Projecting an Architectural Perspective: Euclidean Propositions and Common Practices at Gonçalves Sena’s Workshop
The projection of architectural perspectives, from the drawing plane to the tectonic support, might be the central problem of quadratura. However, specialized treatises tend to treat it through ideal statements. Highlighting the materialization of abstract reasoning, those demonstrations serve the scientific legitimation of quadratura practice, possibly circumscribed to enlightened scientific and artistic circles, rather than the dissemination of procedures aiming to train people in the projection craftsmanship. This may even be the best-kept secret of the quadratura painters. In this context, this paper is focused on Sena’s quadratura (1754), at the church of Santarém’s Jesuit College, where under Vieira’s statements exposed in his treatise on perspective (Tractado de Prospectiva, Codex 5170, Biblioteca Nacional de Lisboa, 1716), we will disclose procedures based upon common practices of the painter’s workshop concerning the outline and projection of the image.
KeywordsPerspective Optics Quadratura painting Projective geometry Virtual architecture
Considering the projection of architectural perspectives as the central problem of quadratura practice, this article aims to unveil practical procedures hidden under theoretical statements widespread through specialized treatises. As such, the scope of quadratura conical projection is taken as practical procedures that allow for the outlining of linear structures onto curved surfaces. Although the theoretical framework is based upon ideal schemes and narratives of scientific demonstration, the common practice at the painter’s workshop is assumed here, along with the material circumstances of the quadratura projection. The recent essays by Hoffmann (2009) and Camerota (2010) have looked at the problem without the constraints imposed by idealized procedures (such as Pozzo’s Modo di far la graticola nelle volte or Danti’s narrative of the fili tesi process), advancing into practical hypotheses. On the other hand, and concerning the primary sources reported here, studies on Inácio Vieira’s treatise (1716) do not fully analyze the quadratura projection resolution, despite their value in rescuing the document, identifying its sources and organizing its main contents (Leitão and Mello 2005; Mello 2002; Raggi 2004; Trindade 2015). In the same order, Gonçalves Sena’s quadratura (1754) was brought into daylight by recent art history studies on Portuguese quadratura (Mello 2001, 2002; Raggi 2004) although they call for further developments of its projective and architectural features.
In this sense, and under the recent considerations on both objects (Cabeleira 2015), a thorough review of Vieira’s manuscript and Sena’s quadratura identifies procedures for creating the image outline and its projection. Their cross-referencing leads to theoretical hypotheses tested upon graphical models that allow us to ascertain its validity as well as to communicate and visualize the aroused assumptions.
Based upon Euclidean propositions, the theoretical enunciation by Vieira coincides with the empirical reasoning undertaken at Sena’s workshop, a consideration that allows us to detach the method from ideal schemes and advance towards a practical solution. Delimiting the exposed reasoning to this issue, unveiling the best-kept secret by quadratura painters, we aim to clarify the quadratura projection practice through a simplified process that sends us back to the very fundamentals of geometric knowledge, the Euclidean propositions.
Approaching the Problem Through Sena’s Quadratura
Although the spatial purposes of Sena’s quadratura are largely achieved, inconsistencies in the projective structure are revealed. Deflections from the fundamentals of perspective are consequent to the limited scientific speculation and artistic practice of perspective in the Portuguese context, as well as to Sena’s knowledge and methodology.
Through this a compositional matrix (Pozzo 1693: 89), is detected from which the architectural image is reset: the plan of the painted architectures was reorganized, the elevations were reconfigured, the employed constructive vocabulary was adjusted and the apparent depth was reduced. Architectural themes were rearranged in order to respond to the desired spatial image (consequent to the supporting built environment), reforming, in the same sequence, the color palette (coordinated with the polychrome of painted woods and inlaid marbles of the supporting space) and figuration (according to the commissioned iconographic program), renewing the whole image. Hence, the relevance of the work results from a modus operandi capable of shaping models according to the specific circumstances of the spatial support and the intention of the imagery. However, if compositional and perspective discrepancies are verifiable, such as deviations from the precepts of a soda architettura and from the management of depths, yet the image embodies its spatial intentions. Arrangements fall into a strictly imagery-based scope, pursuing the perceptual likelihood of the spatial effect rather than its metric truth and constructive logic constrained under perspective precepts. Even so, the painter is able to unify the representation through the engagement of the outlined structure to a single convergence point providing an apparent stability and an effective space induction. Nevertheless, despite his rudimentary mastery of perspective, how did Sena solve and execute the quadratura? How was he able to project the outlined image into the semi-cylindrical soffit of the vault? Without a clear definition of the viewpoint, regulating the outline of the depth and the image projection, how did Sena circumvent the problem of the quadratura projection?
