Adding facades to Palladio’s domestic plans is less difficult than creating the plans themselves. Daniele Barbaro’s maxim “Every three-dimensional form is born from its plan as a tree is bom from its roots” embodies the process we use to recreate Palladio’s facades.
Although Planmaker and Facademaker are computer programs, we can imagine them as a pair of architects interacting as they design a Palladian villa. Decisions as to when the exigencies of the facade must override those of the plan, and vice versa, are arrived at by mutual agreement. As much as possible, our own priorities mirror Palladio’s. Although he eliminated most conflicts between plan and facade by working the two into a unified whole, traces of compromise remain. Our charge to achieve “all possible Palladian villas” means that for any one plan we will design not one but a multitude of facades. By doing so, interestingly, we encounter conflicts rarely found in Palladio’s own designs.
Reflecting their different responsibilities, Planmaker and Facademaker employ opposing methods of design. Planmaker divides and redivides a given rectangle, while Facademaker adds one element to another. The former speaks the language of splits, which as we have seen is a flexible one indeed. In contrast, Facademaker uses a small set of what we call “blocks.” A few simple rules guide it in assembling these blocks. Unlike Planmaker, which operates on a blank slate and can produce millions of possibilites, Facademaker is guided and constrained by a preexisting plan and is capable of far fewer variations. As in the last chapter, we have used Book II of the Quattro Libri as our primary source. We have, however, also borrowed extensively from Palladio’s drawings and built structures.
First of all, then, we must de- and reconstruct the facades in question by viewing them as assemblages, or stacks, of blocks. After identifying all the types of blocks, we will define their possible combinations. Both the types and the combinations are, of course, derived from Palladio. Depending on the particular combination, Facademaker will adjust each block to the plan and then insert it into the facade.1 We begin by dividing the facade into two sections, base and body. These are illustrated in figure 3.1 for the Villa Sarego at Miega di Cologna (Quattro Libri, 2.68). Bases are complex subfacades in their own right, and so we will examine them later. Figure 3.2 illustrates assorted combinations of the four types of body blocks: floor (as in “first floor” or “second floor”), entablature, attic or mezzanine, and roof. A facade need not contain all these block types, and may contain multiple floors or entablatures (fig. 3.2a). The minimal body (fig. 3.2c) consists only of a floor and a roof. In short, our facades simply recombine these four types of blocks.
The essential component of all facades is the floor block. Always located directly above the base, and either one or two stories high, the floor block is the facade’s largest, most detailed, and most prominent block. Figure 3.3 illustrates the five different floor blocks found among Palladio’s villa facades. We call them: (a) column, (b) column with pedestal or dado, (c) colossal column (more than one story high), (d) arch, and (e) center composition.
Palladio’s workhorse, the column block (i.e., either a temple front or a colonnade), forms the first story of his many temple-fronted villas and palazzos. Its close relative (used for the upper story of his temple fronts) is the column with dado block. In it, a dado rests beneath correspondingly shortened columns. The colossal column block is two stories tall, its columns reaching from base to entablature without an intervening cornice or dado. The arch block substitutes arches for columns and is, like the column block, one story high. The center composition block includes all Palladio’s villa facades that are neither columnar nor arcuated; it comes in several varieties (fig. 3.4).2
In figure 3.5a we illustrate examples of the two kinds of entablature blocks (Quattro Libri, 2.68). The molded stringcourses across the facade distinguish the full entablature (below) from the half entablature (above). These terms are only conveniences, of course; the actual difference is that the full entablature includes an architrave while the half entablature has none. Unlike the full entablature, the half entablature may house windows, as in the Villa Badoer (fig. 3.5b).
Attic blocks are generally one-third the height of floor blocks, and they appear as second or mezzanine stories. The plain attic block on the Villa Fomi-Cerato (fig. 3.6a) houses only windows. The Villa Rotonda (fig. 3.6b) has an attic block containing windows and also the pediment of the temple front. Note that the pediment is the full height of the attic block. In addition, some attic blocks in Palladio’s facades have no windows in their center section, as in the Villa Saraceno (fig. 3.6d).
Roof blocks are flexible in size and shape. Two characteristics define a roof, pitch and height, as illustrated in figure 3.7. Most roof pitches in the Quattro Libri range from 15° to 35°, so these angles are our minimum and maximum pitches. But Palladio also gives verbal instructions for roof heights. In Book I he writes that the ideal roof height is the width of the facade. However, only about half of the Quattro Libri roofs follow this rule. The other half vary, as do our roof heights, between and ¼ the width of the facade. Nearly all Quattro Libri roofs fall in this range.
