganization names, a compilation of place-names from various
historical sources provides examination of those variants across several time frames
from the regions of Cuenca and Toledo. Examples of places with several variants
within Villuga’s record, identified by their locations as representing the same place on
the ground, demonstrate the lack of place-name standardization in the record.
Tomaszewsky provided a discussion of automating the process of placename matching in his thesis on Aztec political geography. The heuristic (i.e., trial-anderror) method used in this study of matching place-names contrasts with
Tomaszewsky’s automated approach to the process of populating a gazetteer.
Tomaszewski wrote an interesting historical political geography of Aztec-era Mexico.
He created a complex query script based on pseudeocode and pseudeoSQL languages,
and matched modern place-names to historic names and their related variants. His
query asked whether any matches existed in the modern data set with the historic
names; if not, it checked the variant name data set; the variant spellings produced
many matches.93 In comparison, this study matched Villuga’s place-names over hours
of searching a modern road atlas and querying Google Earth for possible variant
spellings not found in the official atlas.
93 B ia To asze ski, The Re o st u tio of Azte Politi al Geog aph i the Tolu a Valle of
Me i o, (M.A. Thesis, State University of New York, University Center at Buffalo, 2005), 33.
58
Creating a greater density of place-names required further sources of names to
compare to Villuga’s record. The surveyors for the Relaciónes de los Pueblos de
España Ordenadas por Felipe II worked on behalf of King Philip (Felipe) II, in 1575,
identifying important historic, economic, and population information about towns and
villages in the Kingdom of Toledo. The survey, like the Domesday Book compiled
much earlier in William the Conqueror’s England in 1086, required answers to
searching questions from every populated place in Toledo Province, such as the name
of the site, what rare features existed in he province, natural and artificial; what
habitations it had; which rivers passed through it; what monasteries, abbeys, and holy
orders inhabited the place; what rents the ecclesiastics obtained from the place; and
many other political and economy-based questions. The record gives us historical
place-names for each location in Toledo, and provides a comparison to the placenames Villuga listed for the same region. A table of four place-name sources provides
a greater context for Villuga’s variants. It contains four sources, each from a different
time period. This table visualizes each source in an alphabetized format, and, through
database SQL queries, it lists the place-names from one location across temporal
spans. Some place-name values came back empty, likely due to changing
administrative boundaries over the temporal span of the table. Some replicate earlier
name variants, but most demonstrate mild to significant changes over time.
Comparing the place-names from various sources to each other took some preprocessing of the data. Not all sources contained georeferenced place-names, so a
comparison by location proved challenging. A query of all place-name fields from the
59
four sources provided a list from each source, which revealed the trends in placenames: repeats of place-names, minor and major name changes, and missing names.
Alphabetizing the lists formatted them for a SQL SELECT query in the Villuga
Gazetteer database to find all modern place-names from Cuenca Province:
SELECT MODERN_NAME FROM VILLUGA_GAZETTEER WHERE Part_of= 'Cuenca
Province';
Then, a SQL INNER JOIN query found all historic names from Villuga that reside
within modern Cuenca Province:
SELECT VILLUGA_GAZ.Part_of, PRIMARY_HISTORIC_MUNI_NAMES.Historic_name
FROM VILLUGA_GAZ INNER JOIN PRIMARY_HISTORIC_MUNI_NAMES ON
VILLUGA_GAZ.PK_PLACE = PRIMARY_HISTORIC_MUNI_NAMES.PK_PLACE
WHERE (((VILLUGA_GAZ.Part_of)='Cuenca Province'));
The historic names from Villuga relate to the gazetteer through its primary key
column. This relation provided a georeferenced historic name, and tied each historic
name to its present-day province. Villuga did not provide information about a given
place-name’s region or kingdom; he only gave data on the route for each populated
place. Comparing Villuga’s place-names to those from other contemporary sources,
such as the Cuenca Relaciónes topográficas, from 1575, provide some context. The
Cuenca relaciones topográficas only covered places within the jurisdiction of the
Bishopric of Cuenca, the spatial reference for the relaciones, which means, for
example, it did not cover relaciones from Requena, Guadalajara, or Utiel, which
bordered on the Bishopric of Cuenca. The modern province of Cuenca overlaps with
some historic parts of these other regions, and therefore contains some place-names
that once resided in them. This is a common problem in historical GIS. The Villuga
60
Gazetteer database query necessarily queried the historic place-names by their modern
province as the only solution, considering that the historic place-names data in the
database is not restricted by historic province, but relates to the modern place-name
gazetteer, which does restrict place-names by their given modern province.
4.2 Methods of League Analysis
The league, a unit of distance measurement, has many possible definitions.
Travelers may have used it in random lengths, or in specific patterns of use. An
analysis of the league distances in Villuga’s guide supplied an answer to the question
of how to define the league in a GIS for greater understanding of the Iberian
transportation routes. Determining a definition for the league facilitated its use in GIS
analysis of missing place-names. The analysis also extracted from the historical record
spatial patterns in the use of the league. The inquiry sought to find out whether certain
regions used specific unique lengths for the league, or if any patterns of use existed.
When patterns emerged, these informed the analysis of missing places. Scholars may
use these patterns to answer other transportation-related historical questions.
To create a model of Villuga’s routes in GIS, polylines connected all of the
place-names in a given route. Polylines consist of line features in GIS that contain a
series of nodes that connect to each other via lines. To create polylines that represent
each route, a GIS query selected each point in a particular route. Starting at the first
place in each route, a polyline connected to each point that represented a place-name.
These route polylines also function as the basis of a spatial analysis of the kilometer
61
distances per league. ESRI’s ArcGIS 9.3 feature class structure automatically
determines the length of polylines, which the Split Line at Vertices tool then split into
segments, providing the user with the distance of the polyline segments between each
locality.
The new resulting feature class, created from the output of the Split Line at
Vertices tool, contained a length field, providing a kilometer distance for each line
segment in the data set. The output of the Split Line at Vertices tool, and the leagues
Table 4.1. Kilometers per league.
data, converted to text files for use in MS Excel, provided analysis of Villuga’s league
distances. In Excel, added rows represent the starting termini. They provide place
markers for each route’s starting point. They are shown in Table 4.1 as ‘-999.’ Notice
62
that in the ‘Order_type’ field, these rows are identified as ‘Start’ or the starting
terminus of a route. Villuga could not record a valid distance measurement for the
starting termini of each route, since these places reside at the zero-distance (i.e.,
starting) location of each route. The SHAPE_length field divided by the League field
determined the kilometer length of each league Villuga listed in his record.
Spatial Autocorrelation
One of the advantages of incorporating historical GIS into a project is the
ability to search for spatial patterns not previously displayed by the data. 94 Measures
of spatial autocorrelation provide one means by which to test whether or not data
present a pattern of spatial clustering. The basic premise of spatial autocorrelation
builds upon Tober’s “First Law of Geography” in that phenomena that are closer
together in space are more likely to be similar than phenomena that are farther apart. 95
Goodchild described spatial autocorrelation as “a descriptive index, measuring aspects
of the way things are distributed in space, but at the same time it can be seen as a
causal process, measuring the degree of influence exerted by something over its
neighbors.”96 When one tests data for spatial autocorrelation, what could also be
termed as spatial dependence, the results of the test typically indicate one of three
94
Gregory and Healey, Histo i al GI“: st u tu i g,
; G ego a d Ell, Historical GIS, .
appi g, a d a alyzing geographies of the past,
95
Michael Goodchild, “patial Auto o elatio . Concepts and Techniques in Modern Geography 47.
Geo Books, Norwich, U.K., 1986. 3-4.
96
Ibid, 3.
63
possible outcomes. First, the data display positive spatial autocorrelation and therefore
similar phenomena are clustered. Second the data are randomly distributed and display
no spatial pattern. Or third, the data are uniformly distributed across the study area.
This final outcome indicates a form of spatial patterning, but not of spatial clustering
and dependence and is referred to as negative spatial autocorrelation.97
While a number of tests of spatial autocorrelation were available, the Moran’s I
statistic was used in this thesis. Utilizing this measure of global spatial autocorrelation
can identify the presence of clustering but not the location of the hotspots. In general
terms Moran’s I measures the correlation between events across space, thus a positive
score implies that league measurements are not spatially independent of one another.
The formal equation for the Moran’s I statistic is:
∑
∑
∑
∑
∑
̅
̅
(
̅)
where n is the number of observations, ̅ the mean of the attribute values (in this case
league ranges),
is the value at a particular location, and
all other locations.
is the attribute value at
is the spatial weight between i and j determined by the inverse
distances between i and j. The result of the index calculation usually falls along a
range from 1 to -1, where 1 is the highest degree of positive spatial autocorrelation,
97
Arthur Getis, B. “patial Auto o elatio , I Handbook of Applied Spatial Analysis: Software Tools,
Methods, and Applications, eds. Manfred M. Fischer & Arthur Getis (New York: Springer, 2010), 4; Ned
Levine, CrimeStat: A Spatial Statistics Program for the Analysis of Crime Incident Locations (v 3.2a).
(Ned Levine & Associates, Houston, TX, and the National Institute of Justice, Washington, D.C., 2009),
4.47.
64
random patterns return a number near zero, and -1 is the highest degree of negative
spatial autocorrelation.98
A variety of literature informed the choice of the Moran’s I statistic as opposed
to alternative tests for spatial autocorrelation. Getis argued that Moran’s I is the
leading statistic for both testing and measuring spatial autocorrelation, while many
other statistics only work as a test or a measure. Building upon Getis, Levine discussed
that the Moran’s I statistic and the Geary’s C statistic are similar, but that Moran’s I is
slightly more robust.99 Goodchild asserted that these two statistics are equally
satisfactory in most applications, but Moran’s I offers a more intuitive index than
Geary’s C.100 A good criterion for the choice between the two statistics, according to
Getis is the null hypothesis that each statistic offers. Moran’s I uses the null hypothesis
that related objects do not demonstrate covariance (vary in concert) in any consistent
way. Geary’s C uses the null hypothesis that related spatial objects do not differ from
each other in a consistent way. Moran’s I therefore more closely fits the desired
measure for Villuga’s record.
Average Nearest Neighbor Analysis (Nna) is a spatial statistic used in this
study to determined whether each kilometer-per-league distance range clustered in a
spatially significant pattern. Aldstadt explained that this statistic is not as advanced as
98
Getis, B. “patial Auto o elatio ,
283.
99
; Le i e, CrimeStat, 4.49; Aldstadt, B. . “patial Cluste i g,
Levine, CrimeStat, 4.48.
100
Good hild, “patial Auto o elatio ,
.
65
Ripley’s K-function, which provides a multiple-scale calculation.101 However, the Nna
statistic fits well the needs of this study. The Nna algorithm analyzes only distance,
not taking attributes into consideration. Each distance range isolated the various values
into one attribute, and was separately analyzed, making the Nna statistic a good match
for the simple data. Moreover, the desired test result needs to answer a question
dealing with a first-order analysis, describing the intensity of clustering, rather than a
second-order test, which also evaluates “interaction or dependence” between points.102
The basic calculation determines the distance from a point i, to all other points j, in the
data set, and chooses the minimum distance from each i to the nearest j. The algorithm
then sums and divides these minimums by the total number of points (n).103 This
calculation finds the Nearest Neighbor distance. Then an algorithm determines what
the average distance would be under random conditions, the Mean Random Distance,
in other words, a lack of clustering. The calculation then creates a ratio of Nearest
Neighbor distance, d (Nna), per Mean Random Distance, to create the Nearest
Neighbor Index. If the index is less than one, the data cluster spatially, forming hot
spots; if the index is greater than one, the data are dispersed. The Nna test determines
only that the “average nearest neighbor distance is significantly different than what
would be expected on the basis of chance”.104 The distance measurements for the
101
Aldstadt, B. “patial Cluste i g,
102
Ibid, 287.
.
103 Levine, CrimeStat, 5.3.
104
Ibid, 5.4.
66
league in Villuga’s record fit into six ranges. A Nna analysis determined how much
each range clusters together.
The spatial autocorrelation analysis of Villuga’s leagues combined spatial and
attribute patterns that determined whether the data are clustered, dispersed, or
produced a random pattern. To prepare the league distances between each place for the
spatial autocorrelation analysis, the Feature to Point tool in ArcGIS 9.3 derived
centroids for each polyline in the data set. The Moran’s I global spatial
autocorrelation tool produced an analysis of all centroids in the data set, combining
coordinates with attribute data to determine the existence of possible hotspots. The
Average Nearest Neighbor Analysis (Nna) tool does not consider attributes in its
computation. The first step in this process highlighted all places with the same
attribute to prepare them for the Nna tool. This tool analyzed all centroids falling into
each of the six kilometers-per-league distance ranges, one range at a time, to determine
the degree of regional clustering within each range.105
4.3 Methods of Interpolation: Traditional GIS Methods and Fuzzy Rules
This section adapts some GIS methods to the analysis of unknown places. It
considers ways in which this study can apply GIS-based interpolation methods, which
105
Fo fu the e pla atio of these tools, see E“RI’s A GI“ . Desktop Help o li e at
http://webhelp.esri.com/arcgisdesktop/9.3/index.cfm?id=1798&pid=1790&topicname=Feature_To_P
oint_(Data_Management), for Feature to Point tool;
http://webhelp.esri.com/arcgisdesktop/9.3/index.cfm?id=2150&pid=2146&topicname=Spatial_Autoc
orrelation_(Morans_I)_(Spatial_Statistics), fo Mo a ’s I, a d
http://webhelp.esri.com/arcgisdesktop/9.3/index.cfm?id=2147&pid=2146&topicname=Average_Near
est_Neighbor_(Spatial_Statistics), for Average Nearest Neighbor Index (accessed 12 April 2010).
67
identify a possible value for locations where no data exist. Interpolation algorithms
extrapolate these data from a set of control points where the desired data do exist. GIS
conditional statements can assign weights to sectors of a bounding box, based on local
features, which will increase or decrease the likelihood of the unknown place being
located within that region, such as the presence of a river or other source of water. The
condition statement can utilize such barriers existing inside the bounding box as
elevation, water location, proximity to other likely road locations, and the likelihood of
a certain measurement for the league. The interpolation tool would use these features
as parameters for sample points. The inability to use a combination of the historian’s
rules and the vague historical data for computation in GIS, which relies upon crisp
data, produces the major problem with this approach. Also, the study’s lack of more
detailed physical geographic data limits this analysis to a focus on Villuga’s general
league distances between places.
The following methodology answers the need for determining the “very likely”
location of the unknown place-names in Villuga’s guide, and adding an interpolated
surface that defines a continuum of likelihood of location from “very likely” to
“somewhat likely.”
