You are on page 1of 24

Studies of XVI Century Masonry Structures in Oaxaca

Jonathan Dessi-Olive
Bachelor of Science
School of Architecture
University of Minnesota
June 2009 - February 2010

Abstract

Churches from the XVI c. built by indigenous people under the direction of Dominican
monks are found in the Mixteca region in Mexico. These churches follow the principles of
Gothic construction making them unique places to research ancient building methods in
the Americas. While the principles of fabrication, geometry and design of Gothic churches
in Europe have been studied extensively, there has been almost no investigation of these
Mexican examples. This project focuses on Santo Domingo Yanhuitlan, in Oaxaca,
because its complex rib vaults provide an excellent context to use as a case study.

The goal, here, is to analyze the construction of Yanhuitlan’s vaulting system, focusing
on the links between geometry and design, while attempting to gain insight on Gothic
constructive principles and their origins. Explorations of the vaults at Yanhuitlan are
carried out through hand-drawn and computer-aided design studies. This project balanced
iterations of these drawings and analysis rooted in the theories on Gothic buildings in
Europe proposed by scholars such as Viollet-le-Duc, Robert Willis and Enrique Rabasa.

The process used for this analysis through the lens of Viollet-le-Duc’s proposed principles
is applicable to other Gothic vaults. It provides evidence that the form of the vault was a
consequence of the constructive methods used by XVI c. masons. This study also lays
groundwork to draw other conclusions regarding the Gothic methods of construction and
why they were a consequence of a constructive necessity.
Gothic as a construction method and evolutionary consequence

The term architect is relatively new within to the history of architecture. Although rooted
in Greek, literally translating as “master builder”, the term as we know it today originates
around the XVII c. around the time the École des Beaux-Arts was founded in France.
Prior to this, people with a limited knowledge of construction methods carried out both the
design and construction of buildings. Gothic buildings, which irst emerged in the XII c.
and often characterized by their impressive structure, are no exception to this.

Gothic churches are unique because of their


vaults, which employ speciic methods of
construction accomplished through the use
of a framework of stone arches, or ribs. What
made the use of this technology so innovative
was that the ribs transferred all of the loads
on the imposts to vertical piers, completing a
structural system that was not dependant on
the particular construction of the outer wall.
Thus, the vault is reduced to the plan of the
piers and ribs that spring from the impost as
independent arches. With the exception of
the diagonals, all of the arches that make
up the vault are pointed or “broken” arches.
The surface of the vault itself resembles a
thin panel that is draped over the rib system. Diagram of the plan, pier, and rib structural
system of Gothic vaults.

These technologies are considered an advancement on the Romanesque principles of


groin and barrel vaulting. Viollet-le-Duc describes the Gothic as a “new” architecture
based entirely on the methods that immediately preceded it. Gothic architecture was
carried out and established by those who practiced it:

Moreover, this spirit and outlook can ind no stopping place, for no one who has
ever established a principle based upon a valid chain of reasoning can then just
turn around and say to reason: “You cannot take me any farther!” Even the builders
of the eleventh century who were working in the shadow of the cloister recognized
this fundamental principle. A hundred years later, no one at all could any longer
have prevented Romanesque architecture from becoming the new architecture that
came to be called “Gothic.” This new architecture was the inevitable consequence
of the logical development of the old.

(Selections from the Dictionnaire Raisonne, pg 133)


Groin Vault Diagram Early-Gothic Vault Diagram

The inevitability of these “new” principles is a curious point. Viollet-le-Duc claims that Gothic
architecture, which spreads across much of Europe, was an evolutionary consequence
of constructive methods. Thus, the Gothic principles could never have existed in a world
without Romanesque buildings. This contention is certainly not too far-fetched. A quick
comparison of a Romanesque groin vault with a Gothic rib vault would reveal some
similarities, especially in terms of their general form. A groin vault is essentially two semi-
circular barrel vaults that intersect each other on perpendicular axes. The intersections
of these two create the groins. These groins clearly appear to be antecedents of the
diagonal ribs of the Gothic vault. The primary difference between the latest examples
of groin vaults and the earliest examples of rib vaults in Europe, is that a sharply raised
set of arches (ribs) runs along and deines the diagonal groin lines. In fact, Robert Willis
suggests that ribs, such as those diagramed above, were used by the Romans, but more
typically as decorative construction. These raised arches eventually became the major
structural feature of Gothic Vaults.