Even within these limits, however, the two roof characteristics are not independent variables. The pitch in figure 3.8a is within bounds, but the height exceeds the maximum of ¼ of the building’s width. A roof this steeply pitched would have to be hipped, as in figure 3.8b. A more moderately pitched roof could be either pyramidal (fig. 3.8c) or hipped (fig. 3.8d), depending on one’s choice of roof height. By varying pitch and height within these Palladio-derived limits, we can create a multitude of canonical roofs for any facade.
In figure 3.9 we illustrate the two kinds of roof block, with and without pediment, respectively on the villas Mocenigo at Marocco and Godi. The width of the pediment in figure 3.9b, as we shall see, is fixed by the plan. For the most part Palladio makes the pitch of the pediment equal that of the roof. But while roof pitches vary from 15° to 35°, pediment pitches vary only between 25° and 30°. So if the roof is steeper than 30° or flatter than 25°, we pitch the pediment at a different angle from the roof to keep it within range. These adjustments are shown in figure 3.10.
Finally, the main body of a villa facade must rest on a base. Palladio’s base designs are numerous and are related to the villa’s site. For example the Villa Foscari (La Malcontenta), sited on a plain adjacent to the river Brenta, sits on an extremely high 11-foot platform to prevent the piano nobile from flooding. Other bases are relatively low, and of course urban palazzi typically have no bases at all. To allow for this diversity we need a flexible block—a “superblock,” so to speak. If our block description is to be reasonably terse, however, we have to give up the idea of replicating all of Palladio’s bases. We will consider only what we call straight-run bases, and eliminate those bases with steps that are both perpendicular and parallel to the facade. Our aim, which has been to create the rules for all possible Palladian villas, must here be slightly deflected.
Our standard unit of base construction is a set of steps, or a level, defined by a stringcourse. As illustrated in figure 3.11 in four facades from the Quattro Libri, bases may consist of from one to four levels. Table 3.1 defines the number of steps each level may contain. The allowable numbers of steps vary with the number of levels in the base as a whole. Although they do not encompass all of Palladio’s bases, our step parameters do closely reflect most of his step groupings. In addition, the second base level, if tall enough, may contain windows, as in figure 3.11c.
NUMBER OF STEPS IN LEVEL
NUMBER OF LEVELS
Like levels, individual steps come in assorted sizes and styles. Accordingly, we define four step characteristics: height, configuration, termination, and width. As illustrated in figure 3.12, each of these characteristics may assume one of two guises: single or double height, straight or splayed configuration, ended or notended termination, and broad or narrow width. (We call “step ends” what are more technically dwarf walls—low blocks protruding at right angles from the facade that provide parapets for the stairs.) Note that the combination of guises affects every step, regardless of the number of levels or number of steps per level.
The different combinations of these guises define different base styles. For example, the style in figure 3.11c is single/straight/ended/broad. As shown in figure 3.13, the combination of broad, splayed, and ended is nonsensical because the steps would spread past their terminating ends. For this reason we disallow the combinations single/splayed/ended/broad and double/splayed/ended/broad. Four characteristics, each able to assume one of two guises, minus two combinations, result in fourteen different base styles. We illustrate all of them in figure 3.14 using three-level bases.
We now possess a full set of blocks. As they stack together in an infinite number of ways, we need to segregate the combinations, or styles, that rate as Palladian. If our definition of “Palladian” is too inclusive then the term loses its meaning; if too exclusive, it defines only part of what Palladio thought possible. So we have to sort through all the possible combinations and accept or refuse them based on Palladio’s built, published, and drawn works.
To begin with, common sense excludes many combinations (such as two entablatures stacked one atop the other). Convention dictates many other restrictions. For example, all facades contain one roof and one base block. All arcuated or columnar facades contain exactly one pediment, located in either a roof or an attic block. Arch and column blocks always support an entablature, but center composition blocks never do. Correspondingly, entablatures rest only on arch or column blocks. And we use attic blocks for second stories only, never for ground floors.
Some classically acceptable combinations are not found in Palladio’s oeuvre, however, and this poses a dilemma. For one, Palladio pairs a symmetrical plan with a three-story facade only once in the Quattro Libri (and the design is for a palazzo, not a villa). Yet we can build myriad three-story villas from our blocks, and in doing so fabricate a sizable class of facades not grounded in Palladio’s own work. We do not wish to form the majority of our facades from extrapolations, and so we limit all combinations to two stories or less. Also, the half-entablature block, which may contain windows, must lie underneath the roof and not between stories. Of all Palladio’s many one-and-a-half- and two-story facades, not one sandwiches a half entablature between floors. As illustrated in figure 3.15, the entablature windows would disjoin the first and second floors and clutter the facade. We can see why Palladio avoided this particular device.