Figure 5.4 provides a visualization of the following process: first a 4.2
kilometer buffer around the known places gave the approximate likely distance to the
unknown place, Luna, from the known places, Ariza and Alhama de Aragón, both of
which Villuga recorded a one league from Luna. The Buffer tool placed buffers around
the “very likely” location points. A straight line denoted the most likely path between
68
the known places.106 The intersection of the buffer and the path line marked the most
likely location of the unknown place, based on Villuga’s general distances. Two “most
likely” points exist, each representing the same unknown place, based on the distance
of 4.2 kilometers from each known location. The Multiple Ring Buffer tool then
placed several buffers around the “very likely” location points, at 0.5, 1, 1.5, 2, and 2.5
kilometers, each of which represented a zone of likelihood; the smaller the distance,
the more likely the location.
In the Polygon to Raster tool, these buffers provided the input parameters for a
raster layer, with a field for weighting each buffer by the above distances. The Extract
Values to Points tool transferred the weights from the raster layer to a layer of
centroids in each buffer. This tool required two parameters: the raster that held the
weight values from each of the buffer zones, and a vector point layer which received
the weight values from the raster layer. Using these weighted points, the Kriging tool
created an interpolated surface which indicated a continuum of membership in each of
the levels of likelihood for the location of the unknown place-name.
Spatial statistics consist of those statistics that measure the variation in spatial
phenomena. Geostatistics are those spatial statistical methods that measure continuous
geographic features. Geostatistics treat an attribute variable as though it were random.
Any attribute, z(x), at a specific geographic location, is an instance of the property,
Z(x), or the random variable. Geostatistics assume stationarity, or that the mean µ and
106
This path is a st aight li e as the o flies et ee the t o lo atio s. Whe additio al data
related to the directionality of the path or barriers between the two places become available this line
might be altered to better reflect reality.
69
the variance,
, of a given random variable is the same everywhere. Covariance, or in
this case, how much two instances of the variable change together, is measured solely
by the distance between the two instances of the random variable.107
Oliver argued that kriging offers a superior statistic to such alternatives as
splines, inverse distance measures, and trend surfaces, because it takes into account
errors of prediction, for which the later techniques do not. The ESRI version 108 of the
equation for kriging defines a semivariogram, following Matheron’s Methods of
Moments (MoM) estimator,109 which interpolates the z-values at locations in between
the features with known attributes at various lag distances:
∑
where h is a lag distance, or a set distance from any of the points being
measured; Z is the z-value, or attribute variable;
represents each location on the map
with a known z-value, providing an index for each instance of the random variable,
. The semivariance scores are determined by subtracting the z-values at each
location,
from the sum of the z-values plus the lag distance,
squaring them:
and
These scores are summed and then divided
107
For a more in-depth e pla atio of geostatisti s a d k igi g, see Ma ga et A. Oli e , The
Va iog a a d K igi g, I Handbook of Applied Spatial Analysis: Software Tools, Methods, and
Applications, eds. Manfred M. Fischer & Arthur Getis (New York: Springer, 2010), 319-320.
108
E“RI A GI“ . Desktop Help, Ho K igi g Wo ks,
http://webhelp.esri.com/arcgisdesktop/9.3/index.cfm?TopicName=How Kriging works (accessed
August 25, 2010).
109
Oli e , The Va iog a
a d K igi g,
.
70
by twice the sample size, n, to determine the quantity of paired locations at a given lag
distance, h.110 The algorithm, run at multiple lag distances, provides the variance at
each lag distance, which is plotted on a semivariogram. The semivariogram provides
predicted z-values at the unmeasured locations in-between each point of known zvalue.
Maximizing GIS Methods through Adaptation of Fuzzy Rule-Based Modeling
Zadeh claimed that there exists no “standard probability theory nor any
methodology which does not employ the machinery of fuzzy information granulation
[that] can come up with a machine solution [i.e., a solution relying upon the
calculations of a machine].”111 The interpolation algorithms available to GIS have
great merit in locating the unknown places in sixteenth-century Spain, but the
methodology of fuzzy rule-based modeling gives a more robust tool for the context of
Villuga’s imprecise and uncertain distance and place-name data, which rely upon a
“humanistic system” rather than a “mechanistic system.” In the context of humanistic
systems, fuzzy rule-based modeling expands upon what GIS interpolation methods
offer. Moreover, some of Zadeh’s concepts apply to GIS interpolation methods.
Converting fuzzy if-then statements to Map Algebra condition statements allows the
researcher to use GIS to anlyze the historian’s rules to create a raster surface that
determines the most likely location of the unknown place-name. Our fuzzy if-then
110
Ibid, 322.
111
Lotfi A. Zadeh, To a d a Theo of Fuzz I fo atio G a ulatio a d its Ce t alit i Hu a
Reaso i g a d Fuzz Logi . Fuzzy Sets and Systems 90 (1997), 126.
71
statements consider other likely geographical features to expect around Villuga’s
missing places, but access to the data needed for running the resulting GIS conditional
statements is not available for this study.
The following fuzzy if-then statement provides an example:
If Villuga’s league is measured from [some number of km] to [some number of km]
and [location i] is between [some number of km] and [some number of km] of
[location h] and [location i] is between [some number of km] and [some number of
km] of [location j] and [location i] is within [some number of meters] of [river A],
then it is very likely that [location i] is located at [some fuzzy location].
The researcher can convert this statement to Map Algebra format, once a set of
number values defines the parameters of the statement. A historical data source would
list a range of possible data values to use as input parameters. The GIS analysis would
seek for geographic locations with attributes that satisfy the condition statement and its
parameters (fuzzy rule-based modeling refers to this as ‘inference,’ or executing the
conditional statement in order to determine, in this case, the most likely location). 112
The numbers chosen represent arbitrary choices for the sake of example. One way to
determine these input numbers could rely upon ranges of data previously derived
through GIS analysis of historical sources, such as this study’s analysis that
determined ranges of kilometers per league. Using data ranges derived from the
historical sources, the Map Algebra condition statement would read as follows:
con(km to unknown place1 >= 5.5 AND km to unknown place1 <= 6.6 AND km to
unknown place2 >=5.5 km AND km to unknown place2 <= 6.6 AND
location_i_river >= 10 [meters] AND location_i_river <= 750 [meters],1,0)
112
Zadeh, To a d a Theo
of Fuzz I fo
atio G a ulatio , 115.
72
In the condition statement’s syntax, the con operator denotes a condition statement; the
statement ‘km to unknown place > = 5.5 AND km to unknown place < = 6.6’ (which
measures the distance range from 5.5 kilometers to 6.6 kilometers from a known
place), provides the condition for which the algorithm tests. The output gives the 1
directly after the test condition if the statement is true, and the 0 (which comes directly
after the 1) if the statement is false, giving it no weight in the output. Running the
condition statement would produce a weight value in each cell, which denotes whether
a given cell fulfills the tested conditions. Extracting the weights from the raster to
vector points would provide parameters for interpolating the location of the unknown
place, using a interpolation algorithm such as Kriging.
4.5 Methods of Population data: Transferring Data Between Gazetteers
In order to further contextualize Villuga’s routes, the Villuga Gazetteer
incorporates population data from the same period. Some of these population data
were already available in gazetteer form. Researchers of place-names may desire to
use gazetteer data from various sources to form their own specialized gazetteers.
Frequent variations exist between sources in both place-names for locations and
coordinates for those locations. For example, two online gazetteers may provide placename and coordinate data for one given place that are likely not standardized, and thus
will contain minor variations between each other’s record for any one place.
73
In an earlier study, a gazetteer called the Censo Gazetteer113 mapped many
place-names from the 1591 Censo de Castilla, along with population data, to their
geographic coordinates. Using the Censo Gazetteer’s data in the Villuga Gazetteer
required matching them by each record’s longitude and latitude coordinates. The two
gazetteer data sets sometimes consulted different sources for coordinates for the same
place. Madrid in one gazetteer may have coordinates that came from the Alexandria
Digital Library Gazetteer, and in the other gazetteer it may have coordinates which
came from the US National Geospatial Intelligence Agency’s GEOnet Names Server
(GNS). Both data sets were georeferenced, and therefore GIS provided a suitable
format for combining them. The Near tool in ArcToolbox provided the solution, by
comparing one gazetteer to the other, identifying which coordinates in the Censo
Gazetteer nearly matched coordinates in the Villuga Gazetteer, by location, proving
they likely represented the same place. Each matching set of coordinates also had a
place-name.
The Near tool in ArcToolbox has a query function that sets a ‘near’ threshold.
Upon running the tool at a maximum matching distance of 100 meters, any two points
in the two gazetteers with similar coordinates resided less than 5 meters apart, and
most resided at less than 1 meter. Therefore, all of the matched places represented
duplicates of the same populated place. From the Near tool’s output, the matches were
exported to a .dbf file. A query of the Near tool output file selected all of the matches,
or the place-names that existed in both the Villuga Gazetteer and the .dbf file. The
113
Robert Hibberd, Censo Gazetteer, [Unpublished gazetteer], 2008.
74
Switch Selection tool then highlighted all the place-names that remained in the Villuga
Gazetteer that did not match place-names in the .dbf file.
The above process created two tables: one table with all of the place-names
from the Villuga Gazetteer for which the Censo Gazetteer had matching place-names
and their coordinates, entitled Censo Matched; and another table with all of the placenames for which the Censo Gazetteer did not match the Villuga Gazetteer, entitled
Censo Unmatched. Neither table yet contained the population data from the Censo
Gazetteer; they only contained place-names. The Modern Names fields in both tables,
which contained identical modern place-names, provided an ID field which tied the
Censo Gazetteer to the Censo Matched table. A SQL RIGHT JOIN statement in MS
Access combined the matched names to the population data:
SELECT * FROM Censo_Gazetteer RIGHT JOIN Censo_Matched ON
Censo_Gazetteer.MODERNNAME=Censo_Matched.MODERN_NAME;
This query connected the Censo Gazetteer and the Censo Matched table, and
retrieved all of the place-names from both tables, including all place-names from the
Censo Matched table that did have population data in the Censo Gazetteer, as well as a
list of those names in the Censo Matched table which resided in the Censo Gazetteer
without population data. The RIGHT JOIN identified which of them did have
population data in the Censo Gazetteer, and which did not.
In ArcGIS, a SQL SELECT statement highlighted all of the place-names in the
Villuga Gazetteer that the Censo Matched table also contained:
75
SELECT * FROM VillugaGazetteer WHERE NEAR_DISTANCE > 0 AND
NEAR_DISTANCE <= 5
The Villuga Gazetteer (in feature class format, a file format made specifically for
ArcGIS that contains attribute and spatial data) contained the output fields from the
Near analysis. Once the SQL SELECT query highlighted all entries that matched its
conditions, the Switch Selection tool selected all place-names in the Villuga Gazetteer
that the Censo Matched table did not list. The Censo Unmatched table, derived from
the last SQL query, did contain all of the place-names that the Switch Selection tool
highlighted (i.e., all place-names in the Villuga Gazetteer which the Censo Matched
table did not list). This list of place-names had no population data available in the
Censo Gazetteer. This data set was exported to MS Access. Population data from the
hard copy of the 1591 Censo de Castilla was added to each record in the Censo
Unmatched table (i.e., the records from the Villuga Gazetteer that the Censo Gazetteer
did not contain). Each table required an identical schema, or database format.
76
CHAPTER 5: DISCUSSION OF RESULTS
Matching the modern variants of Villuga’s historical place-names produced a
georeferenced network of transportation routes for the Iberian Peninsula in the
sixteenth century. Converting league distances to kilometer distance ranges provided
a tool for the analysis of the location of missing places in Villuga’s record, using fuzzy
rules to make the historical data compatible with GIS. Using GIS tools, multiple
gazetteers were merged to attach population data to the transportation network.
5.1 Gazetteer Results
Some regions of modern maps of Spain use the regional language variant of a
place-name rather than the Castellano counterpart that Villuga recorded in his
handbook, which adds difficulty to the process of place-name matching. Table 5.1
provides some examples of the variations in place-names from 1546 to the present
day. Villuga did not represent Valenciano and Gallego in his handbook, likely in part
because he published his book in Castile. The ascendance of Castellano to the status of
a literary language in the twelfth and thirteenth century preceded its emerging cultural
dominance, which grew during Castile’s growth in political power in the fifteenth and
sixteenth centuries.114 Villuga’s use of Castellano for his handbook causes some
difficulties with the identification of place-names, since he did not represent them in
their native language. For example, Villuga changed Moixent, near Valencia, to
114
Téofilo Ruiz, Spanish Society (New York: Longman, 2001), 12, 15-16.
77
mojent; Molins de Rei became molin derreche. The 1834 Reorganization dramatically
altered even the Castellano place-names. The Liberals wanted to differentiate between
the many places with the same or similar names. Additions to the place-names, often
taking the form of an extra descriptive word at the end of the name, reflected regional
elements of landscape, or the new political hierarchy formed by the changed
boundaries (e.g, the name “Chinchilla,” became “Chinchilla de Monte Aragón”). This
temporal variability between Spain’s sixteenth-century place-names and its presentday variants presented a challenge to forming the Villuga Gazetteer.
Table 5.1 provides a few examples of the differences between the sixteenthcentury and modern place-name variants. The modern locations of Villuga’s placenames appear in figure 5.1, a map derived from the Villuga Gazetteer. This map
reveals much about Villuga’s data set that he did not explain in his narrative format. It
shows clusters of routes and populated places, whether the regions connected in closeknit or far-flung patterns. One can easily discern which regions were heavily traversed
or sparsely traveled. The map demonstrates how the topography of the peninsula
clearly constrained the transportation network. Studies of the Castilian economy may
use this data as part of a spatial data infrastructure to contextualize the historical
documents about trade and transportation in the heart of the Castilian global
monarchy. Hot spots on the map provide likely hubs of economic activity and power.
78
Figure 5.1. Map of Jua Villuga’s Routes, f o
Mode
“ou es.
The modern variants of Villuga’s place-names did not always appear perfectly
matched by both place-name and expected location. Attaching confidence levels to
each modern variant provided one solution to the problem. The various confidence
levels are discussed in depth in chapter 4. Figure 5.2 provides a map of the resulting
confidence levels. The place-names with the greatest confidence level appear in the
highest category,
79
Figure 5.2. Map of Confidence levels.
labeled with a 1; the place-names with the lowest confidence level appear with a 4.
The place-name Gálvez provides an interesting example of the usefulness of storing
the many variants of a place-name by location in a gazetteer, and comparing them
across their temporal changes, giving them a greater density of variants to analyze.
80
The surveyors for the Toledo relaciones, who recorded the responses to the king’s
inquiry from each municipality’s representatives, wrote it down as la villa de Galves,
and claimed that the people of the place knew of no other name variant, but the name
appears in the record as also Galvez. Another place, Hormigos y La Higuera Del
Campo, started as two separate places, but became “one people.” Peña Aguilera is also
called las Ventas con Peña Aguilera, because two ventas, small roadside inns, had
resided in this area.