While Viollet-le-Duc’s contention is certainly valid in explaining the lineage of the


Gothic principles, it does not explain its widespread use across Europe and beyond.
The Romanesque building principles were used throughout Europe and were applied in
similar ways - in terms of their constructive principles. The interesting phenomenon is that
the progression from Romanesque to Gothic reveals very similar adaptations from one
to the other. Robert Willis discusses this issue with some depth in his Remarks on the
Architecture of the Middle Ages, especially of Italy. In his introduction he presents three
explanations of the emergence of the Gothic:
One class of writers wishing to conine all Gothic styles to our country, and treating
the continental specimens with contempt, while others seem to expect to ind the
origin of Gothic architecture in some individual building, or evade the history of its
invention by asserting it to have been derived from the East; and a third hypothesis
is, that different nations beginning from the classical specimens, and gradually
transforming them, were all led in the course of ages to the same Gothic style,
although by different roads.

If this be taken merely as a phrase expressing the vicissitudes of architecture


in each country, without regard to the mode in which they took place, it is
undoubtedly true; and we may add, that after discovering the identical Gothic style,
and wondering from it in different manners, they were all led to the same revived
classical form, for a chronological review of the buildings of any one country in
Europe compared with another, presents exactly this phenomenon; they start
from the classical style and return to it and midway are found practicing the same
Gothic, and all the intermediate steps are different; but this was not done without
mutual communication.

(Remarks on Architecture of the Middle Ages, pg 6-7)

While the vaults across Europe vary slightly based on their country of origin, their overall
principles of design and construction are generally common throughout. Willis explains
in the second volume of his Remarks that this commonality is due to the fact that the
principles deining the design of Gothic buildings are actually quite simple. Willis would
most likely contend that while there is a clear lineage between the groin vaults and the
early rib vaults, the use of these “raised arches”, as described before, was not a signiicant
departure from the precedent. The ribs were a logical solution for emphasizing and more
clearly expressing the form of the groin vault. In fact for the most part, the Gothic vaults
continued to be built with very little change up until the very latest periods of rib vaulting,
just prior to the Renaissance. In these late periods, Gothic vaults of Europe are evidence
that the basic principles of rib vaulting were so well understood, they could be applied to
create vaults with much more extravagant arrangements of ribs. This eventually let to the
use of the pointed arch in the lateral arches as well as the ridge and intermediate ribs that
in turn produced far more elaborately shaped vaults.