Considering that the outcome does not result from a happy coincidence, a possible resolution and materialization of the architectural perspective is pursued through simple assumptions, both theoretical and technical, rather than the commonly accepted and disseminated procedures.
Euclides and the Hidden Key of Quadratura Projection
The influence of quadratura’s theoretical models can be seen in the Portuguese panorama via Inácio Vieira’s lessons at the Aula da Sphera of the Jesuit College of Santo Antão in Lisbon (Leitão 2008). Compiling the theory on the sciences of vision, the Tractado de Prospectiva (Vieira 1716) confronts the practical statements of Andrea Pozzo’s, Perspectiva pictorum et architectorum 1693–1700 (Pozzo 1693, 1700), with the theoretical ordination of Claude Dechales, Cursus seu mundus mathematicus 1674 (Leitão and Mello 2005; Dechales 1674), directing the discourse towards the artistic/scientific prodigy of visual deceit. It is precisely from this theoretical body that a resolution of the quadratura projection emerges, fleeing away from instituted technical models.
Illustrating the procedure with the 282nd and 283rd schemes (Fig. 4) Vieira draws the architectural composition plan coincident with the picture plane. Through delineation of the plan, vertical lines would be drawn converging into F (main convergence point), and the element’s depth would be dictated by the foreshortened image of CE, EH, HI e IJ. So, bearing in mind that the represented space is made up by a succession of homothetic figures, managed by the main convergence point, it is possible to draw the different perimeters of the illusory architecture. However, the presented theory is still limited to the image outlined in the picture plane, remaining unclear about its projection onto curved, irregular or compound surfaces.
It is then, at Das abobedas, e planos irregulares (Vieira 1716: 283) (‘from vaults and irregular planes’), that Vieira explains the image projection onto curved surfaces. Repeating almost entirely statements from Dechales (1674: 519–520), which in turn holds its origins in the practices of Dubreuil (1649: 48), Vieira supports the procedure with an abstract grid. That is, a geometrical structure that, once drawn over the prototype, will determine the vault organization (Vieira 1716: 283) in the same number of squares although deformed by the conical projection.
Although a modus operandi concerning the quadrature projection is not immediately clarified, the debate touches upon three canonical possibilities: “To prescribe this in an easy way, strings may be used… or with view and nothing else, or at night by putting a light in the viewpoint…” (Vieira 1716: 283–284). These theoretical procedures all have evident drawbacks to an effective implementation: the scaffolding installation that, ultimately, should result from a set of beams permeable to the passage of ropes, light and sight; and the quality of the light source that should be strong and stable enough to project shadows with the necessary accuracy.
Following this reasoning the conical projection onto curved surfaces is solved through recognition and materialization of a Euclidean proposition. Hence, the possibility of outlining drawn structures onto the curved surface emerges using up the free space between the scaffold and the vault.
By this simple and demystified process, the quadratura projection is outlined as approachable, feasible and accurate given its operational substance. First, the grid projection onto the curvature is performed entirely above the scaffolding. Second, curvilinear projections, consequent to the conic transformation of linear segments, are operated by the intersection of a projective plane and the curved picture surface. Third, linear projections, whose transformation is strictly relative to its metric nature, are achieved by employment of an auxiliary drawing. Through this procedure the operation renounces the spatial materialization of the eye, although obedience to a viewpoint is conceptually required.
According to this sequence, and although Vieira’s statements result from a path trodden by Dechales (1674: 520), the emancipation from the widespread models based upon employment of extended ropes or light sources from the viewpoint should be valued. An attainment of the quadratura projection rooted in Euclidean propositions beside the common resources of the painter’s workshop.
The Painter’s Workshop Practice
Returning to Sena’s quadratura, it is accepted that the painted image was previously solved in the plane, where the sources of the imagery were combined and regulated under a single point of convergence. At the same time, abstract alignments would be established in order to support the image projection onto the tectonic picture surface. But how would the painter actually transfer the draft onto the vault soffit?
Due to its compositional symmetry (in four quadrants) it is possible to solve the projection out of ¼ of the image and, after that, obtain its totality by means of card and spolvero techniques. Simultaneously, assuming the hypothesis of projection already considered, it seems tangible that the operation is freed from the constraints of a projective center where a light source was placed or ropes would be attached. This disaffiliation of the projection process from the projective center is consistent with Sena’s drafting process (whose assembly of images seems to suppress fundamental conditions of a projective space) tending towards the implementation of procedures equivalent to the Dechales/Vieira statements.