MODERN_NAME Villuga_name
Salses-le-Château salsas
Perpignan
perpiñan
Perpignan
perpiñan
Le Boulou
albolo
Le Perthus
el pertus
La Jonquera
xunqueras
Figueres
figueras
Bàscara
abascara
Gerona
girona
Hostalrich
astarlíd
Sant Celoni
sancelonij
Llinars del Vallès linas
La Roca del Vallès larroca
Moncada I Reixac moncada
Table 5.1. Modern and historic
names from Villuga Gazetteer.
Puebla de Almoradiel was new, perhaps 300 years old, in 1575 when the
king’s agents conducted the survey, but the inhabitants did not know who founded it.
Talavera de la Reina has no less than ten variant names, alluding to a long and varied
81
history from antiquity, encompassing Roman and Muslim occupation and other
changes in its history. The recorders of Valdeverdeja stated that this place had never
had another name, but the record lists it also as Valdoverdeja. The people of
Villanueva de Alcardete gave the place its name after they abandoned Alcardete a
short league from the new place, due to the old settlement’s very close proximity to
the ribera (bank) of Ciguela.115 Precision in recording place-names somewhat escaped
the recorders of this survey.
Including the relaciónes de Cuenca into the Villuga Gazetteer involved more
difficulty in filtering the data: the analysis revealed duplicates of many names,
particularly Gil García, which may or may not represent the same place. In comparing
the modern place-name variants to those from Villuga, many major and minor changes
appear. El Fresno became El Frasno; this minor change can return a null query from
some gazetteer search engines. Vaydes became Baides; Antepoçuelos became
Ciempozuelos. Guaraf became Garraf; la torre dembara became Torredembara; las
cuevas became Cuevas de Vinroma. These place-names, when georeferenced (i.e.,
matched to a geographic location) easily matched with their historic variants.
Table 5.2 provides a sample of the side-by-side comparison of each placename through the various sources, representing dates from 1546 to the present, for
115
C. Viñas & R. Paz. Relaciones Histórico-Geográfico-Estadísticas de los pueblos de España hechas por
iniciativa de Felipe II: Reino de Toledo. Vols. 1-3. (Madrid: Instituto Balmes, de Sociología; Instituto
Juan Sebastian Elcano, de Geografía; Consejo Superior de Investigaciones Científicas, 1963), 416-419,
467, 212, 242, 444, 647, 729.
82
Cuenca. If the historical names fell outside of the modern province, they did not
appear in the result.
VillugaCuencaNames
NamesCuencaRelaciónes
HistNames1834
Modern_name
Alarcon
Alarcón
Valadiego
null
Alarcon
Albaladalejo del
Cuende
Alarcón
Albaladejo del
Cuende
Almodovar
alcaçar de huete
Almodóvar
Alcázar del Rey
Almodovar del Pinar
Alcazar del Rey
Almodóvar del Pinar
Alcázar del Rey
arcuaz; arquas
null
Arcas
Arcas
Alguisuellas
null
Arguisuelas
Arguisuelas
barchi; barchin
Barchín del Hoyo
Barchin del Hoyo
Barchín del Hoyo
vilinchon; valenchon
Belinchón
Belinchon
Belinchón
Buenache
Buenache
Buenache de Alarcon
Buenache de Alarcón
Campillo
null
Campillo de Altobuey
Campillo de Altobuey
Cardenete
null
Cardenete
carraschosa d huete
null
Carrascosa del Campo
el castillo
cervera
chillaron; chilaron;
gillaron
Castillo de Garcimuñoz
null
Castillo de Garcimuñoz
Cervera
Cardenete
Carrascosa del
Campo
Castillo de
Garcimuñoz
Cervera del Llano
null
Chillaron de Cuenca
Chillarón de Cuenca
cuenca
alcañavale; cañavete
el histo; el hito
Cuenca
El Cañavate
null
Cuenca
El Cañavate
El Hito
Cuenca
El Cañavate
El Hito
el pedernoso
El Pedernoso
El Pedernoso
El Pedernoso
el provencio
El Provencio
El Provencio
El Provencio
fuentes
null
Fuentes
Fuentes
agua valdon; guavaldon
Gabaldón
Gabaldon
Gabaldón
Table 5.2. Density of a Sample of Place-names for Cuenca Province.
The MS Access query ensured that the place-names in the
‘VillugaCuencaNames’ matched the names in the ‘Modern_name’ field. Place-names
in the other two fields matched to the place-names in the ‘VillugaCuencaNames’ and
83
‘Modern_name’ fields based on their similarity to place-names in those fields, rather
than through explicitly georeferencing them separately.
The resulting geodatabase provides users with the modern remnants of
Villuga’s routes, and a picture of which place-names have disappeared since the
sixteenth century. It provides query functionality for searching the geodatabase by
specific place-names or their routes. It allows users to visually search and analyze
sixteenth-century Spanish and Portuguese transportation patterns through spatial
queries. The outline of best practices in gazetteer design aid the gazetteer designer in
avoiding common problems associated with gazetteers, such as ambiguity between
similar place-names from various regions of a country, or from various time frames
(e.g., Villuga’s book recorded the name ‘monreal,’ multiple times, which appears in
more than one province across Iberia).
Place-name changes over time made the use of database technologies more
difficult. Many place-names from the density analysis (table 5.2) have little temporal
variation, such as Alarcón. Albaladejo del Cuende, and Almodóvar del Pinar have
significant toponimical changes, but still resemble the historic variants when compared
to them. Belinchón and Arcas both exhibit minor variations between the modern and
1546 variants that do not present a major stumbling block when evaluating place-name
matches by hand. However, they do pose a problem for exact searches in databases
and online gazetteer server clients. Shifting administrative boundaries influence a
place-name’s temporally dynamic administrative ontology. Google Earth and Google
Maps can run searches for names similar to the one provided by the user. These
84
programs found many modern variants of Villuga’s historical place-names. For
example, if the user types the place-name, ‘Gabaldon,’ or ‘agua valdon,’ the query will
return ‘Gabaldón,’ and other similar names.
In the 1834 Reorganization, Spanish place-names changed significantly. Table
5.2 presents several examples: alcaçar de huete became Alcázar del Rey. The
inhabitants used this name as early as the 1575 Relaciónes topográficas survey of
Cuenca. According to the survey, Alcázar del Rey belonged to the jurisdiction of
Huete, and to the realengo (royal jurisdiction).116 The Reorganizers changed buenache
and almodovar to Buenache de Alarcón and Almodovar del Pinar, standardizing them
to produce a name unique to all of Spain. Caraschossa de huete became Carrascosa
del Campo. The survey stated that in 1575 it belonged to the Crown, and had once
resided within the jurisdiction of Huete. The people of Huete purchased Huete and all
places within its jurisdiction from its owner, Don Pedro Buil, and “gave it to the royal
crown of the king of Castile.”117
A GIS facilitated the analysis of the place-names of Villuga’s record and their
variants across time and various historical sources. More complex studies will benefit
from integrating the common elements of the gazetteer standards discussed, such as
time stamps. The use of multiple historical sources for place-names, to create a greater
density of place-names, provided the best, most common historic variant of a given
116
Eusebio-Julián Zarco-Bacas y Cuevas, Relaciones de Pueblos del Obispado de Cuenca.
(Cuenca, Spain: EXCMA. DIPUTACION PROVINCIAL DE CUENCA, 1983. Original publisher:
Felipe II, Rey de España 1575), 535.
117
Ibid, 212.
85
place-name. While the efforts of the nineteenth-century Spanish Liberals to
standardize Spain’s place-names had a vast impact on those names, the place-names
continue to exhibit the effects of the powers of Spain’s cultural and political
polycentrism, decentralization, and linguistic variability. Modern Spanish place-name
variants exhibit the effects of both standardization and linguistic regionalization.
Searching for the modern variants of Villuga’s place-names revealed many historic
human places that modern maps do not record. The method of locating these unknown
place-names involved utilizing Villuga’s vague league distance data to produce
kilometers-per-league distance ranges, through cartographic classification, and using
these ranges as an essential part of text-based fuzzy rules, adapted to GIS condition
statements.
5.2 Results of League Analysis
The Moran’s I analysis of the path centroids revealed that the global spatial
autocorrelation pattern is neither dispersed nor clustered. The ArcGIS version of the
tool reported a Moran’s I index of 0.13, with a z score of 0.44. Therefore, no
significant clusters of kilometer distances per league exist across the entire data set.
An Average Nearest Neighbor analysis (Nna) tested whether a larger spatial scale
revealed spatial autocorrelation in the kilometers-per-league distance ranges (Table
5.3). It revealed that a high level of regional clustering existed in all ranges except the
last, which demonstrated a highly dispersed pattern.
86
Ranges in Km per
Average Nearest Neighbor
Z-
P-Value
Range
League
Index
Score
0 – 3.4
0.59
-9.208
0.000
138
3.41 – 5.12
0.55
-19.033 0.000
496
5.13 – 6.8
0.616
-18.9
0.000
661
6.81 – 10.7
0.64
-13.355 0.000
378
10.8 – 18.06
0.76
-3.59
0.000331 63
18.07 – 43.53
1.756
3.54
0.000397 6
Counts
Table 5.3 Average Nearest Neighbor Analysis of League Distance Ranges
None of the ranges had a center; they all scattered in regional clusters
throughout the Peninsula. This makes the Nna quite useful to this study. While other
statistics are based upon on a known geographic mean (i.e., the average location of all
the points being considered), the Nna is not, and therefore fits the needs of the
regionally-based data clusters.
The index values less than one in the Nna index signify clustering of the
distances in a given range, and the numbers above one denote dispersal of those
Table 5.3. Average Nearest Neighbor Analysis of league distance ranges.
distances. Therefore, the place-name locations in all distance ranges but the last
demonstrated significant clustering. The z-scores for all but the last range signified
that the distances in each range clustered significantly below the mean distance for
each range. The P-values indicate a significant unlikelihood that these patterns existed
merely due to randomness. The range from 4.2 to 5.6 kilometers per league, the
87
distances most frequently cited in historical sources, presented no major classification
breaks, and these distances fit into the same range. The ranges with the highest counts
appeared on the more level elevations; the higher of the two, 5.13 – 6.8 kilometers per
league, appeared to cluster in slightly higher elevations than those of 3.41 – 5.12, the
lower of the two. The last two ranges had the lowest counts. Analysis of the
kilometers-per-league ranges provides the historian with important regional patterns of
measurement for the Spanish league. These ranges provide more data for locating the
missing place-names from Villuga’s guide by giving a more concrete format for the
measurement of Villuga’s league distances, which GIS can utilize. It also provides
economic historians with measurement tools for use in determining the cost of
transportation, which is an important aspect of an economy, in various parts of the
Iberian Peninsula.
88
Figure 5.3. League distance ranges.
89
5.4 Results: GIS and Fuzzy Rule-Based Modeling
The interpolation process outlined in this study revealed that the 4.2 kilometer
buffer chosen to represent Villuga’s league distance did not accurately measure
the distance to Luna from its neighbors. Figure 5.4 demonstrates the map that resulted
from the interpolation analysis. While this analysis proved helpful, it still relied upon a
crisp data format, incompatible with the data Villuga supplied in his record. The
combination of fuzzy if-then statements with conditional statements improved upon
the use of interpolation tools. The introduction of a domain expert who makes
deductions based on historical data about how to weight certain parts of the condition
statements (or the points used as a foundation for the interpolation weights of a
Kriging analysis) adopts some principles of fuzzy rule-based modeling to the process.
The if-then statement and the GIS conditional statement have compatibility, the latter
built upon the former, once the analyst chooses a range of data for each parameter of
the conditional statement, as demonstrated in chapter 4. The output of the GIS analysis
provided an interpolation surface for the unknown place that generalized the exactness
of the buffers. The points received a weight based on which buffer they fell into, and
the end effect overrode the buffers' exactness, at least to a minor extent. The resulting
interpolation surface presents some difficulties: it did not have the immediate breaks
between categories; but, it relied upon precisely defined buffers, and a less-than
random location within each buffer zone. The results illustrated that travelers likely
used a distance greater than 4.2 kilometers for the league at the location of Luna.
90
This exercise also demonstrated that the historian can adapt fuzzy rule-based
modeling to the use of GIS as an analysis tool, and also to illustrate how fuzzy rulebased modeling can modify ‘crisp’ data analysis to add more of the human (i.e., fuzzy)
decision process. The input parameters of kriging and other ‘crisp’ interpolation
methods rely, however, on ‘crisp’ data. Sharp grades of membership will exist
Figu e . . I te polatio a al sis of a u k o
pla e’s lo atio .
between each level of classification on the map; each parameter for a spatial location
resides either fully in one class or fully out of it. However, the use of ranges in the GIS
condition statements aids in calculating with vague data. The kilometers-per-league
91
ranges provide one useful example from this study of the applicability of ranges to
GIS analysis of vague data. Fuzzy rule-based modeling, however, allows for each
parameter for a spatial location to reside partially in one class and partially in another
class. This study also produced an example of how to convert an if-then statement,
written by the domain expert, into a GIS condition statement (page 72).
5.5 Results of Population Analysis: Spanish Urban Patterns Prior to Royal
Madrid’s Impact
The spatial analysis of two gazetteer sources combined their place-names and
other data by comparing each place’s geographic coordinates, and by running queries
in a relational geodatabase structure to identify which place-names from the Villuga
Gazetteer matched those in the Censo Gazetteer table. The process of making the
population data useable in a GIS consisted of matching the Villuga Gazetteer’s placenames to a previously-existing gazetteer, which contained population data for some of
Villuga’s places; then, determining which names still needed population data, and
recording that data from the hardcopy text. To cartographically visualize the
population data, a relational join in ArcMap appended the population data to the
gazetteer table, to facilitate queries between the two tables. The resulting new census
data table, formed by SQL queries, holding population data for each place-name in
Villuga’s record, allowed visualization and analysis of town-level data from the 1591
Censo de Castilla. A kernel density and cartographic analysis of the population data
92
provided a look at which places or significant clusters of places exerted the most
influence in the economy as population centers during Villuga’s time.
The lands of the Castilian Crown in sixteenth-century Iberia had a settlement
pattern that fit the description of a polycentric urban region, illustrating how its
dispersed patterns precluded a strong centralized system, but rather created a
pluralistic format of economic, political, and religious interconnectedness and
interdependence.