Site-plan of Santo Domingo Yanhuitlan


Case Study: Santo Domingo Yanhuitlan

Digital rendering of Yanhuitlan

This discussion now turns to a speciic example of a Gothic church found in an unlikely
place. The Dominican church of Santo Domingo Yanhuitlan is a masonry temple with
large dimensions that involved relevant constructive challenges during the XVI c., notably
because it was built using Gothic construction methods. Yanhuitlan is located in the
Mixteca Region, about one hundred miles outside the capital of Oaxaca. The Mixteca
region is a cultural, economic and political region in southern Mexico that crosses parts
of the states of Puebla, Guerrero and Oaxaca. It is a rather isolated region both in terms
of communication and culture, primarily because it is located at the convergence of
two of the country’s mountain ranges, the Neo-Volcanic Belt and Sierra Madre del Sur.
This region of Mexico is a unique place
to research ancient building methods in
the Americas. It contains a number of
churches from the XVI c. built by the hands
of indigenous people under the direction of
Dominican monks. These heroic churches
follow the Gothic tradition of construction
and contribute to Oaxaca’s rich architectural
history. While the principles of fabrication,
geometry and design of Gothic churches in
Europe have been studied extensively, there
has been almost no investigation of these
divine examples in Oaxaca. This project focuses on the church of Santo Domingo
Yanhuitlan because its complex rib vaults provide an excellent context to use as a case
study. The goal is to study the building construction of Yanhuitlan, focusing on the links
between geometry and design, while attempting to gain some insight on the origins of the
principles of Gothic construction.
The analysis of Yanhuitlan begins by looking at the underlying geometrical relationships
between the elements of the vault. Robert Willis, argues that there must be a simple
geometrical system that serves as the basis of laying out these elements. This system is
necessary according to Willis especially when it comes to locating the intermediate ribs.
The following exercise will demonstrate that the plan of a typical bay at Yanhuitlan is no
exception to this.
The geometrical manipulations begin with a square with diagonal ribs. Let ABCD
be a perfect square. AC and BD are the diagonal ribs. If one connects the midpoints
of the square ABCD, the result is a smaller square EFGH. EG and FH are the ridge
ribs.

A F B

E G

D H C
By repeating this process of connecting midpoints three times, the results are three
even smaller squares JKLM, NOPQ, and RSTU. Each of the points of the square
R, S, T, and U are the bosses of the diagonal ribs.

A F B

J O K

S T

E N P G

R U

M Q L

D H C
Next, line segments are drawn from the spring points ABCD through these newly found points
RSTU. Segments AR and AT are drawn through and extended to the next ridge rib. In other
words, AR is drawn through EG and terminates at FH. These line segments establish the angle
in plan of the intermediate arches, or tiercerons, and where they intersect the ridge ribs. These
intersection points in turn become the location for the bosses.

A F B

S T

E G

R U

D H C

Once all four spring points are treated in the same manner, the design of the vault (in
plan) is complete. The resulting diagram is a sequence of squares with a four-pointed star
over it. Overlaid on the plan of Yanhuitlan, it becomes clear that the person in charge of
the design of the church used a similar process to layout the elements of the vault in plan.

A B

S T

R U

D C
Now that the position of each rib has been found via these simple geometrical manipulations,
the next step in the analysis is to determine the height of the keystones on the ribs.
The curvature of the ribs should be determined in section. At Yanhuitlan, the transverse
and former arches – which have the same curvature – present a major complication
in the understanding of the cross section of the vault. The height of the keystones of
these elements appear to have fallen victim to the “experimental method” inherent to
Gothic vault construction. Beyond that however, the lateral arches of a Gothic vault are
signiicant because according to Viollet-le-Duc, they are a “vital, free and elastic feature”
of the frame. He proposes a convincing method for determining an approximate height of
the transverse and former arches based on the plan in his Dictionnaire Raisonne.

Let ABCD be a perfect, or almost perfect, square (for the approximation would
sufice); this square comprehends two bays of a nave, AE, BF, and EC, FD.
The diagonals AD and BC give rise to this vault. Now these two diagonals are
the diameters of two perfect half circles, rotated on the plan. Being of the same
diameter, these two half circles necessarily meet at point G, which is the master
keystone. […] Taking a length shorter than the radius GA, but longer than half of
the width of the nave AB, and carrying this radius to the perpendicular HK, we
trace the broken arch AKB: this is the transverse arch of which the horizontal
projection goes to AB or CD.

(Selections from Dictionnaire Raisonne, pg 149-50)


Viollet-le-Duc suggests this method was more commonly used to obtain the curvature of
the formwork for the construction of stone arches. Thus, if the purpose of this research
were to design and build a Gothic vault, the use of this method would be suficient for
its construction. However, this study is analyzing existing conditions of a speciic Gothic
vault, and requires a method more suited to this approach. Viollet-le-Duc proposes a
second method of manipulating the section based on the size of the voussoirs of the
arch. This method describes what he believes was employed by builders to determine the
curvature of the lateral ribs.

