Ancient Technical Manuscripts: the Case of 17th-century Portuguese Shipbuilding Treatises

During the 16th- and 17th-centuries European seafaring
underwent an incredible transformation driven mainly
by the exploration of newly discovered lands, the contact with
previously unknown cultures, and the increase in maritime
commerce. The social changes that resulted from this cultural
revolution affected the long chain of events entailed by the
construction of oceangoing ships and determined a number of
technical innovations in the construction of ships. From an
original oral tradition, where apprentices learned from masters
the intricacies of shipbuilding techniques; this process evolved
into a more formal field as masters began to follow guidelines,
materials used, and construction sequences in a more systematic
way, forming a corpus of information that was soon compiled
in manuscripts known as shipbuilding treatises.
The first shipbuilding treatises, understandably, were written
by mathematicians, priests and other learned men, reflecting a
reality where shipbuilders probably were largely illiterate. Rich
in technical descriptions, shipbuilding treatises play a key role
in Nautical Archaeology both for scholars and students. Scholars
access these manuscripts for several reasons. For example,
working on the reconstruction of sunken ships, they can provide
vital information for reassembling fragments and damaged
timbers of ship remains recovered from underwater excavations.
Also, their contents often help our understanding of shipbuilding
techniques. In addition, they can be used to compare different
construction traditions both from geographical and
chronological standpoints. Moreover, they are great sources
for understanding the evolution of shipbuilding (Figure 2 shows
our treatises browser).
Nautical Archaeology students—although not carrying out the
complex tasks of ship reconstruction—are exposed to treatises
early in their studies. For them treatises are a good source to
understand the basic terminology and concepts they will be
using during the rest of their studies. Current teaching practices
are constrained to browse physical copies of the original
sources; with obvious restrictions such as limited number of
copies and access, or unknown language—depending of their
provenance, they were written in different languages, and
difficult technical terminology. In fact, treatises inherit most
of the limitations of printed books.
Providing ways in which shipbuilding treatises can be used
digitally by both scholars and students is an attractive
interdisciplinary effort for a number of reasons. First, is the
opportunity to investigate the variety of ways in which
manuscripts’ contents (both texts and illustrations) can be
structured and classified. Second, is the opportunity to make
original-source material available at the location of an
excavation; the treatises description of the construction of the
physical objects can provide valuable information about
fragments of ship timbers that are recovered from an excavation
site. This effort clearly draws techniques from the earlier
projects in the digital humanities that examine the characteristics
of digital representations of paper-based texts. However the
linkage to physical artifacts opens up additional possibilities
and considerations.
Digital humanities projects involving manuscripts or printed
texts have been related, for the most part, to literature and
historical records; some Know: Responding to the
Computational Transformation of the well-known examples
include the Canterbury Tales Project.1, the Rossetti Archive2,
and the Perseus Digital Library 3. We have been involved in
creating collections of this form as well for Cervantes4,Donne,5
and Picasso6. In the context of Nautical Archaeology, the
manuscript of Michael of Rhodes captures the knowledge
obtained by this 15th century seaman during his 4- decade-long
career7. The presentation of the manuscript, oriented to a general
audience, hints at the value that such materials will have to the
professional archaeologist when representations and tools are
provided that meet his scholarly needs.
Until relatively recently, ships were the most advanced and
complex transportation means designed. Nautical treatises hold
the key to understanding their technical complexity. In essence,
the collection of treatises represents the technical manuals
describing the components, their use, and the steps taken in
manufacturing of the ship. Several characteristics of treatises
make understanding them a very challenging task. Language
is a major problem; in order to better understand their contents,
it is necessary to provide translations and explanations of concepts, pieces, and sequences. Beyond the multiple languages
in which the treatises were written, they also come from diverse
geographical locations and span centuries, making terms,
concepts, and descriptions difficult to understand. Different
units and standards of measurement—a key aspect in technical
descriptions—raise problems about not only comparing treatises
with different provenance, but also translating them into modern
scales; units of measurement used in the treatises are not
necessarily the ones used by archaeologists to measure
recovered evidence. To tackle these problems, we have
developed a multilingual glossary, in which terms include their
corresponding translation and definition into ten languages
(which can be expanded as needed). The incorporation of
“roles,” enables us to expand characteristics related to the terms,
for example spellings and synonyms. Our framework allows
multiple values per role as well the addition of more roles as
they are required (see figure 1).
However, despite common features shared by physical
fragments and their corresponding descriptions in the texts,
fragments obtained from individual ships have important
differences because of the differing physical conditions that
they have been exposed to. Damaged and incomplete ship
remains require the adoption of an encoding scheme to describe
and quantify uncertainty; textual descriptions do not encompass
“uncertainty.”