Unlike Spain’s later urban concentration in Madrid, many sixteenth-century
Spanish cities had similar levels of dominance through population levels, which also
somewhat denote their economic pull. These patterns began a slow change, to varying
degrees, when in 1561 the Castilian Crown transformed Madrid into the Royal
Household and Court, and the first permanent location of their government, although
they moved it to Valladolid again for a brief period. The Court’s move to Madrid
would likely have the least impact on Iberia’s religious traditions, which depended
more upon the institutions of the Catholic Church than upon the Court. Population data
can act as a surrogate for administrative and economic activity in a place, revealing
something about its centrality in an urban system. This study’s GIS constructs,
according to the available sources, the Castilian population prior to Madrid’s
dominance to create a picture of what the Spanish economy and urban system looked
like before Madrid became the location of the royal court. The cartographic analysis of
the Castilian population (figure 5.1) reveals many polycentric urban regions
93
throughout Castile, composed of clusters of urban centers roughly equal in economic,
political, and religious power.118
Narrative Descriptions of Spain’s Polycentric Urban Structure
Some of the routes of Spain as described by Juan Pedro Villuga span the
Peninsula, traveling, for example, from Alicante to Santiago de Compostela, from
Valencia to Lisbon, or from Leon to Sevilla. The map of all these routes, however,
reveals a regionally-based travel system, with a few important interregional connecting
routes, a few of which connected to the international travel system, such as sea routes
at San Sebastian, Lisbon, Sevilla, Alicante, Valencia, and Barcelona. The connecting
routes between south and north, east and west, provided the lifelines between separate
population zones, with very few people in between these zones. The routes in
geographical context reveal the considerable constraints placed by the topography and
climate of Spain upon the mobility and interconnectedness of the Peninsula’s various
regions. As stated above, these constraints influenced the degree to which interregional
trade flourished.
The discussion focuses on the regionalized nature of the system of routes
Villuga described, and how the routes connected to each other. It explores the issues
travelers dealt with along their routes, which demonstrate the polycentric and
dispersed nature of the population along the routes. Specifically, it explores the
118
Gladys R. Dobbs, The I dia T adi g Path a d the Colo ial “ettle e t De elop e t i the No th
Caroli a Pied o t. (PhD diss., University of North Carolina-Chapel Hill, 2007), 6.
94
importance of bandits and ventas to our consideration of Spain’s sixteenth-century
polycentric urban system.
The map of Villuga’s routes produces a picture of a regionally-based system,
with some important long-distance connections. Valladolid to Sevilla, Leon to Sevilla
(which contains several other intermediate routes), Granada to Cuenca, and Alicante to
Santiago (and its intermediate routes) provided the only routes connecting southern
Spain with the northern regions. The northern routes varied greatly in length and
population. The connecting routes between southern and northern Castile provided the
lifelines between separate population zones, with a very sparse population in between
these zones. The main Castilian population density zone ended south of Toledo, with a
considerable sparsely-populated stretch before one reaches the southern, Andalucian
population density zone.
Villuga provided limited routes to northwestern Iberia, as well. There exists
only one road each that connects Castile to Galicia and Asturias. Only four roads
connect Castile to Portugal. Only two roads connect eastern Spain to Andalucia, as
they emerge from other more elaborate networks in the east. Only two roads connect
Andalucia directly to Castile. Toledo and Valencia connect only via one route.
The routes in the northern regions display the geographic constraints upon
travel in these areas. The Cantabrian Mountains had one main route pathway going
east to west; the other northern region routes provided the necessary routes going
through the mountains north to south. The routes into the Cantabrian Mountain region
made only 4 tracks. The many routes through Burgos merged into three of these. As
95
one traveled north, starting at about 42.8 degrees latitude, the mountains began and the
major east-west routes ended. One path took some east-west routes through this region
that ran just south of the 42.8 degrees latitude demarcation line, and then west to
Santiago de Compostela.
One should not overstate the limits of interregional networks. Some of the
routes-if you look at the long-distance patterns, rather than focusing on all of the
smaller routes which overlap the longer ones-did cluster at a peninsular scale; they ran
through set paths of long distance from regional center to regional center. The analysis
only includes explicit population measurements from Castile. The kernel density map
(figure 5.7) visualizes the only place the routes significantly clustered: through the
Valladolid-Madrid-Toledo cluster of population; and in Andalucía, around SevillaMálaga-Córdoba-Granada cluster of population. All of the other routes ran long
distances to connect these population complexes. Other clusters of intermediate
significance existed in the Burgos, Cuenca, Valencia, Murcia, Barcelona-Zaragoza,
Pamplona, Vittoria-Gasteiz, Palencia, Lisbon, Estremoz, and Castelo Branco placename regions.
However, there existed an elaborate but sparse network of routes traversing the
space between central and eastern clusters of places and populations. Upon quick
examination of the map, one can see 13 routes between these regions of Iberia. In
contrast, five routes connected central Iberia to the western regions of Portugal and
Galicia. Most of the routes going towards the west ventured no farther west than the
meridian running near Sevilla, Cáceres, and Plasencia, at about -6 degrees longitude.
96
Martín Benito provided a description of the difficulties of travel through the
Zamora region which included precarious wooden bridges over rivers (one of the most
dangerous obstacles in a traveler’s path through Spain), dangerous mountain passes
that, due to their high peaks and dense trees, harbored bands of robbers; travelers also
found them treacherous in the winter blizzards.119 Physical violence, explained Téofilo
Ruiz, offered a manifestation of the upheaval of day-to-day existence in the sixteenth
century.120 Legal violence presented a sign of many governments’ difficulty in
controlling their populations in an organized fashion. Banditry caused systemic trouble
in southern Spain and in Catalonia during the sixteenth century.121 In 1608 a group of
Moriscos (Moors nominally converted to Christianity) were accused of armed banditry
in the Valencian region, and eleven of them were condemned to die in absentia.122
Contrastingly, in 1852, English visitor Richard Ford described the Spanish travel
routes thus:
Of the many misrepresentations regarding Spain, few are more inveterate than
those which refer to the dangers and difficulties that are there supposed to beset
the traveller. This, the most romantic, racy [sic], and peculiar country of Europe,
may in reality be visited by sea and land, and throughout its length and breadth,
with ease and safety, . . . the steamers are regular, the mails and diligences
excellent, the roads decent, and the mules sure-footed; nay, latterly, the posadas,
119
José Ignacio Martín Benito, Cronistas y viajeros por el norte de Zamora. (Benavente, Spain: Centro
de Estudios Benaventanos <<LEDO DEL POZO>> (CECEL – CSIC), 2004).
120
Ruiz, Spanish Society, 1400-1600, 167.
121
Ibid, 166.
122
Ibid, 182.
97
or inns, have been so increased, and the robbers so decreased, that some
ingenuity must be evinced in getting either starved or robbed. 123
He supported this argument by discussing the development of the guardias civiles, a
mounted police which patrolled the road frequently, after their founding in 1833, by
Martinez de la Rosa, whom bandits had also attacked, as he made his way to Madrid
and his government office as head of the political party in power at the time: “the
guard [of Señor de la Rosa’s train], at the first notice, [threw] himself on his belly,
with his face in the mud, in imitation of the postilions [government postal runners],
who pay great respect to the gentlemen of the road.”124
Ford, in 1852, also described the Spanish ventas, or small roadside inns, as
places of roguery and misery of accommodations (leaving behind his earlier
optimism). He explained that the term was a Latin word intended to imply the selling
of goods, but the ventas never actually sold any.125
The great bulk of the Peninsular family, not being overburdened with cash or
fastidiousness, have long been and are inured to infinite inconveniences and
negations; they live at home in an abundance of privations, and expect when
abroad to be worse off; and they well know that comfort never lodges at a
Spanish inn.126
Ford continued his description of the venta, describing it as a place to tie one’s animals
for the night; beyond a bed of pest-infested hay in a loft above the animals’ stalls, and
123
Richard Ford, Spaniards and Their Country. (New York: George P. Putnam, 1852), 40.
124
Ibid, 189.
125
Ibid, 177.
126
Ibid, 189.
98
meager offerings from the venta’s cook, it offered little else; at the venta, “hay de
todo”: you could have the world, as long as you brought it with you. 127 Ford’s
description of the venta, inasmuch as he described it accurately, must have only
applied more completely in the sixteenth century, before the advent of any kind of
railway system, when the considerably smaller Spanish population traveled less
frequently. The Spanish travel route did not provide tourist comfort because travelers
mainly used it for short- to moderate-distance travel. The majority of the unknown
place-names in this study consist of ventas located in areas with very low densities of
other towns or villages. They likely filled the role Christaller pointed out for smaller
central places; they were filling a need in Spain’s network of places, in geographic
locations that fell outside of the maximum distance for the central goods of the nearest
central places. This is especially true for ventas located in Andalucía north of Sevilla,
where population and settled places hardly existed. Reconstructing Spain’s sixteenthcentury commercial and pilgrimage routes provides an essential context for analysis of
the economic, political, and religious makeup of Spain in this period.
127
Ibid, 189.
99
Figure 5.6. Castilian Populatio Ce te s i
, alo g Villuga’s Routes.
100
Figu e . . Ke el De sit of Castilia Populatio Ce te s of
101
, alo g Villuga’s Routes.
Polycentric Life in Sixteenth-Century Iberia
Among the multiple causes for the decline of the Castilian Crown in the
seventeenth century, transportation network constraints had a significant impact. As
Ringrose pointed out, after 1561 Madrid’s dominance of the economy gradually
reached such high levels at times, that it taxed the limits of Spain’s already stretched
transportation system. Spain lacked the development of its transportation system
enjoyed concurrently by France, which had an extensive canal system, and England,
which had far-reaching navigable rivers which facilitated a much stronger nationallevel transportation infrastructure.128 Ringrose argued that Spain’s lack of a sufficient
transportation system perpetuated a mainly local economy that to some degree had
reached into the market economy. 129 In addition to the large amounts of goods
gravitating along the routes toward Madrid, moderate amounts of goods and animals
flowed along the long-distance routes, mostly consisting of luxury items and the
Mesta’s vast flocks of sheep. These flocks provided the backbone of a considerable
market in renowned Spanish wool. But for the most part, goods flowed on regional
and local bases, making markets highly regionalized, due to topography and weather.
The coastal provinces relied on seaside connections to emergency food sources during
famine, and could thus specialize in export crops, but remote parts of the interior had
128
Ringrose, Transportation and Stagnation, xxii.
129
Ringrose, Transportation and Stagnation, 18-24.
102
difficulty looking far outside themselves during famine.130 Most goods from the
interior could not bear the cost of transport to the coast, for international trade; wool
provided one exception. The Europe-wide economic transformations did not reach as
far into the interior of the Castilian Crown as they did along its coast.131
The peninsula is best seen as a large mosaic of self-sufficient local economies
buttressed by short-range exchanges of basic commodities. This mosaic was
criss-crossed by a web of economic connections. Its strands represented the
meager inflow of manufactures and amenities, the export of the few goods that
could bear the cost of transport, and the movement of commodities controlled by
the landed elites. Occasionally a few strands joined where a market, bishop,
governor, or court dispensed services. Near the coasts, the web was denser as
port towns provided the mercantile functions made possible by the sea. 132
Here Ringrose likely spoke of a Spain with Madrid as the royal Household and Court, but
these patterns existed prior to 1561, as well. The population relied heavily upon its farms,
and few could afford to venture into transport as a full-time business. Spain’s inclimate
weather and lack of sufficient pasturage lands for the pack animals (also insufficient in
number) that supplied the mainstay of the system, restricted working in the transportation
system to seasonal work that did not pay high enough returns to allow most farmers to
safely leave the agricultural sector to expand the transportation sector. The majority of
the population involved in transport of goods operated seasonally within a locally- or
regionally-focused economy, mostly sustained by subsistence with some local and
regional market activity. The drain on the transportation system by Madrid, which
130
David Ringrose, Madrid and the Spanish Economy, 1560-1850. (Los Angeles, CA: University of
California Press, 1983), 5.
131
Ibid, 6.
132
Ringrose, Transportation and Stagnation, 6-7.
103
dominated the use of the infrastructure, eventually by about 70 percent of its capacity (by
the 1750s; it would have been less pronounced in the late sixteenth century), hampered
further expansion of the transportation system to assist the growth of industry. 133 The
transportation system’s limitations aided in Spain’s polycentric urban structure. Into the
seventeenth century when Madrid was building royal roads to service the whole of the
peninsular kingdom, the local populations still maintained most of the roads, and
spontaneous trails connected local towns and villages.134 Before 1561 the limited
transportation structure aided the development of a polycentric system of population and
economy. After 1561, Madrid gradually pulled much of the limited transportation system
into its own use, which focused on bureaucratic consumption of incoming goods, rather
than on a trade relationship between Madrid and its hinterland. Therefore, the limited
market economic activity focused mostly on Madrid, which did not offset its
consumption with production. The remaining economic activity in the Castilian Crown’s
peninsular lands operated within a constrained system focused on a combination of barter
and market activity on the local and regional scales. Essentially, Madrid as the location
of the royal court gradually converted the Spanish economy from polycentric to a
monocentric focus on supplying Madrid via ever-increasing percentages of the available
transportation capacity. The limited remaining capacity of the transportation network
served the majority of the geographic extent of the Peninsula. This resulted in
perpetuating and even extending the focus on economies of relatively small scope, as
133
Ibid, 90.
134
Ibid, 17.
104
most of the transporters consisted of farmers and an ever-smaller number of pack animals
who could only take brief excursions away from their fields.135
Vassberg argued that historians have underestimated the market system of
sixteenth-century Castile. While geography hampered transportation systems,
extensive regional markets did exist, as well as connections across the Iberian
Peninsula, and with markets in Italy, Flanders, and other foreign lands.136 Villagers in
sixteenth-century Spain focused mainly on an autarchic, subsistence-level focus for
meeting their needs, save for the occasional connections they made to the marketplace:
first, on a barter basis with other villagers, and then on a regional level, sending wood,
charcoal, or other locally available commodities into other regions in a market-based
exchange for that region’s particular commodities which the villagers could not find
elsewhere.137 Specialization of production required a strong transportation network.138
Without this essential, Spain’s market network remained largely an augmentation of
and support for the local and regional focus of the economy. Villagers bought iron and
salt outside the village by selling their extra grain or animals, but on a small scale; “the
early modern Spanish agriculturalist was sufficiently involved in the market economy
to justify calling him a ‘farmer,’ rather than a ‘peasant.’”139 While some producers
135
Ibid, 57-62.
136
David E. Vassberg, . The Village and the Outside World in Golden Age Castile: Mobility and
Migration in Everyday Rural Life. (New York: Cambridge University Press, 1996), 25.
137
Ibid, 25.
138
Ringrose, Transportation and Stagnation, viii.
139
Vassberg, The Village and the Outside World, 25.
105
focused more on the autarchic level, many areas of Spain did have some degree of
specialization of crops for a market-based economy, particularly in Castilla y León.140
Many markets and fairs operated in Spain, such as the fair in Medina del Campo, and
several others; the villagers of the market’s region produced most of the goods one
might find there.141 To emphasize the importance of the regional fairs, Vassberg
pointed out that while fairs in major market towns like Medina del Campo pulled its
goods from a large and far-flung network of places, regional fairs continued to operate
alongside the larger fairs. However, Palomas, near Badajoz, gives us further evidence
of a far-extending market. This village with only 200 households in 1575 sent goods to
fairs at a distance between 5 and 16 leagues, or from about 23 to 90 kilometers away.