If Sena’s expertise on perspective seems rudimentary it suggests the need to verify the possibility of performing the quadratura image projection by application of pictorially common procedures. From the resources of the painter’s workshop it is possible to highlight the application of reference grids (used in the calculation and manipulation of scale besides the combination and transfer of imagery models), the employment of card and spolvero techniques (to transfer and replicate defined configurations) as well as the use of accepted drafting tools (such as lines or ropes in the guidance and control of the image layout).
The use of geometrical grids serves the delineation of architectural perspectives, regulating the preliminary draft (enlarging and reducing Pozzo’s images, composing, combining and manipulating the scale of represented architectural features according to an apparent depth) and simultaneously supporting the projection onto the vault surface (being a common procedure applied at the image transference among preparatory drawings and the picture plane, in the case of easel painting).
The employment of a quadrettatura expedient (at the root of the quadratura term) is thus a technical condition that, beyond the scope of easel painting, integrates the quadratura purposes and practice. It regulates the designed composition and its projection onto the tectonic surface, as enunciated by the specialized treatises that bail out the statement in order to communicate the geometric transformations consequent to a conic projection.
On the other hand, the application of card and spolvero techniques, intrinsic to the practice of mural painting (at which Sena had a large experience), would allow the artist to transfer the outlined data from ¼ of the vault onto the remaining surface, shortening the whole process.
As for the applied instruments (such as ropes, plumb line, ruler and compass), these allow the quadratura painter to respond to the essential problems caused by the projection of linear entities, coordinated under practical geometrical procedures, onto curved surfaces.
Starting from this inscribed grid over the perspective draft, Sena would project it onto the vault through the definition of projective planes without the necessary materialization of a projective center. Operating exclusively above the scaffolding, the painter would be assisted by a plumb line (suspended from the center of the vault) and a layout of the compositional grid (inscribed over the scaffold through extended ropes or by direct sketch). Applying the Euclidean proposition, underlined by the procedure (Vieira 1716: 286), the plumb line and a segment of the materialized grid, at the scaffold, would define the required projective plane for projection of the geometric entity onto the vault.
However, while in Vieira’s treatise the application of the 18th proposition from Euclid’s eleventh Book is based upon the definition of the projective plane through two parallel lines (taking into account the 35th proposition of Euclid’s first Book), the methodological proposal in Sena’s procedure is based upon the plane definition through two intersecting lines (the 2nd proposition from Euclid’s eleventh Book). Thus, the parallel lines of Vieira’s statement are the distance line and a vertical line of the vault suspended grid, while Sena’s intersecting lines are the distance line and a vertical line of the perspective image outlined at the scaffold plane.
After the projection of the compositional grid, the transference of the outlined image is operated under evaluation of the proportional relations between the perspective image and its referential grid, either in the supporting draft or its homologous projection onto the vault. This is a procedure that does not require a complex scaffold (according to logics of flexibility and visual permeability) once the whole operation takes place exclusively above it. On the other hand, the transference of the outlined image onto the rest of the vault intrados can be operated through card and spolvero techniques, whose procedures are certainly controlled by the painter.
The projection of the compositional grid results from the conjunction of the horizontal and vertical alignments. However, it can be pointed out that the features considered are only related to the quadratura perspective structure and its projection, leaving aside pictorial aspects highly responsible for the image’s effectiveness. Thus, besides the delineation of linear entities, the quadratura spatial effect arises from the pictorial potentialities, like the skill in material imitation, the use of color in the emphasis of depth and segregation of illusory spatial layers, the modelling of the chiaroscuro and the conformity between represented and physical light.
Quadratura Projection Between the Ideal and Real Constraints
As the problem of projection, between the plane draft and the curved picture surface, might be the key issue of quadratura, it is important to review the modus operandi followed by some authorities whose procedures may be at the origin of Vieira’s statements.
In Le due regole (Vignola and Danti 1583) the Fili tesi method solves the conical projection into the tectonic surface through a matrix of horizontal and vertical alignments of the architectural image.
According to this, the horizontal alignments are calculated through an auxiliary drawing (determining the intersections of the visual rays with the vault profile), and delineated onto the vault by beating strings regarding the non-transformation of its linear configuration. On the other hand, the projection of vertical alignments will result into a curvilinear transformation. As such, their projection is supported by extended ropes, materializing visual rays, and confirmed through the eye. For each vertical segment, both extremes are signed into the vault: one coincides with the horizontal picture plane from which the image is generated, and the other correspondent to the farthest point of the segment. Joining a rope from the first point until the viewpoint and suspending a plumb line from the second point, the method applies, implicitly, the same Euclidean proposition identified by Dechales and Vieira. Thus, the plumb line should intersect the oblique rope, evidencing the segment’s projective plane. Once the accurate projection of the second point is verified, it is possible to outline the entire vertical entity with a flexible catenary adaptable to the surface curvature.