In the previous paragraphs we have briefly outlined the
relationship between physical archaeological evidence and
written descriptions in the treatises. However, treatises in
themselves have properties that make them unique. For
example, an important question is how similar or different are
treatises in terms of the sequences, construction techniques,
and materials used. An initial approach would suggest that
probably the encoding used in their description could be used
to quantify the degree of similarity.
Since treatises are “technical manuals,” illustrations are essential
in their understanding; therefore, we adopted a two-step process.
First, illustrations have to be segmented, an illustration Ij, can
be composed of a set of components C = {c1, c2,… ck}, where
a component ci has a list of properties P = {p1, p2 …pk}.
Second, each component might have a description within the
text, thus a linkage between the two is required. To make things
more complex, a component cj can be formed from a subset of
components, a step that resembles a recursive property, where
the ship as a whole is formed by small parts, which in turn are
composed by smaller ones, and so on. Figure 3 depicts the
interface for capturing coordinates in images linking them to
terms from the glossary.
Conversely, components can be mapped to other
representations; a good example is a model created in 3D
rendering software such as Rhino. We have done preliminary
tests, exporting geometric data from Rhino models into XML
and linking them to both 2D slides of the model and their
corresponding occurrences in the treatises.
Although linking text and images has been extensively studied;
the context of treatises raises a series of complex issues. For
example, the text of a treatise could be segmented in a variety
of ways based on different needs; assemblage sequences,
materials used, and section of the ship being described. This in
turn raises some interesting questions, for example: could the
components being included in part of the text give a hint of
what that section is about, or what section of the ship it
describes? How could components in different treatises be
compared?
Our current collection includes digital images and transcriptions
of three of the most relevant late 16th and early 17th century
Portuguese treatises: 8 Fernando Oliveira’s Livro da Fabrica
das Naos (dated to 1580), João Baptista Lavahna’s Livro
Primeiro da Architectura Naval9 (dated between 1608 and
1615), and Manoel Fernandez’s Livro de Tracas de Carpintaria
(dated to 1616). We expect to add more manuscripts as
permissions from holders are granted.
The treatises’ dual role as historically-significant text and as
formal specification of elements of ship design affords an
opportunity to investigate the relationship of techniques
developed within the context of textual studies to applications
with physical objects and virtual 3D models. Further, the
treatises provide the foundation for our development of the
Nautical Archaeology Digital Library (<http://nadl.ta
mu.edu/>), which will center on providing resources in
support of archaeologists’ work and on dissemination of
expedition artifacts10. The combination holds promise of
extending the reach of the digital humanities.
Figure 1: The multilingual glossary interface depicting terms, translations,
and definitions in various languages. Figure 3: Treatises markup utility allows to mark certain areas of the image
and associate them with terms and categories in the glossary.
1. The Canterbury Tales Project Accessed 2006-11-12. <http:/
/www.canterburytalesproject.org/>
2. The Rossetti Archive Accessed 2006-11-12. <http://www.
iath.virginia.edu/rossetti/>
3. The Perseus Digital Library Accessed 2006-11-12. <http://
www.perseus.tufts.edu/>
4. Eduardo Urbina, Electronic Variorum Edition of the Quixote
Accessed 2006-11-12. <http://www.iath.virginia
.edu/rossetti/>
5. Gary Stringer, Digital Donne. Accessed 2006-11-12. <http:/
/www.csdl.tamu.edu/donne>
6. Enrique Mallen, The On-line Picasso Project Accessed 2006-11-12.
<http://picasso.tamu.edu/>
7. Michael of Rhodes Accessed 2006-11-12. <http://www.i
ath.virginia.edu/rossetti/>
8. Special thanks to Academia de Marinha, Lisbon Portugal for
providing the facsimiles. Accessed 2006-11-12. <http://ww
w.marinha.pt/Marinha/PT/Menu/DescobrirM
arinha/Actividade/AreaCultural/academia
/>
9. Lavahna, Joao Baptista. Livro Primeiro da Architectura Naval, c.
1610. Facsimile, transcription, and translation into English, Lisbon,
Academia de Marinha, 1996.
10. Dissemination of archaeological artifacts also has been the focus
of other significant efforts including ArchaeoML, used in OCHRE
Accessed 2006-11-12. <http://ochre.lib.uchicag
o.edu/>formerly XSTAR, and ETANA Accessed 2006-11-12.
<http://www.etana.org/>