In 1529 Salamanca secured rights to hold a tax-free market every Thursday which
allowed all villagers within a 12-league radius of the market to enjoy its tax-free
status.142 The villagers of Palomas and the villages around Salamanca might have
secured the services of muleteers or carters who worked as full-time transporters, but
they more likely would have designated a team of muleteers from their own village to
take a brief respite from their agricultural work to travel with animals and vehicles
from their own village. Arrieros ordinaries (specialized muleteers) traveled as
140
Ibid, 26.
141
Ibid, 27.
142
Ibid, 27-28.
106
designees from their own villages to regional markets with mixed cargoes of
expensive goods and basic commodities.143
Pastoralism provided a key industry in the Iberian economy. Phillips and
Phillips argued that large-scale sheep herding in Spain complemented agriculture,
providing a more efficient use of the land in the face of environmental constraints, also
restraining the growth of arable during times of rapid population growth. They argued
that pastoral uses of the land claimed an unusual importance even in comparison with
other Mediterranean regions. The king established the Mesta, a royal stockraisers’
association, and granted it unrestricted passage through lands in 1273, which the
Crown renewed in 1414. The king granted rights to the Mesta to alleviate disputes
over lands and animals among herders, and also land disputes between herders and
local landowners. 144 Ringrose argued that wool provided “the only product that could
bear the cost of transport to the sea and still be profitable.”145
Ferdinand and Isabella made some effort to aid Castilian industry. However,
the large amount of edicts they issued that focused on the economy contained
insufficient levels of coherence, and they had already committed themselves to
supporting the stockraising industry. They did this for at least one reason: tax revenue
143
Ringrose, Transportation and Stagnation, 27-8.
Carla Rahn Phillips and William D. Phillips, Jr., Spai ’s Golde Fleece: Wool Productio a d the Wool
Trade from the Middle Ages to the Nineteenth Century (Baltimore The Johns Hopkins University Press,
1997), 8-9, 36-37.
144
145
Ringrose, Madrid and the Spanish Economy, 5.
107
grew considerably in connection with the rise in strength of the pastoral economy.146
Industries such as metallurgy, paper production, soapmaking, and fish salting all had
their place in the growing sixteenth-century Iberian economy.147 The textile industry
had its moments of success. Granada, Cuenca, Segovia, and Toledo all produced silk
or woolens. Public authorities refused the exportation of approximately one-third of
raw materials. The 1580s marked one highpoint for textiles, but soon the international
market favored other producer regions outside of Castile, which then turned primarily
to exporting raw materials.148 Geographical variations in Iberia also affected the
Catholic Kings’ efforts to stimulate the increase of grain production. While the meseta
region in southeastern Castile and parts of Andalucia supplied a surplus of grain, the
northern Iberian regions imported it to fill their food needs, and grain from the meseta
faced multiple transportation challenges, such as customs barriers and transportation
deficiencies.149 Spain and its industries remained mostly focused on local and regional
spheres during the sixteenth century, constrained, among other causes, by a small,
dispersed population and an insufficient transportation system.
Braudel compared Iberia to an island, with many smaller islands within itself, a
place of inaccessibility and unique personality. “It would hardly be an exaggeration to
say that Portugal, Andalusia, Valencia, and Catalonia were a series of peripheral
146
John Lynch, Spain 1516-1598: From Nation State to World Empire (Cambridge: Blackwell, 1994), 23.
147
Phillips, Ti e a d Du atio ,
148
Henry Kamen, Golden Age Spain, 30-31.
149
.
Henry Kamen, Spain 1469-1714: A Society of Conflict. (New York: Longman, Inc., 1983), 49-51.
108
islands attached to the Iberian mass through Castile.”150 The Spanish interior’s isolated
position from the sea and the rest of the European commercial system forced it to work
within the context of “limited resources, poor transport, and unreliable harvests. Such
circumstances restricted market-oriented agriculture and encouraged regional selfsufficiency”.151 Ringrose explained that “self-sufficiency was based on overlapping
regional commodity exchanges that were sometimes not far from barter . . . exchanges
rarely transcended a 50 to 75-mile radius created by terrain and transport that could
double the price of wheat within 100 miles.” 152 “For the most part,” commented
Philips, “the population centers of La Mancha were far apart and nucleated near the
sources of fresh water.”153
Camacho Cabello recorded that among the most populous areas of the modern
Comunidad de Madrid and Castilla-La Mancha, according to the 1591 Censo, were:
Alcalá de Henares (15,553), Almagro (21,991), Guadalajara (23,897), Madrid
(28,201), Ocaña (21,251), Talavera de la Reina (16,223), and Toledo (42,857).154 In
addition, many other localities throughout Castile had significant populations, as seen
in the maps in figures 5.6 and 5.7 (below) that in 1591, thirty years after Madrid
150
Fernand Braudel, The Mediterranean and the Mediterranean World in the Age of Philip II. Vol. 1.
(New York: Harper & Row, Publishers, 1966), 161.
151
Ringrose, Madrid and the Spanish Economy, 164.
152
Ibid, 165.
153
Carla Rahn Philips, Ciudad Real, 1500-1750: Growth, Crisis, and Readjustment in the Spanish
Economy. (Cambridge, Massachussetts: Harvard University Press, 1979), 6.
154
José Camacho Cabello, La población de Castilla-La Mancha (Siglos XVI,XVII, XVIII): Crisis y
renovación. (Toledo: Junta de Comunidades de Castilla-La Mancha, 1997), 102-117.
109
became the royal court, clustered together according to Dobbs’s definition of
polycentric urban regions, in formations of interdependent urban clusters. Examples of
polycentric urban clusters appear clearly on the maps in figures 5.6 and 5.7: Ciudad
Real, Daimiel, and Almagro; Salamanca, Zamora, and Valladolid; Badajoz and
Alburquerque; Córdoba and Ecija; Lorca and Murcia; Burgos and Villadiego; Segovia,
Madrid and Toledo; from Cáceres to Plasencia; from Sevilla to Antequera, from
Cuenca to Villarrobledo; and others.
Camacho Cabello provided population data from earlier in the sixteenth
century that further strengthens the polycentric patterns evident in figures 5.6 and 5.7.
He stated that the Castilian population grew by 15.8 percent between 1575 and 1591;
but while Madrid province grew by 12 percent, Toledo province grew by only 7.1
percent; all of Castile’s other provinces also grew at a slower rate than that of Madrid
during this time.155 These figures help to correct the maps in figures 5.6 and 5.7, more
accurately revealing the population levels at the time of Villuga’s record. They also
support the notion of the growth of Madrid as the center place in a gradually more
mono-centric Castile after 1561. Madrid’s dominance grew only gradually over an
extended period of time.
Some examples from Spanish history demonstrate its economic, political, and
religious polycentrism. In chartering Ciudad Real, first known at its thirteenth century
founding as Villa Real, Spanish King Alfonso X of Castile sought a counterweight to
155
Ibid, 109-115.
110
the Military Order of Calatrava, a crusading order which held considerable power in
the region around Almagro. They had gained the land as royal grants, in connection
with their capture of it during the Reconquest, but the king disliked their
counterbalancing effect on his power. While Alfonso X allowed the villagers to utilize
wood, and other resources on the lands of the Order of Calatrava, when King Alfonso
XI proved too weak to control them, the knights of Calatrava assumed authority over
the place, refusing Villa Real the use of wood in the surrounding hills, which the Order
owned.156 The brief arrival of the Chancillería (high judicial court) in Ciudad Real
from 1494 to 1505 in Ciudad Real aided its ascendancy as a village, but before and
soon after, when the Chancillería moved to a new location, Toledo province held
jurisdiction over it, and represented it in the Cortes.157
Ciudad Real’s intermittent states of productivity and drought made its cash
crops, wine and oil, tenuous resources and separated it from the larger market
economy, since Ciudad Real had to get its grain supply either from an outside source,
or from its own struggling soil.158 Its industry, commerce, and trade all augmented
Ciudad Real’s subsistence economy, but wasn’t strong enough to help it completely
break out of an emphasis on subsistence and more fully join a market economy. 159
156
Philips, Ciudad Real, 1500-1750, 11.
157
Ibid, 15.
158
Ibid, 44-45.
159
Ibid, 49.
111
In the mid-fifteenth century, John II of Castile spoke of his absolute royal
power, which he used to write laws as valid as those coming from the Cortes.160 But in
the reign of the son of John II, Henry IV of Castile (1454-74), he being the next king,
“anarchy, fomented by unscrupulous nobles, dominated much of the reign, and in an
infamous tableau known as ‘the Farce of Avila’ an effigy or statue of the king was
deposed,” followed by the real king himself.161 This revealed that the Castilian kings’
absolute power still eluded them, diminished by the powers of the regional nobility.
Ferdinand and Isabella are generally known as the unifiers of Spain, and yet,
into Richard Ford’s day in the nineteenth century, the several kingdoms of Spain
appeared, as he described them, “an ‘unamalgamated bundle’” with regional and local
laws holding the true authority; the king’s decrees would arrive in a far-flung area of
the peninsula, ritually obeyed and “laid aside for ‘obedience without compliance.’”
While Ford may have exaggerated the limitations of the royal authority, FernándezArmesto argued that the Spanish king’s claim to absolute power referred to the right to
dispense justice, rather than to legislate, and that this absolute power really referred to
the king’s standing as answering to God only.162
160
Fernández-A esto, Felipe, The I p o a le E pi e, I Raymond Carr, ed. Spain: A History. (New
York: Oxford University Press, 2000), 111.
161
Ibid, 113.
162
Ibid, 121.
112
“Spanish rulers,” argued Fernández-Armesto, “could not have taken a uniform
approach to their realms, even had they wished to do so.”163 Essentially, the Spanish
rulers, like the contemporary French monarchs, did not even desire to rule all parts of
their kingdom in uniformity, but rather according to regional and local customs. The
kings relied upon intermarriage with noble families and patronage of pliant nobles,
without whom the king’s power on regional and local scales would have had a lesser
sway.164 “The real power struggle was not between Crown and nobles but between
nobles and towns.”165 The Castilian Crown and the nobility supported each other; the
latter propped up the former and by so doing kept regional or local power according to
long-standing polycentric traditions.
Even the common people of the Crown claimed some authority. Bartolomé de
las Casas argued that “citizens are not under the power of the sovereign, they are under
the power of the law,” and Philip II’s Nueva Recopilación, a new reiteration of
Castilian law, determined that “royal commands contrary to divine and natural law, or
against conscience, the church, or the faith; or uttered in anger, were of no force.”166
One may ask who was going to enforce this statement, other than local nobility or the
common man himself. Mackay argued that beyond the king’s requirement to answer to
the nobility and other elites, through patronage and clientage, the king had to answer
163
Ibid, 122.
164
Ibid, 122-126.
165
Ibid, 127-128.
166
Ruth Mackay, The Limits of Royal Authority: Resistance and Obedience in Seventeenth-Century
Castile. (New York: Cambridge University Press, 1999), 1.
113
to the common man. “Kings did not have to contend just with estates and ambitious
nobles; they also had to answer to men and women who did not equate their own
humble status with powerlessness.”167
Alonso García explained that before Madrid became the royal court, it
interfaced with a complex “urban and commercial web.”168 Important bankers and
merchants operated within the city before 1561. The Castilian Crown had earlier made
one of its primary revenue sources the alcabala, or general sales tax. Queen Isabella
gave the towns the right to gather the alcabala as a locally-assessed lump sum, or
encabezamiento, giving local elites power over the collection of a considerable
percentage of royal revenue.169 A general encabezamiento took shape in 1536 in
Madrid and morphed into a more inclusive assessment, including certain rents on lands
and other taxes, in 1547. The king had two bodies of government which governed the
encabezamiento, one for making decisions and the other for practical executions of the
royal decrees concerning the encabezamiento; in other words, the king did not have
sufficient power, at that time, to make his pronouncements absolute; the prominentlyheld ideal of an absolute monarchy did not fit completely with the polycentric reality
in sixteenth-century Castile.170 The king left much of the management of tax collection
to the city and territories themselves, in the form of a receptor (territorial treasurer), a
167
Ibid, 4.
168
David Alonso García, Una Corte en Construcción: Madrid en la Hacienda Real de Castilla (15171556). (Madrid, Spain: Miño y Dávila, 2005), 41-44.
169
Kamen, Spain 1469-1714, 48, 87.
170
Alonso García, Una Corte en Construcción, 45.
114
royal officer.171 The king’s social network contextualized and partly defined his
authority. Local collection of the encabezamiento often took the form of auctions,
which diminished the amount actually collected, but the receptor had to work within
the local system, alongside the local tenants. The king’s trusted agent thus had to yield
to local, rather than royal, authority. In Madrid, two traits made a man eligible for this
office: his membership in the local oligarchy, and his links to those groups that could
offer the king liquidity. Alonso García argued that one must see the receptor del rey as
integrated into a complex social network with ties to many scopes of society.172 This
social network formed the backbone of the king’s royal authority, which authority was
limited by his need to collaborate with and concede governing authority to nobles and
towns.173
Negotiations for commercial or political rights depended on consent (or at least
forced agreement) among the parties to the contract, within the context of feudal
fidelity and obligations between vassal and lord, in unequal relationships. Debt and the
search for credit and liquidity formed inherent parts of the economic system.174
Therefore, when the regidores (magistrates) of Madrid needed favors from the king,
171
Ibid, 46.
172
Ibid, 51.
173
For more information about this collaborative relationship, see J.B. Owens, By My Absolute Royal
Authority: Justice and the Castilian Commonwealth in the First Global Age. (Rochester, NY: University
of Rochester Press, 2005).
174
Alonso García, Una Corte en Construcción, 59.
115
they offered him large sums of money, and their fealty.175 The power of a city to get
representation before the Crown could enlarge its status; they often sought this
objective by offering the king services. To negotiate the encabezamiento required a
man of high quality. In the first quarter of the sixteenth century, Bernardino de
Mendoza became the receptor of Toledo when the king appointed his father-in-law as
receptor general del reino (general treasurer of the kingdom) in Madrid. His positions
in the all-important social network gave him the power needed at this time to gain the
money and credit essential to a high status. Francisco de Vargas had many links, inside
and outside of Madrid. When he gave the king money in return for royal favors, the
king in return acceded to some decisions made from below. Francisco de Vargas
operated a position of influence with the king due to his links to Madrid, and with
Madrid due to his links to the king.176
In 1517-18, when representatives of Madrid’s government sought the king’s
grant of a reorganized encabezamiento, they sought it out of the context of their just
dues, while the king sought to make it a sign of his grace towards his kingdom,
granted in that context and not in the one in which Madrid sought it. Some of the
other cities in the kingdom sought common cause with Madrid before the king, by
sending Madrid notice that they would seek the king’s aid with updating their
encabezamiento.177 The cities themselves enjoyed autonomy in some ways, such as in
175
Ibid, 51.