Anyhow, given the articulation between the projective plane and the picture surface, it is possible to avoid the viewpoint as the control center of the projection. According to the essay by Di Marzio (1999), the rope connecting the lower extreme of the vertical segment with the projected viewpoint, and the vertical plumb line suspended from the projected viewpoint, are enough to define the projective plane. Hence the method chases the scientific legitimacy of the quadratura practice rather than the disclosure of a practical procedure. Looking at Danti’s statement, and taking into account Di Marzio’s reasoning, the proposed variation is not far from the method presented by Vieira.
However, if the method exposed in Le due regole (Vignola and Danti 1583) is rooted in the specificities of a compositional matrix, Le Moyen universel (Bosse 1653) sets an universal model, either theoretical or practical, regulating the quadratura projection under constant projective rules.
In Bosse’s method the image is subordinated to an abstract grid transformed by the conic projection. The same conceptual basis taken by Pozzo (1693), through the definition of three square grids, or adopted by Dechales (1674) and Vieira (1716), defining a square mesh suspended from the curved surface.
From the ascertained methods, the quadratura projection may be based upon delineation of a compositional matrix, attached to the represented bodies (Vignola/Danti), or an abstract grid, juxtaposed to the draft image (Bosse/Pozzo and Dechales/Vieira). Focusing on the last case, it should be mentioned that the positions of the grids considered have differences although they may generate matching results.
As such, while in Bosse’s procedure the baseline of the interposed grid matches with the vault width, in Vieira a correspondence of measure arises from a higher alignment of the projection surface, demanding the calculation of the grid lower limit constriction A requirement that, being forgotten by Vieira, reveals the metric awareness of Bosse’s perspectival construction, while Vieira seems to depreciate the value of measure due to a strict management of an outlined image. However, Vieira’s enunciation is reportedly based upon application of Euclid’s Proposition, the 18th of the eleventh Book, in a clear methodological coincidence with the distinct proposals disseminated throughout the specialized treatises. An enunciation that departs from dependence on the viewpoint, decreasing requirements for its execution (scaffolding installation, ropes extension and control from the viewpoint) at the same time that it responds to the required accuracy in outlining the referential drawing structures.
The model taken from Vieira’s treatise seems to converge with the recent speculations of Hoffmann (2009) and Camerota (2010). In the attempt to unveil Pozzo’s practical procedures concerning the quadratura projection, especially the control of the delineated lines onto the vault curvature, both authors refer to the use of two ropes attached to the projection of the viewpoint at the center of the vault: one to guide “uprightly the rule” in the delineation of lines, and the other, as a suspended pendulum, to prevent any variation of the established direction. This is a model based upon statements and illustrations by Emmanuel Maignan (1648: 372), given for the need to control deformations of meridian lines, but neglected by the quadratura specialized treatises, other than Dechales (1674) and Vieira (1716).
Through the consideration of the theoretical and technical components examined here, the applicability of the method is surprising, giving to the quadratura projection a resolution distant from the elaborate theoretical assumptions expressed in specialized treatises. Those assumptions arise as speculative exercises in the research and demonstration of projective rules, mastered by high artistic and scientific erudition centers. As such, the wider practice of the quadratura projection would be solved through common instruments, procedures and knowledges acquired at the painter’s workshop. An example of this parallelism between theory and practice is the case of the projection of straight lines into curved surfaces stated by Dechales and Vieira. A procedure solved through the spatial conformation of the 18th proposition of Euclid’s eleventh Book and the method that the painter would have applied to connect pairs of points—the projection of the viewpoint in the vault, and points of the linear entities coincident with the impost line—with rope and ruler in order to guide the delineation.
Sena’s quadratura at Santarém’s Jesuit College openly asserts the imagery and resources of the Jesuit mission: “The painting is a brilliant mirror of virtues, a rhetorical scourge of this world of addictions, a spirituous incentive to the moral perfection and orientation of the wise, and catholic men.” (Benedicto 1791: 14). But if the work reflects the desired virtues of its commissioners, Sena is not necessarily a virtuoso on perspective. Even so, this humble painter is able to solve the quadratura projection through the application of Euclidean propositions and pictorially common procedures.