They traced the two semicircles: the intrados ABC and the extrados DEF. Then,
on the diameter AC, they tried to ind the center, O’, of a circular arch uniting point
A of the intrados to point E of the extrados of the semicircular arch. Placing a joint
at EG instead of a keystone, in order to avoid the effect of equilibrium, they cut
the stones of the new arch, AE, following the normal lines of the curve AE – that
is to say, tending toward the center, O’. If breakage still occurred in the transverse
arches thus composed of two diagonal curves, AE, the builders proceeded with
the arch, AE, as with the semicircular arch; that is to say, they moved the center,
O’, back to O” on the diameter, in a way that resulted in an arch uniting point A
with point G.

(Selections from Dictionnaire Raisonne, pg 142-43)


This alternate method is clearly more suited for testing existing conditions of the transverse
arches at Yanhuitlan. Therefore, we must use this alternate method to determine if there is
a logical process for drawing the transverse arch of a Gothic vault to better understand it
in section. As-built documentation has been graciously provided by the Instituto Nacional
de Antropología e Historia (INAH).

We are trying to determine the curvature of the transverse arch, which lies
perpendicular to the nave. The known points are A, and C of the intrados and
D and F of the extrados. The section parallel to the nave provides crowns of the
intrados and extrados of the transverse arch: E and G, repectively. Let OG be
an imaginary line segment between the keystone G and point O, which is found
by projecting G down to the level of the impost. Next, semicircles ABC and DEF
can be traced on the section. We can then carry on in the same manner as before
to ind O’ which can be mirrored on both sides of O to ind the origins of the two
segmented arches AE and EC for the intrados, and DG and GF for the extrados.

Section Parallel to Nave Section Perpendicular to Nave


Showing Former Arch Showing Transverse Arch

E G G
E E
B

A C
D A O O’ C F

G
E

Assuming the former and transverse arches have identical curvatures, we are able to
compare the curvature of the transverse arch determined, using Viollet-le-Duc’s method
with the rough outline of the former arch provided in the documentation from INAH. By
overlaying the two, it is clear that Viollet-le-Duc’s method can successfully determine
the curvature of the transverse and former arches of an existing Gothic vault. With the
curvature of the transverse and former arches conirmed using Viollet-le-Duc’s proposed
methods, the other elements of the vault can be analyzed, beginning with the diagonal
arches.
Traditionally, the diagonals are known to be the only arches of a Gothic vault to have a
semi-circular curvature. A simple manipulation using descriptive geometry can be used
to verify this at Yanhuitlan. By translating the diagonal rib in plan into a section elevation,
the correct curvature can be visualized. Next a horizontal line (drawn in solid red) is
drawn across, connecting the imposts and a vertical line (also solid red) is drawn from the
keystone to midpoint of the horizontal line. Finally a circle is drawn in a thick red dotted
line to show that the curvature of the vault is the same as half of this circle.

A22
rota
punta
punta
rota
B1

The ridge ribs span the vault from the keystone of a lateral arch to its opposite, and through
the master keystone, at the intersection of the diagonals. Little has been documented in
terms of there being a clear logic to the curvature of these arches. What can be assumed
is these elements are just as “vital, free and elastic” as the lateral arches. However,
the curvature of the ridge ribs depends on the lateral arches. It is simply a segmented
arch through three points. While there may exist Gothic vaults that were designed with
a clear logic for these elements, it is safe to say that at Yanhuitlan, the ridge ribs are the
resultants of the former and transverse arches.
Similarly, the intermediate ribs, or tiercerons, are also resultants of the other elements.
While there are two different types, or curvatures of tiercerons at Yanhuitlan, they are
very similar in terms of how their curvature can be determined. Earlier, we found how
the tiercerons are located in plan. The tiercerons are located in section by translating
their intersection with ridge ribs and the spring points of the diagonal ribs. The curvature
of these elements simply follow the surface of the vault that is created by the lateral,
diagonal and ridge ribs. Once again, there may be an embedded logic to the tiercerons,
but to determine this at Yanhuitlan, additional measurements and further analysis would
be necessary to draw any solidiied conclusion about them.
Conclusion