176
Ibid, 63-67.
177
Ibid, 80-81.
116
their management of funds, as far as they could conflate their powers with those of the
Crown, and they had some fiscal power over the collection and use of rents from their
localities. The Castilian Crown was obliged to work within a system of a
conglomerated and pluralistic political system, with many types and scales of
treasuries, such as the royal, the ecclesiastic, and the municipal, all with separate and
disputed jurisdictions.178 It is true that the king required his subjects to pay taxes, but
he gave them with the understanding that he granted other rights in return, such as
local control over rents, which provided tools in economic exchange, social control,
and political ascension.179
Spain’s polycentric nature also had some impact on religious localism. This
localism did not necessarily change over time as a result of Madrid becoming
gradually more dominant. This aspect of Spain’s polycentrism likely resulted more
from the reach and particular administrative institutions of the Catholic Church in the
sixteenth century, and their degree of strength in the Iberian Peninsula at that time,
rather than stemming from the political reach of Madrid or the Royal Court, although
the Court did reach into church affairs to important degrees. Iberia’s particular
polycentric structure did have an impact on the strength of the Church’s authority in
Iberia. Isolation and difficulty of movement had its impact on the religious lives of
Iberia’s dispersed localities.
178
Ibid, 120.
179
Ibid, 187.
117
The Catholic Church sought to unite the whole church in one form of worship
through “the universal cults of the Virgin Mary and the suffering Christ,” but failed to
completely make the change. The people of the various Spanish localities, for
example, adapted the universal cults to their own local traditions, in addition to tying
into the universal cults.180 In Toledo, the religious brotherhoods took as their basis the
given occupations of their members, such as blacksmiths, day laborers, silk spinners,
and so on. The brotherhoods despised each other’s saints, and one saint might have
more than one helping function depending on which brotherhood sought their services,
all within the various parishes just within the jurisdiction of Toledo.181
A common notion stated that God’s grace was more powerful in some places
than in others. This idea supported the development of shrines and pilgrimages;
moreover, it supported the localization of religious practice.182 At times a local shrine
would become important to the surrounding region, but most shrines and sacred
monuments enjoyed importance only on a local scale, and each shrine required
continual support from the local populations, if the people did not wish for unpleasant
consequences from their patron saints.183 Diocesan legislation, such as that which
supported the reforms of the Council of Trent and the work of the Inquisition, sought
to support the universal form of Catholicism on the local scale, to the detriment of
180
Mackay, The Limits of Royal Authority, 133.
181
Ibid, 150.
182
Ibid, 150.
183
Ibid, 177.
118
local customs, but with mixed success.184 Religious vigils in churches and shrines,
held overnight (which tended to the pilgrims’ riotous behavior); charity feasts on fast
days; reluctance to go to confession; occasional lay laxity concerning the church’s
standards on fornication; and combativeness against one’s priest, all discouraged or
forbidden by the church and its diocesan visitors to the far-flung localities of Spain,
revealed that local religious customs held their influence in the face of the church’s
centralizing efforts.185 Local members of the church sometimes took up the habits of
brotherhoods to which they did not belong, and claimed to be gathering alms on the
behalf of the brotherhood. Therefore, the church issued a restriction on wearing the
hermit’s habit without the approval of the diocese. The Council of Trent even
reinforced the powers of religious localism by affirming the use of local sacred places,
locally renowned relics, and special patronage of a particular village from a given
saint.186
Women often formed beatrías, or unrecognized and unofficial religious orders,
the adherents claiming a direct connection with the divine, without the need for the
strictness involved in the monastic orders. The church considered these beatas
threatening and they required that all beatrías be converted into official convents or
members of third orders. The spreading of stories of “false miracles” and selling of
unverified relics caused the church to demand notarization of testimonies and
184
Ibid, 162.
185
Ibid, 166-167.
186
Ibid, 169, 179.
119
certification of relics. These certificates often relied on vague evidence. Locals in
Toledo extended feast days to 13 more days than Rome had dictated, and out of the
necessity of the poor, and for need to bring in the harvest, many of these days the
people did not observe; in some places in Cuenca the bishops had refused to grant
more vowed days; the people observed them anyway, preferring their own feast days
to those ordered by the church.187
Villuga’s record, combined with other historical data, revealed valuable
information about the urban, transportation, and population structures of sixteenthcentury Castile. The revealed polycentric urban structure manifested itself in the
economic, political, and religious life of the time period. The structure affected the
reach of central powers into the finer scales of Iberian life, on the village or regional
levels. Although some significant economic activity did exist on coarser scales, such
as the peninsular or international level, the polycentric structure of Iberian urban
regions had a constraining influence on the integration of the local economy into the
larger structures.
187
Ibid, 170-172.
120
CHAPTER 6: CONCLUSION
The use of digital technologies such as GIS, and the use of cartographic
analysis methods and spatial autocorrelation techniques have the potential to make
important contributions to the development of historical studies, by aiding in the
integration of space, place, and time, and by providing tools to manage large amounts
of data. The potential resulting growth of historical understanding through the practice
of geographically-integrated history will prove worthy of the effort. Maps have
provided essential aids for visualization and analysis (e.g., the Mercator map’s use in
navigation). Statistical summaries, such as those of mass immigration, have in the past
provided useful tools in historical analysis, but historians’ utilization of databases,
spatial statistics, and GIS technology has a short history. The new digital technologies
mentioned in this study have their difficulties and challenges when applied to
historical research. Also, GIS research has only produced a rudimentary
implementation of fuzzy analysis of vague data.188 Major barriers to integrating GIS,
information technologies, and statistics into historical studies include the learning
curve involved, the considerable time required for input of historical data into
databases and GIS formats, finding funding to provide the needed time, and the
analysis of vague and uncertain data. However, the increased understanding of the
historical narrative is worth the effort, and many scholars and organizations have in
recent years been advocating a global analysis of the historical processes of human
188
Michael F. Goodchild, . “tatisti al Pe spe ti es o Geog aphi I fo
Analysis 40 (2008): 323.
121
atio “ ie e. Geographical
activity. As stated, GIS provides one of the best integrative tools for a pursuit of this
global analysis.
ESRI’s ArcGIS 10 provides some tools for the use of fuzzy data, and work
moves forward in GIS research to provide for the inherent uncertainty or
incompleteness in all types and sources of GIS data. 189 Database design continues to
evolve towards platforms more useful to the historian’s complex data.190
Geospatial approaches to, and GIS tools used for, organizing place-name and
population data proved a useful augmentation of traditional database practices. For
example, the Near analysis of two sources of gazetteer data facilitated a quick
integration of the two sources of data. The efforts of many organizations to standardize
and improve gazetteer design practices are aiding in the development of the
geosciences. GIS tools and spatial organization of this study’s data demonstrated that
the forces of linguistic, cultural and political regionalization in Spain thrived at the
height of the Castilian Crown’s global power.
Geostatistical methods, or
specifically the spatial autocorrelation tools used, evaluated Villuga’s league data and
demonstrated that the kilometers-per-league distance ranges produced were highly
189
Ga L. Rai es, et al., I o po ati g E pe t K o ledge: Ne Fuzz Logi Tools i A GI“
ArcUser, Spring 2010, 8-13.
190
,
For information on a new database, Intentionally-Linked Entities, see Vitit Kantabutra, A new type of
database: Intentionally-Linked Entities—a direct implementation of databases from the
Entity/relationship model. Unpublished PowerPoint presentation. Accessed 10 December 2009:
http://progeny.isu.edu/~vkantabu/ILE/. See also Kantabutra Patent application of Vitit Kantabutra for
Title: Intentionally-Linked Entities: A general-purpose database system, 2009[b]. Accessed 10
December 2009: http://progeny.isu.edu/~vkantabu/ILE/.
122
positively autocorrelated, validating the hypothesis that the distance ranges clustered
into regional patterns of use. This analysis applied a method of data generalization to
the GIS analysis of Villuga’s league data. The analysis of population data and
historical sources provided a cartographic visualization of Castile’s polycentric urban
system prior to the rise of Madrid as the Castilian Royal Household and Court.
6.1 Gazetteer Creation
Gazetteers provide a useful tool in connecting informal data about geographic places,
such as text-based place-names, to formal mathematical descriptions of places. The
best practices documented by gazetteer authorities provide a foundation for designing
an efficient gazetteer and avoiding common problems that researchers have already
solved. The gazetteer of Villuga’s guide, combined with other historic data sources of
place-names, provided a greater density of place-names to compare in their temporal
changes across several centuries. This density of Iberian place-names across time
provides a ground-breaking study of the impact of Spanish naming policies in the
nineteenth and twentieth century. It organizes place-names along the routes of the
transportation infrastructure recorded in the sixteenth century, and tracks their changes
over time, facilitating a connection between sources of historical data across the time
spanning from the sixteenth century to the present day. For example, trade records
from different time periods may identify a place by more than one place-name. The
Villuga Gazetteer connects these varying names by their spatial location,
contextualizing them for greater clarity. The greater density of place-names aids the
123
historian in identifying place-name matches across historical sources. The search for
Villuga’s place-names revealed many that no longer appear on modern maps.The 1834
reorganization of Spain’s administrative institutions, followed by the 1978
restructuring, had a significant impact on Spanish place-names. Use of GIS in
organizing place-name variants across several different time periods provided a greater
density of place-names for the historian to compare in an analysis of place-name
change over time. The greater density of place-names aids the historian in identifying
place-name matches across historical sources. The search for Villuga’s place-names
revealed many that no longer appear on modern maps. In possible future studies,
adding further place-names from other regions of the Crown’s territories will further
expand the gazetteer’s usefulness.
6.2 League Analysis
The league had a general and variable standard of measurement in the sixteenth
century. Historians analyzing the Iberian historical record may find league data that
provides valuable information about transport and its impact on economies, but they
face the challenge of the league’s vagueness. The map of Villuga’s league distances
indicates spatial distributions of general ranges of kilometer measurements for the
league in sixteenth century Spain. A spatial analysis verified regional patterns of
dominance for each range of league measurements, except the last range. By finding
the general ranges of league distances across the Iberian Peninsula, the autocorrelation
analysis determined the most likely distances for places to be apart from each other.
124
The outlined methodology provides a method of applying vague and coarsely-defined
data (Villuga’s league distances) to a GIS analysis, relying upon generalization of the
data through the use of a range of data values, rather than a more crisply defined
definition of the league. These results provide an important tool in locating the missing
place-names in Villuga’s historic record, and can provide guidance to historians who
seek missing places from other historical sources. Historians can transform ill-defined
spatial data from their sources into ranges for use in GIS. This approach can aid in
overcoming a major obstacle in using computers to analyze imprecise data produced
by humans. Fuzzy rule-based modeling takes this approach further; this study applies
it to historical data and GIS analysis tools.
6.3 Traditional GIS and Fuzzy Rules
The integration of fuzzy logic principles augmented traditional GIS methods.
Using this study’s method, historians can assign a greater role to their knowledge of a
historical setting when employing GIS conditional statements. They can take into
account their knowledge of certain aspects of a time or place, such as a traveler’s need
for a stopping place to have a sufficient amount of pasture for his animals. A
combination of existing GIS tools and fuzzy if-then statements facilitate the utilization
of a historian’s knowledge about a topic under analysis.
This method expanded upon existing GIS statistical methods by generalizing
them for compatibility with Villuga’s vague league distances.191 This study introduced
191
“ ithso , Fuzz “et I lusio ,
.
125
a methodology which integrated fuzzy logic principles with traditional GIS
interpolation principles. The condition statements produced by this technique take into
account the vague and variable nature of historic data, and provide the historian with
the ability to produce parameters of membership for geographic data in fuzzy if-then
text-based statements. These GIS conditional statements, augmented by fuzzy rulebased modeling, will provide the most likely coordinates for the unknown place-names
of Villuga’s record, thus providing the reader with a good example of the usefulness of
soft computing in GIS. Fuzzy rule-based modeling provides greater inclusion of the
domain expert into the GIS interpolation methodology. This method applies to other
ill-defined historical data, and its analysis in GIS.
6.4 Population and Urban Patterns along Transportation Routes
This study has provided greater understanding of the economic, urban, and
social history of Spain from the reconstruction of its major sixteenth-century
roadways. By visualizing the human population at a critical transition point in Spain’s
past the student of history and economics may gain greater comprehension of this
aspect of Spanish history. This visualization of Castile’s past routes and population
centers has provided further empirical examples of the theory of polycentric urban
regions, and a fuller picture of Iberian social interactions and interconnections.
This study uses a gazetteer database to demonstrate Spain’s urban pattern in
the sixteenth century. This result would not materialize from a reader analyzing the
data sources which provided the basis of our GIS analysis. Only a database form
126
would suffice for the organization of this large data set of place-names, their distances
from each other, and their populations. Moreover, only a GIS map format would
sufficiently organize the data across space and reduce the historian’s cognitive burden
to a manageable level.
The GIS analysis of Castile’s population along its main routes provided a
picture of what the economic life of Castile looked like in the late sixteenth century,
and further population data from secondary texts demonstrated that several polycentric
urban regions functioned in Villuga’s time, providing a localized context for
economic, religious, and political power. Combining gazetteers through spatial
analysis provided a more automated way to utilize many sources of digital geospatial
data. Historical narratives illustrated the polycentric nature of Castile’s urban
structure, showing its locally-focused and regionally-interdependent economic context,
while providing evidence of some trans-peninsular connections for the transporter of
such things as the rare luxury good, the shepherd, or the export. The outcome of the
population analysis will aid historians in the study of Madrid’s impact on the Spanish
population distribution and economic structures over time. The uncertainty and
incompleteness of the data will have some impact on the outcome, but the available
data show sufficiently the anticipated polycentric urban pattern on the landscape of
Iberia’s routes. The major historiographic question of the cause of the decline of the
Castilian Crown’s economic and political power may benefit from the ability of GIS to
integrate many historical sources, visualize them in a geographical context, and apply
spatial and cartographic analysis to the data, providing further insight into the
127
historical narrative’s context on many different scales. This study’s visualization and
analysis of Castile’s Iberian transportation networks and population provides a critical
context toward answering this complex historical question.
128
References
Aldstadt, Jared, “B.4 Spatial Clustering,” In Handbook of Applied Spatial Analysis:
Software Tools, Methods, and Applications, edited by Manfred M. Fischer & Arthur
Getis, 279-300. New York: Springer, 2010.
Alexandria Digital Library. Website. University of California at Santa Barbara, 2005.
http://www.alexandria.ucsb.edu/ (accessed March 10, 2010).