Gonçalves Sena (1713/1790) was a self-taught spirit whose apprenticeship and artistic update arises from the study of coeval treatises (Benedicto 1791: 8), necessary to the scientific and iconographic codification of the pictorial practice. He started as an easel painter, developing, progressively, capacities as a mural painter and finally as a quadraturist.
If a straight line be at right angles to any plane, all the planes through it will also be at right angles to the same plane (Euclid 1956, vol. III: 302).
- Benedicto, Joaquim Duarte. 1791. Elogio do grande apelles portuguez, Luiz Gonçalves de Senna. Lisboa: Officina patriarcal de Francisco Luiz Ameno.Google Scholar
- Bosse, Abraham. 1653. Moyen universel de pratiquer la perspective sur les tableaux, ou surfaces irrégulières ensemble quelques particularitez concernant cet art et celuy de la graveure en taille douce. Paris: Bosse.Google Scholar
- Cabeleira, João. 2015. Arquitecturas imaginárias. Espaço Real e Ilusório no Barroco português. Ph.D. thesis, Universidade do Minho.Google Scholar
- Camerota, Filippo. 2010. Il teatro delle idee: prospettiva e scienze matematiche nel seicento. In: Mirabili Disingani, eds. Richard Bösel and Lydia Salviucci Insolera, 25–36. Roma: Istituto Nazionale per la Grafica, Artemide.Google Scholar
- Dechales, Claude. 1674. Cursus seu mundus mathematicus. Lyon: Officina Anissoniana.Google Scholar
- Di Marzio, Daniele. 1999. La Sala Clementina in Vaticano. Procedimento per la costruzione diretta della prospettiva su superficie curve: ipotesi teorica e verifica sperimentale. In La Costruzione dell’Architettura Illusoria, ed. Riccardo Migliari. 152–177. Roma: Gangemi Editore.Google Scholar
- Dubreuil, Jean. 1649. La Perspective pratique. Paris: Melchior Tavernier.Google Scholar
- Euclid. 1956. The Thirteen Books of the Elements. 3 vols. Trans. Thomas Heath. New York: Dover.Google Scholar
- Hoffmann, Volker. 2009. Matematica e mistificazione: Andrea Pozzo spiega la costruzionedel suo affresco in Sant’Ignazio a Roma. In L’Arte della matematica nella prospettiva. Atti del Convegno Internazionale di Studi.Roma, Ist. Svizzero, 9 ottobre 2006. Urbino, 10–11 ottobre 2006. 151–160. Perugia: Cartei and Bianchi Edizioni.Google Scholar
- Leitão, Henrique. 2008. Sphaera mundi: a ciência na aula da esfera: manuscritos científicos do Colégio de Santo Antão nas colecções da BNP. Lisboa: Biblioteca Nacional.Google Scholar
- Leitão, Henrique and Magno Mello. 2005. A pintura barroca e a cultura matemática dos jesuítas: o Tractado de prospectiva de Inácio Vieira, S.J. (1715). Revista de História da Arte 1: 94–142.Google Scholar
- Maignan, Emanuel. 1648. Prospetiva horária. Roma: Philippi Rubei.Google Scholar
- Mello, Magno. 2001. Os tectos pintados em Santarém durante a fase barroca. Santarém: Câmara Municipal.Google Scholar
- Mello, Magno. 2002. Perspectiva pictorum: as arquitecturas ilusórias nos tectos pintados em Portugal no século XVIII. Ph.D. thesis, Universidade Nova de Lisboa.Google Scholar
- Pozzo, Andrea, 1693. Perspectiva pictorum, et architectorum. Tomo I. Roma: Giacomo Komarek Boemo.Google Scholar
- Pozzo, Andrea, 1700. Perspectiva pictorum, et architectorum. Tomo II. Roma: Giacomo Komarek Boemo.Google Scholar
- Raggi, Giuseppina. 2004. Arquitecturas do engano: a longa construção da ilusão. Ph.D. thesis, Universidade de Lisboa.Google Scholar
- Trindade, António Oriol. 2015. A Pintura Integrada em Tecto e Abóbadas e a Perspectiva Linear. Lisboa: Faculdade de Belas-Artes da Universidade de Lisboa.Google Scholar
- Vieira, Inácio. 1716. Tractado de Prospectiva. Codex 5170. Biblioteca Nacional de Lisboa.Google Scholar
- Vignola, Jacopo and Ignazio Danti. 1583. Le Due Regole della Prospettiva Pratica. Roma: Francesco Zanetti.Google Scholar