The plan of a typical bay of the vaults at Yanhuitlan is essentially a four-pointed star
vault on a square plan. This coniguration, according to Enrique Rabasa, is very common
among XVI c. vaults. Although this is not a strict coniguration for Gothic vaults, it makes
the initial design of the vault a more straightforward process, and in turn allows for a
simpler construction process. Earlier, we found that the design of Yanhuitlan involves
the manipulation of this initial square plan via geometrical relationships. This simplicity
seems to be inherent to the basic principles of Gothic architecture. The case study at
Yanhuitlan proves to us that even in the later periods, Gothic buildings followed these
basic and fundamental principles of design and construction. This is likely due to the fact
that Gothic Architecture, from its inception was a creation of common people, as Viollet-
le-Duc explains:

Gothic Architecture was in its entirety a creation of lay people, and it stemmed from
nothing else but a rigorous analogical application of the system of construction
inaugurated by the Romanesque architects. Those who imagine that Gothic
architecture is some kind of oddity or eccentricity or exception in history fail […] to
understand its elementary principle.

(Selections from the Dictionnaire Raisonne, page 133)

Based on what is known about the construction of Yanhuitlan, the use of these
“elementary principles”, as Viollet-le-Duc calls them, would have been the only
option based on the existing conditions, and the workers available. It is known
that the construction of Mexican Gothic churches of this time were built under the
supervision of, and often designed by the monks. These monks arrived from Spain at
a time when Gothic was the dominating architecture. They had been immersed and
probably moved by these buildings and it was only natural for them to erect similar
structures in the “new world”. Yanhuitlan required very little skill on the part of the person
charged with the task of designing it, so it is of no surprise that a square plan was chosen,
if nothing else because of these elementary concepts.

The geometrical simplicity of the design is one of the inherent beauties of the Gothic
as a system: the “designer” determines which elements should be included (tiercerons,
ridge ribs, etc.) and lays them out schematically, in plan, using basic geometries and
common tools. Thus, with just a pencil and a straight edge the monks in charge of hiring
the indigenous people of the Mixteca region to build the structure, could quickly and
easily design the rib system in plan using a similar method as was found in the analysis
of Yanhuitlan.

The ridge ribs and tiercerons present some problems. Robert Willis believes there should
be a method that would be used from a constructive standpoint to control the curvature
of the tiercerons. He argues that, since the number of tiercerons varies from one vault to
the next, much of the effect and character of the vault depends on the curvature of these
elements. Based on the indings at Yanhuitlan however, it seems more likely that the
tiercerons are the main focus in plan. They are more dependent on the other elements
in section, than the other way around. Viollet-le-Duc’s view of the former and transverse
arches as “vital, free and elastic” seems much more itting of the spirit of the constructive
method.

The main complication that arises is in terms of the curvature of the ribs. As was found in
the analysis of this paper, only the diagonal ribs form semi-circular arches. All the others
form pointed, equilateral arches, or have curvatures that depend on the other elements.
Theorists throughout history have attempted to understand Gothic structures in terms of
their cross-sections in search for some sort of order, or simpliied geometrical system that
would help manage the curvature of the ribs. As we have seen, Viollet-le-Duc proposed
a very convincing method for inding the correct curvature of the former and transverse
arches. His method based on the size of the voussoirs seems to validate the existing
conditions at Yanhuitlan and seems worthy to test on other Gothic structures.