___. Gazetteer Server Client Website. University of California at Santa Barbara, 2005.
http://alexandria.ucsb.edu/gazetteer/ (accessed May 18, 2010; no longer
operational).
Alonso García, David. Una Corte en Construcción: Madrid en la Hacienda Real de
Castilla (1517-1556). Madrid, Spain: Miño y Dávila, 2005.
Alvarez Junco, José & Adrian Shubert, Spanish History Since 1808. New York:
Oxford University Press, Inc., 2000.
Barreiros, Juan de. “Viajes.” In Viajes de extranjeros por España y Portugal: desde
los tiempos más remotos hasta comienzos del siglo XX, comp, trans, ed. J. García
Mercadal . Vols. 1-3. Valladolid: Junta de Castilla y León, 1999.
Batten, David. “Network Cities: Creative Urban Agglomerations for the 21st Century.”
Urban Studies 32 (1995): 313-327.
Berman, Merrick Lex. “China Historical GIS, Georeferencing Historical Placenames
and Tracking Changes over Time.” Presented at the Georeferencing Workshop,
Harvard University, Cambridge, Massachussetts, March 2008.
___. “Places in the Past: What’s in a Name?” Presented at PACSL GeoHistory
Network Conference, Philadelphia, PA, December 2005.
Bodenhamer, David J., “History an d GIS: Implications for the Discipline,” In Anne
Kelly Knowles, ed. Placing History: How maps, spatial data, and GIS are
changing historical scholarship, 219-233. Redlands, CA: ESRI Press, 2008.
Bolstad, Paul. GIS Fundamentals: A First Text on Geographic Information Systems.
White Bear Lake, MN: Eider Press, 2008.
Braudel, Fernand. The Mediterranean and the Mediterranean World in the Age of
Philip II. Vol. 1. New York: Harper & Row, Publishers, 1966.
129
Cabello, José Camacho. La población de Castilla-La Mancha (Siglos XVI,XVII,
XVIII): Crisis y renovación. Toledo: Junta de Comunidades de Castilla-La Mancha,
1997.
Calabria, Antonio, The Cost of Empire: The Finances of the Kingdom of Naples in the
Time of Spanish Rule (New York: Cambridge University Press, 1991).
Carr, Raymond, ed. Spain: A History. New York: Oxford University Press, 2000.
Christaller, Walter. Central places in Southern Germany.Translated by Carlisle W.
Baskin. New Jersey: Prentice Hall, Inc., 1966.
Christian, William A. Local Religion in Sixteenth-Century Spain. Princeton: Princeton
University Press, 1989.
Coppola, Emery, Jr., Owens, J.B., Szidarovszky, Ference. “Fuzzy Rule-Based
Modeling of Degrees of Trust in Cooperation-Based Networks: Close Research
Collaboration among Domain Experts (Historians) and Mathematical Modelers.”
(presented at the TECT Strategic Workshop, Technical University of Madrid, 25-26
Sept. 2008).
http://mapas.topografia.upm.es/Dyncoopnet/presentations/Madrid_fuzzy_owens.pdf
. (Accessed 27 February 2010).
Craglia, Max. “Building INSPIRE: The Spatial Data Infrastructure for Europe,” In
ArcNews, Summer 2010.
http://www.esri.com/news/arcnews/spring10articles/building-inspire.html (accessed
June 22, 2010).
Cram, George F. Cram’s Quick Reference Atlas and Gazetteer of the World.
Containing 105 Newly Engraved Maps and Over 40,000 Index Entries with the
Latest Areas and Census Statistics. Edited by Dr. Eugene Murray-Aaron. New
York: George F. Cram, 1906.
Dobbs, Gladys R. “The Indian Trading Path and the Colonial Settlement Development
in the North Carolina Piedmont.” PhD diss., University of North Carolina-Chapel
Hill, 2007.
Enggass, Peter M. “The Spanish league: A geographical conspiracy,” Journal of
Geography 70 (1971): 407-410.
130
ESRI ArcGIS 9.3 Desktop Help, “How Kriging Works,”
http://webhelp.esri.com/arcgisdesktop/9.3/index.cfm?TopicName=How Kriging
works (accessed August 25, 2010).
Federal Geospatial Data Committee Historical Data Working Group, “Digitising
Historical Maps and Charts,” FGDC Website,
http://www.fgdc.gov/participation/working-groupssubcommittees/hdwg/DigitMaps/Alexdig/?searchterm=ADL. (Accessed 21 Oct.
2009).
Fernández-Armesto, Felipe, “The Improbable Empire,” In Raymond Carr, ed. Spain: A
History, 111-128. New York: Oxford University Press, 2000.
Ford, Richard. Spaniards and Their Country. New York: George P. Putnam, 1852.
Frank, Andre Gunder, ReORIENT: Global Economy in the Asian Age. Los
Angeles/Berkeley, CA: University of California Press, 1998.
García Mercadal, Juan. Viajes de extranjeros por España y Portugal: desde los
tiempos más remotos hasta comienzos del siglo XX, comp, trans, ed. J. García
Mercadal . Vols. 1-3. Valladolid : Junta de Castilla y León, 1999.
Getis, Arthur, “B.3 Spatial Autocorrelation,” In Handbook of Applied Spatial
Analysis: Software Tools, Methods, and Applications, edited by Manfred M. Fischer
& Arthur Getis, 255-278. New York: Springer, 2010.
Goodchild, Michael F. “Statistical Perspectives on Geographic Information Science.”
Geographical Analysis 40 (2008): 310–325.
___. “Spatial Autocorrelation.” Concepts and Techniques in Modern Geography 47.
Geo Books, Norwich, U.K., 1986.
Goodchild, Michael F. & L.L. Hill, Introduction to digital gazetteer research,
International Journal of Geographical Information Science, 22 (2008): 1041.
Google Earth. http://earth.google.com/intl/en/ (accessed July 7, 2010).
González,Tomás. Censo de población de las Provincias y Partidos de la Corona de
Castilla en el siglo XVI. De Orden del Rey N. S., Madrid en la Imprenta Real, año
de 1.829. Re-issued as:Censo de Castilla de 1591: Vecindarios. Edited by Annie
Molinié-Bertrand. Madrid, Spain: Instituto Nacional de Estadística, 1984.
131
Gregory, Ian N. and Paul S. Ell. Historical GIS: Technologies, Methodologies, and
Scholarship. New York: Cambridge University Press, 2007.
Gregory, Ian N. and Richard G. Healey. “Historical GIS: structuring, mapping, and
analyzing geographies of the past.” Progress in Human Geography 31(2007): 638–
653.
Hibberd, Robert, Censo Gazetteer, [Unpublished gazetteer], 2008.
Hill, Linda, Georeferencing: The geographic associations of information. Cambridge:
MIT Press.
Hill, Linda, Qi Zheng, “Indirect Geospatial Referencing through Place Names in the
Digital Library: Alexandria Digital Library Experience with Developing and
Implementing Gazetteers.” In the author’s online index,
http://www.alexandria.ucsb.edu/~lhill/paper_drafts/ (accessed March 13, 2010).
Instituto Nacional de Administración Pública. Subdivisión en Partidos Judiciales de la
Nueva Subdivisión Territorial de la Península e Islas Adyacentes, 1834. Madrid:
Instituto Nacional de Administración Pública, 2000.
Kamen, Henry, Golden Age Spain. Atlantic Highlands, NJ: Humanities Press
International, Inc., 1988.
___, Spain 1469-1714: A Society of Conflict. New York: Longman, Inc. 1983.
Kantabutra, Vitit, A new type of database: Intentionally-Linked Entities—a direct
implementation of databases from the Entity/relationship model. Unpublished
PowerPoint presentation. http://progeny.isu.edu/~vkantabu/ILE/ (accessed
December 10, 2009).
___. Patent application of Vitit Kantabutra for Title: Intentionally-Linked Entities: A
general-purpose database system. 2009[b]. http://progeny.isu.edu/~vkantabu/ILE/
(accessed December 10, 2009).
Katz, Aaron, Matthias vom Hau, and James Mahoney. “Explaining the Great Reversal
in Spanish America: Fuzzy-Set Analysis Versus Regression Analysis.” Sociological
Methods & Research 33 (2005): 539-573.
Knowles, Anne Kelly, ed. Placing History: How Maps, Spatial Data, and GIS are
Changing Historical Scholarship. Redlands, CA: ESRI Press, 2008.
132
___. “GIS and History,” In Anne K. Knowles, ed. Placing History: How Maps, Spatial
Data, and GIS are Changing Historical Scholarship, 1-25. Redlands, CA: ESRI
Press, 2008.
Kuratowski, K. Introduction to Set Theory and Topology. New York: Pergamon Press,
1972.
Levine, Ned. CrimeStat: A Spatial Statistics Program for the Analysis of Crime
Incident Locations (v 3.2a). Ned Levine & Associates, Houston, TX, and the
National Institute of Justice, Washington, D.C., 2009.
López García. J.M. (dir.) El impacto de la Corte en Castilla: Madrid y su territorio en
la época moderna. Madrid, Spain: EUROCIT, 1998.
Lynch, John. Spain 1516-1598: From Nation State to World Empire. Cambridge:
Blackwell Publishers, 1994.
Mackay, Ruth. The Limits of Royal Authority: Resistance and Obedience in
Seventeenth-Century Castile. New York: Cambridge University Press, 1999.
Martín Benito, J.I. Cronistas y viajeros por el norte de Zamora. Benavente, Spain:
Centro de Estudios Benaventanos <<LEDO DEL POZO>> (CECEL – CSIC),
2004.
Mauro, Frédéric and Geoffrey Parker. “Spain.” In Charles Wilson and Geoffrey
Parker, eds. An Introduction to the Sources of European Economic History 15001800. London: Methuen & Co. Ltd, 1980.
Mapa Oficial de Carreteras. Madrid, Spain: Ministerio de Fomento, 2000.
Meijers, E. “Polycentric Urban Regions and the Quest for Synergy: Is a Network of
Cities More than the Sum of the Parts?” Urban Studies, 42(2005):765-781.
Minnesota Population Center, University of Minnesota. National Historical
Geographic Information System. http://nhgis.org/ (accessed June 22, 2010).
Nordlund, Carl, Social Ecography International trade, network analysis, and an
Emmanuelian conceptualization of ecological unequal exchange, Ph.D. diss., Lund
University, 2010.
Open Geospatial Consortium, “Best Practices Documents: Gazetteer Service Application Profile of the Web Feature Service Implementation Specification,”
133
Open Geospatial Consortium, http://www.opengeospatial.org/standards/bp 2006
(accessed March 13, 2010).
Organization of American Historians, LaPietra Report: A Report to the Profession,
2000, http://www.oah.org/activities/lapietra/final.html (accessed on July 3, 2010).
Owens, J.B., “Using GIS to Explore the Nonlinear Dynamics of Historic Systems,”
Paper to be presented at the meeting of The Royal Geographic Society in London,
UK, September 2010.
___.“A Multi-national, Multi-disciplinary Study of Trade Networks and the Domains
of Iberian Monarchies during the First Global Age, 1400-1800,” In SSPHS Bulletin,
33-2 (2008), 23-30.
___. "Toward a geographically-integrated, connected world history: Employing
geographic information systems (GIS)." History Compass, 5, 6 (October 2007):
2014-2040; doi: 10.1111/j.1478-0542.2007.00476.x.
___. By My Absolute Royal Authority: Justice and the Castilian Commonwealth in the
First Global Age. Rochester, NY: University of Rochester Press, 2005.
___.“Dynamic Complexity of Cooperation-Based Self-Organizing Commercial
Networks in the First Global Age (DynCoopNet): What’s in a name?” In David
Alonso García and Ana Crespo Solana, eds., Self-organizing Networks and
TradingCooperation: GIS Tools in the Visualization of the Atlantic Economic
Expansion (1400-1800), (forthcoming).
Pérez Asensio, Ester . et al., “Introducción: Proyecto DynCoopNet,” presented at the
Reunión Nacional de la IDE de España, Madrid, 24-25 de febrero de 2010.
Phillips, Carla Rahn. “Time and Duration: A Model for the Economy of Early Modern
Spain,” The American Historical Review, 92 (1987), 531-562.
___. Ciudad Real, 1500-1750: Growth, Crisis, and Readjustment in the Spanish
Economy. Cambridge, Massachussetts: Harvard University Press, 1979.
Phillips, Carla Rahn, and William D. Phillips, Jr., Spain’s Golden Fleece: Wool
Production and the Wool Trade from the Middle Ages to the Nineteenth Century.
Baltimore: The Johns Hopkins University Press, 1997.
Raines, Gary L., et al., “Incorporating Expert Knowledge: New Fuzzy Logic Tools in
ArcGIS 10,” ArcUser, Spring 2010, 8-13.
134
Ringrose, David. Transportation and Economic Stagnation in Spain, 1750-1850.
Durham, N.C.: Duke University Press, 1970.
___. Madrid and the Spanish Economy, 1560-1850. Los Angeles, CA: University of
California Press, 1983.
Ruiz, Téofilo F. Spanish Society, 1400-1600. New York: Longman, 2001.
Shekhar, Shashi and Hui Xiong, Encyclopedia of GIS. New York: Springer, 2008.
Smithson, Michael. “Fuzzy Set Inclusion: Linking Fuzzy Set Methods with
Mainstream Techniques” Sociological Methods & Research 33 (2005): 431-461.
Skinner, G. William. “Cities and the Hierarchy of Local Systems,” in The City in Late
Imperial China. ed. G. William Skinner. Stanford: Stanford University Press, 1977.
Tobler, Waldo. “On the First Law of Geography: A Reply.” Annals of the Association
of American Geographers, 94(2004): 304–310.
Tomaszewski, Brian. “The Reconstruction of Aztec Political Geography in the Toluca
Valley of Mexico.” MA Thesis, State University of New York, University Center at
Buffalo, 2005.
United States National Geospatial Intelligence Agency GEOnet Names Server.
http://earth-info.nga.mil/gns/html/index.html (accessed July 5, 2010).
University of Portsmouth, UK. Great Britain Historical Geographic Information
System. http://www.port.ac.uk/research/gbhgis/ (accessed June 22, 2010)
Vassberg, David E. The Village and the Outside World in Golden Age Castile:
Mobility and Migration in Everyday Rural Life. New York: Cambridge University
Press, 1996.
Verkuilen, Jay. “Assigning Membership in a Fuzzy Set Analysis.” Sociological
Methods & Research 33 (2005): 462-496.
Villuga, Juan P. Reportorio de todos los caminos de España. Medina del Campo,
1546. Reprint: New York: Kraus Reprint Corporation, 1967.
Viñas, Carmelo & Ramon Paz. Relaciones Histórico-Geográfico-Estadísticas de los
pueblos de España hechas por iniciativa de Felipe II: Reino de Toledo. Vols. 1-3.
Madrid: Instituto Balmes, de Sociologia; Instituto Juan Sebastian Elcano, de
Geografía; Consejo Superior de Investigaciones Científicas, 1963.
135
Whebbel, C.J.F. “Corridors: A theory of urban systems.” Annals of the Association of
American Geographers, 59 (1966): 1-26.
Yuan, May “Mapping Text,” in The Spatial Humanities: GIS and the Future of
Humanities Scholarship, ed. David J. Bodenhamer, John Corrigan and Trevor M.
Harris (Bloomington and Indianapolis: Indiana University Press), 109-23.
Zadeh, Lotfi A. “Toward a Theory of Fuzzy Information Granulation and its Centrality
in Human Reasoning and Fuzzy Logic.” Fuzzy Sets and Systems 90 (1997): 111127.
___. “Outline of a New Approach to the Analysis of Complex Systems and Decision
Processes.” IEEE Transactions on Systems, Man, and Cybernetics SMC-3 (1973):
28-44.
Zarco-Bacas y Cuevas, Eusebio-Julián. Relaciones de Pueblos del Obispado de
Cuenca. Cuenca, Spain: EXCMA. DIPUTACION PROVINCIAL DE CUENCA,
1983. Original publisher: Felipe II, Rey de España 1575.
136
Appendix A: Glossary
Areal Feature: a map feature, or object, such as a polygon, that covers a measurable
area.
Attribute data: information about an object’s attributes. Examples include the address
of a building, or the age of a human.
Boolean: a binary data format, denoting true or false. 1 represents true; 0 represents
false. No other data definitions exist when data are assigned a Boolean value.
Buffer: In GIS, a circle of equal distance around a geographic feature.
Central Place Theory: A geography theory that postulates a model of urban growth
and interconnection, based on the economic relations between populated places
established in a hierarchical relationship. The particular geographic reach of goods and
services from a place in a central location determines its degree of influence among its
surrounding populated places.
Centroid: The geographic center point of an areal feature (i.e., polygon).
Computing with Words (CW): the process of converting text-based descriptions of
phenomena into mathematical computations.
Conditional statements: in GIS, they are used to determine whether certain
geographic features pass a test. A combination of conditions can be tested for.
Crisp data: Precise or exactly defined data. Approximations do not fit this definition.
Disambiguation, of place-names: the process of making each place-name in a
gazetteer unique from all other variants.
Feature Type: the classification or categorization of a map feature or object.
Mountains, rivers, streams, populated places, and administrative units are examples.
Gazetteers use a feature type classification for all objects, in order to distinguish them
from each other.
Footprint: the area on the earth’s surface that an object inhabits; its geographic
location. This can be expressed in GIS by a point, line, polygon, or raster image. It can
consist of a series of geographic coordinates.
Fuzzy data: imprecisely-defined, incomplete, approximate, or vague data.
137
Fuzzy Rule-Based Modeling: as applied to this study, using fuzzy conditions to
determine membership of a geographic feature in a set or number of sets.
Fuzzy Set Theory: A generalization of set theory that allows each feature of a set to
have membership in more than one set at a time. It allows for the use of fuzzy,
approximate, or incomplete data in machine computing.
Gazetteer: A list of geographic coordinates assigned to textual attributes of those
coordinates. Coordinates, a place-name, and a feature type are the three essential
elements of a gazetteer.
Geographic Information Systems (GIS): a type of information system that assigns
geographic coordinates to attribute data for database storage, data integration,
cartographic visualization, and analysis.
Georeferencing: assigning a geographic location to any kind of data, thus facilitating
a map of the data.
Humanistic data: Data produced by a human being, with approximate or summary
definitions and categorizations.
Interpolation (GIS): Based on points of known measurement, it computes the most
likely measurement for points on a map that do not have known measurements. For
example, to create elevation models, a sample data set is gathered at a number of
points, and from these points the rest of the model is created.
Knowledge Organization System (KOS): Hill’s term for a system that organizes data
into a classification scheme (Hill 2008, 92, 113).
League: a general unit of distance. Used throughout Europe, it had many
measurements for its distance, such as 4.2 kilometers per league, or 5.6 kilometers.
Mechanistic data: Data produced and processed by a machine. Precision is required.
Parameter: the input data used in many computations. Specific data used to fill in the
input data requirements for an equation or computational algorithm. Many GIS tools
use parameters.
Place-name: The text-based label of a geographic location.
Point: a set of longitude and latitude coordinates on a map, representing a nondimensional location. In GIS, it is a type of vector data.
138
Polycentric Urban Regions: a geography theory that postulates that in certain cases,
an urban region will be dominated by no one city, but that a group of equallydominant cities will work together in a complementary and synergistic relationship to
accomplish the leadership roles of an urban area.
Polygon: a series of longitude and latitude coordinates that form a two-dimensional
shape that in GIS represents an object such as a building. In GIS, it is a type of vector
data.
Polyline: a series of longitude and latitude coordinates that form a one-dimensional
line that in GIS may represent an object such as a road, or a river. In GIS, it is a type
of vector data.
Raster: data in a grid format, with rows and columns. Digital photographs provide an
example. Each pixel represents the intersection of a row and column. Each pixel (also
called a cell) can hold data values that GIS can analyze.
Schema: the organization scheme chosen for a database. A database’s fields, data
types, and other elements are grouped into a schema (Shekhar, et al. 2008, 226).
Set: a grouping of objects with identical or similar attributes. Venn diagrams, for
example, visualize a number of sets and their relationships to each other.
Set Theory: the mathematics of sets.
Soft Computing: Computations using imprecise or approximate data.
Spatial autocorrelation: the measure of similarity in the attributes of objects that are
a given distance from each other, and how those attributes change in relation to the
objects’ increase or decrease in distance from each other.
Spatial data: data containing the geographic location of an object, or its footprint on
the earth’s surface.
Theory of Fuzzy Information Granulation (TFIG): Zadeh’s explanation of how
humans granulate, or generalize, information for quick use in summary format.
Variant, of place-name: version of a place-name that may be different over time for
the same geographic location. Also, similar names across multiple locations.
Vector Data: in GIS, this type of data consists of points, lines, and polygons, all
consisting of one or more sets of geographic coordinates.
139
Appendix B: Place-name Tables for Toledo and Cuenca Provinces
Toledo:
VillugaToledoNames
HistNames1834
MODERN_NAME
Ajofrín
Cebolla
null
null
Ajofrin
Cebolla
Alba Real de
Tajo
null
al monacid; al monacit
Almonacid
Almonacid
bogas
borrox
burujon
null
Borox
Burujón
null
Borox
Burujon
Cabañas de la
Sagra ó
Miralcazar
la venta de diezma; las
ventas d diezma
acebolla; cebolla
la venta estinel; la venta
etinel
null
cabañas; cavañas;
cavanes
ToledoRelaciones
Ajofrin
null
Albarreal de Tajo
Almonacid
Almonacid de
Toledo
Villanueva de
Bogas
Borox
Burujón
Cabañas de la
Sagra
caçarrubios
camarena
camines
çaranque
Cabañas de la Sagra
Casarrubios del Monte;
las Casas de los Rubios;
Casarrubios
Camarena; Carmarena
Camuñas
Carranque
cedilo; acedillo
censeya; ceseya
el tovoso
Cedillo
null
El Toboso; Toboso
Cedillo
null
null
el viso
escalona
fuen salida
gismonde; guismonde
guadalerce; la venta
guadaleca; la venta
guadalerza; la venta
guadalerce; la venta
guadlherça
holias; olias
El Viso
null
null
null
El Viso
Escalona
Fuensalida
Quismondo
Casarrubios del
Monte
Camarena
Camuñas
Carranque
Cedillo del
Condado
Seseña
El Toboso
El Viso de San
Juan
Escalona
Fuensalida
Quismondo
null
Olias
null
Olias
Guadalerzas
Olías del Rey
horgaz
Orgaz
Orgaz
Orgaz
140
Casarrubios del
Monte
Camarena
Camuñas
Carranque
hueca; huecas
junquillos
Huecas; Hocar
Yuncos; Palomequexo
Huecas
Yuncos
la calçada
null
null
la calçada
la mata
null
La Mata; San Pedro
la puebla
Puebla de Almoradiel
la puebla de don fadriqe
null
null
La Mata
Puebla de
Almoradiel
Puebla de Don
Fadrique
Huecas
Yuncos
La Calzada de
Béjar
La Calzada de
Oropesa
La Mata
La Puebla de
Almoradiel
La Villa de Don
Fadrique
la puente el arçobispo; la
puete el arçobispo
Puente del Arzobispo
Puente del
Arzobispo
El Puente del
Arzobispo
la venta
madrilexos
maqueda
marquaraque
null
Madridejos
Maqueda; Maceda
Mascaraque
null
Madridejos
Maqueda
Mascaraque
Las Vegas (de San
Antonio)
Madridejos
Maqueda
Mascaraque
miguel esteban; miguel
estevan
Miguel Esteban
Miguel Esteban
Miguel Esteban
null
Mocejon; Mocejón
Nambroca; las
Nambrocas
Noves; Nove; Noveldes
Ocaña; Olcañia
null
Mora
Mocejon
Mora
Mocejón
Nambroca
Noves
null
Oropesa
null
Paredes
Nambroca
Novés
Ocaña
Oropesa
Paredes de
Escalona
San Silvestre
Santa Cruz de la
Zarza
San Silvestre
Santa Cruz de la
Zarza
talavera
San Silvestre
Santa Cruz de la Zarza;
la Zarza
Talavera de la Reina;
Cobriga; Talabriga;
Tavira; Elbora; Libora;
Talavera; Aguas; el Ber;
Talabrica; el Bora
Talavera
Talavera de la
Reina
tembleque
Tembleque
Tembleque
Tembleque
mora
mosejon; musejon
nambroca
noves
ocaña
oropesa
paredes
san silvestre; sant
silvestre
santa cruz dla çarça
141
toledo
Torrijos
Toledo
Torrijos
Toledo
Torrijos
la venta
villa cañas; villa cañis
null
Villacañas
villa seca
Villaseca de la Sagra
null
Villacañas
Villaseca de la
Sagra
villamiel
yepes
yevenes; yvenes
yllescas
la venta
Villamiel
Yeles
Yebenes
Illescas
Yuncler
null
Yepes
null
Illescas
Yuncler
Toledo
Torrijos
Venta de la
Zarzuela
Villacañas
Villaseca de la
Sagra
Villamiel de
Toledo
Yepes
Los Yebenes
Illescas
Yuncler
Cuenca:
VillugaCuencaNames
NamesCuencaRelaciónes
HistNames1834
Modern_name
alarcon
Alarcón
valadiego
null
almodovar
alcaçar de huete
Almodóvar
Alcázar del Rey
Alarcon
Albaladalejo del
Cuende
Almodovar del
Pinar
Alcazar del Rey
Alarcón
Albaladejo del
Cuende
Almodóvar del
Pinar
Alcázar del Rey
arcuaz; arquas
null
Arcas
Arcas
alguisuellas
null
Arguisuelas
Arguisuelas
barchi; barchin
Barchín del Hoyo
Barchin del Hoyo
Barchín del Hoyo
vilinchon; valenchon
Belinchón
buenache
Buenache
campillo
null
Belinchon
Buenache de
Alarcon
Campillo de
Altobuey
Belinchón
Buenache de
Alarcón
Campillo de
Altobuey
cardenete
null
carraschosa d huete
null
el castillo
Castillo de Garcimuñoz
Cardenete
Carrascosa del
Campo
Castillo de
Garcimuñoz
Cardenete
Carrascosa del
Campo
Castillo de
Garcimuñoz
142
cervera
null
null
Cervera
Chillaron de
Cuenca
Cervera del Llano
Chillarón de
Cuenca
chillaron; chilaron; gillaron
cuenca
alcañavale; cañavete
el histo; el hito
Cuenca
El Cañavate
null
Cuenca
El Cañavate
El Hito
Cuenca
El Cañavate
El Hito
el pedernoso
El Pedernoso
El Pedernoso
El Pedernoso
el provencio
El Provencio
El Provencio
El Provencio
fuentes
null
Fuentes
Fuentes
agua valdon; guavaldon
Gabaldón
Gabaldon
Gabaldón
honruvia
la venta talayuelas
Honrubia
Hontecillas
Honrubia
Hontecillas
horcajada
null
la huerta
null
Horcajada
Huerta del
Marquesado
Honrubia
Hontecillas
Horcajada de la
Torre
Huerta del
Marquesado
?
Iniesta
Iniesta
el alberca; el alverca
La Alberca de Záncara
La Alverca
la cierva;
null
La Cierba
parra
null
La Parra
La Cierva
La Parra de las
Vegas
la pesquera; la pexquera
null
Pesquera
La Pesquera
la solana
null
null
la laguna
null
La Laguna
La Solana
Laguna del
Marquesado
las mesas
Las Mesas
Las Mesas
Las Mesas
mira
Mira
Mira
?
Monreal
Monreal
Mira
Monreal del
Llano
la mota; la mota el cuervo
null
la montilla; la matilla
null
Mota del Cuervo
Motilla de
Palancar
Mota del Cuervo
Motilla del
Palancar
noales
null
Noales
Nohales
nueva; ñeva
null
Noheda
Noheda
palomera
null
Palomera
Palomera
reyllo
Reillo
Reillo
Reillo
la venta de salcedon; la venta
de cacidon
null
Sacedoncillo
Sacedoncillo
çaelises; sahelizes
Sahelices; Sailices
Saelices
Saelices
san clemente
San Clemente
la xarilla
null
San Clemente
San Lorenzo de la
Parrilla
San Clemente
San Lorenzo de la
Parilla
143
Iniesta
La Alberca de
Záncara
s. maria dlos llanos; s.maría
delos llanos
Santa María de los
Llanos
Santa María de los
Llanos
Santa María de
los Llanos
tarancon
Tarancón
Tarancon
Tarancón
toralva;torralva
Torralba
Torralva
Torralba
la venta la hosilla
null
Tortola
Tórtola
valde moro
null
villa escusa de aro
Villaescusa de Haro
Valdemoro
Villaescusa de
Haro
Valdemoro-Sierra
Villaescusa de
Haro
el villar de cañas; villar de cañas
null
Villar de Cañas
Villar de Cañas
el villar de domingo; el villar
domigo garcia; el villar de humo
null
Villar de Domingo
García
Villar de Domingo
García
el villar del horno
null
el villar de saz
Villar del Saz de Arcas
Villar del Horno
Villar del Saz de
Arcas
Villar del Horno
Villar del Saz de
Arcas
villa ruvia
Villarrubio
Villarrubio
Villarrubio
la venta
Villavieja
null
valverde
Villaverde
Villaverde
Villas Viejas
Villaverde y
Pasaconsol
villora
Víllora
Villora
Villora
vindel; vindiel; vendel
null
Vindel
Vindel
çafrilla
Zafra
Zafrilla
Zafrilla
144