It is believed that pointed equilateral arches, like the lateral arches at Yanhuitlan, were irst
invented to prevent sag, which was common to Romanesque vaults. The adoption of the
“broken arches” was the result of observations made by the builders of the deformations
they found. Therefore, these arches were much more of a reaction to a need. The builders
of Yanhuitlan were limited to the only method available to them: that of experimentation.
The “true revolution in the art of building” that the pointed arch represents is due to the
fact that it was a consequence of a constructive necessity. In other words, the result was
simply a rigorous application of the construction system that preceded it.

During construction, those who worked on a Gothic building had very little to work with: at
most a plan, according to Rabassa. He remarks, that “the medieval builder did not know,
nor did he need to know how to make a cross section of the vault” thus reducing the
general design of the vault to the plan and the height of the keystones. This ignorance,
if it may be called that, is what makes buildings such as Yanhuitlan, fundamentally
Gothic. The uncertain and experimental nature of the process is exemplary of how
different architecture, as a practice, was before the architect. Architecture, was about
this rigorous application of known construction methods. It was about learning through
making; experimenting, failing, and slowly progressing by learning from the failures. It was
entirely dependent on the particular construction as well as the methods and resources
that governed and limited the structures. These governing principles were the evolution
of a constructive method. Lance LaVine would likely support this statement. In his book,
Mechanics and Meaning in Architecture, he states:

This particular use of technology does not automatically replace old techniques
with new, more eficient means to accomplish tasks, as some would profess.
Architectural technology gathers up its history in its progress […] The signiicance
of these technologies might be better deined as an accumulation of the human
inventiveness rather than as a linear progression of measurable eficiency.”
(pg 12-13)

In this analysis we were able to ind a relevant and accurate method for locating the
vault’s elements in plan as well as apply a method proposed by Viollet-le-Duc to ind the
curvature of the lateral arches through drawing. These exercises gave us insight into the
design process of those who participated in the construction of Gothic vaults. It would
be dificult for architects and builders today to imagine a process even remotely similar
to that of the Gothic practitioners. Some might argue that the founding of the École des
Beaux Arts and with it the establishment of the discipline of Architecture, in some ways
made it inconceivable to design and construct a building without drawing it beforehand,
as it is thought to be the case with buildings like Yanhuitlan. The École, made drawing the
way we understand and communicate the ideas of a building prior to construction and this
remains a practice that exists in the discipline today.
Bibliography

Ibarra-Sevilla, Benjamin. Abstract for Studies in Fabrorum Geometry: Stone cutting and
its application in Architecture. 2009.

Kubler, George. Mexican Architecture of the XVI Century. Yale University Press, 1942.

LaVine, Lance. Constructing Ideas. Kendall Hunt Publishing, 2008.

LaVine, Lance. Mechanics and Meaning in Architecture. Minnesota: University of


Minnesota Press, 2001.

Rabassa, Enrique. Gothic and Renaissance Design Strategies in Stonecutting.

Rabassa, Enrique. Principles and Construction of Groined Vaults.

Sanabria, Sergio. From Gothic to Renaissance Stereotomy: The Design Methods of


Philibert de l’Orme and Alonso de Vandelvira. Technology and Culture, Vol. 30, No. 2,
Special Issue: Essays in Honor of Carl W. Condit (Apr., 1989), pp. 266-299. Available in
JSTOR website.

Willis, Robert. Remarks on the Architecture of The Middle Ages Especially of Italy.
Cambridge: Pitt Press, 1835. Pg 1-112.

Willis, Robert. On the Construction of the Vaults of the Middle Ages. London: Longman,
Brown, Green, and Longmans, 1842.

Viollet-le-Duc, M. Dictionnaire Raisonne de L’Architecture Francaise du XIe au XVIe


Siecle. Paris: B. Bance, Editeur, 1854. Pg 103-193.
Appendix
D1
nte
ja
pin
4
A2 D1

24
C

You might also like