Lastly, although some of Palladio’s arcuated temple fronts are one and a half stories, none are two full stories. From our blocks, however, we can build three such facades, as illustrated in figure 3.16. The absence of facades like that in figure 3.16a from Palladio’s entire domestic oeuvre is conspicuous because his basilica at Vicenza is similarly two stories and fully arcuated. He apparently considered this arrangement appropriate for grand civic structures and inappropriate for merely domestic ones. Therefore, we exclude it.
We also have purely stylistic reasons for striking the combinations in figures 3.16b and c. Although we followed Palladian proportions (which we will examine later) exactingly in drawing these facades, in both of them the first and second stories differ substantially in height. Having both stories of equal height would not work because the taller arches would be spindly, while shorter columns would violate the rules governing the proportions of the classical orders. Additionally, in figure 3.16c a visually heavy floor teeters on slender supports, and the facade appears to have distended base-to-roof arches.
Palladio’s facades do not correlate either the height or the shape of the roof with a particular body style. Similarly, no conventions govern base height or style. So we do not link the design of these two malleable blocks (base or roof) to any specific combinations.
Sixteen block combinations now remain. The five styles of floor block—columnar (column, column with dado, and colossal column), arcuated, and compositional—define the three major Palladian facade styles. The combinations we will build on each of these constitute only the Palladian variations on these major themes; other reasonable but un-Palladian variations remain. Some combinations differ only slightly from others—substitution of a half entablature block for a full entablature block, for example—but each combination accommodates a wide variation of roof, base, door, and windows. Figure 3.17 illustrates an example of each combination as constructed by Facademaker using assorted permutations of doors, windows, roofs, and bases.
But we have yet to look at the connection between facade and plan that we cited earlier. As described so far, blocks are static objects, save for certain tightly defined variations, and function only as vertical components. To begin forging the plan-facade link, we contrast two villas. The Villa Emo (Quattro Libri, 2.55) and the Villa Pisani (Quattro Libri, 2.52), illustrated in figure 3.18, are nearly mirror images in plan: nine-square in extent, with the center square divided into three spaces. Excluding wall widths, Emo measures 60 × 55 Vicentine feet, Pisani 59 × 55.
As illustrated by these examples, the facade has two distinct aspects. Its vertical aspect, or block combination in our language, is not linked to the plan, but its horizontal aspect most definitely is. Blocks are not static objects, for in width and latitudinal arrangement they are extremely fluid, and they can adjust in height as well. In order for facade to spring from plan, we must establish lines of communication between Facademaker and Planmaker—what things about the plan does Facademaker need to know? Or, paraphrasing Barbaro, precisely how is a tree bom from its roots?
First, we define the terms “center zone” and “outer zone.” As demonstrated on the Villa Ragona (fig. 3.19; Quattro Libri, 2.57), one can divide Palladio’s facades into a center zone and two outer zones that are mirror reflections of each other. The center zone houses most facade details, such as columns, doors, arches, and pediments. Outer zones house only windows.
Palladio derives the center zone’s width from the plan. Most often, he matches it to the width of the forward central room, as in the Palazzo Antonini (fig. 3.20a; Quattro Libri, 2.5). If that center zone proves too narrow for a minimum of four columns and three intercolumniations, then he expands it to equal the width of the widest room on the center axis, as in the Villa Comaro (fig. 3.20b; Quattro Libri, 2.53). If this room is also too narrow, or should there be no wider room on the center axis, then he aligns the center zone with the three front center rooms, as in the Villa Angarano (fig. 3.20c; Quattro Libri, 2.63). The outer zones are simply those portions of the facade outside the center zone; in figure 3.20c there are no outer zones at all because the center zone is as wide as the entire plan.
Opposite design philosophies govern the two types of zone. Palladio determines placement of the outer zone’s only detail, the window, by the plan. In figure 3.20, all the outer zone windows align with door axes in plan. But in the center zone (the land of abundant detail), facade overrules plan. There, window location is a byproduct of column placement, which is in turn calculated by Palladian formulas. In this zone, ideal column spacing is everything; alignment of windows with the plan is nothing.
There is, however, a practical exception to the facade’s hegemony in center zones. In the Villa Foscari (fig. 3.21; Quattro Libri, 2.50), the two windows that would normally flank the door are missing. Why? The center zone width and resulting column placement left those two windows squarely in front of two walls. Windows that open onto walls are an architectural absurdity, so Palladio simply erased such windows; note that he did not sacrifice or rearrange columns to solve this problem. There are thus traces of facade-plan conflict here.
Planmaker, then, passes to Facademaker the four pieces of information illustrated in the former’s nearly completed plan in figure 3.22 (it lacks only windows). First is the plan’s width; after all, plan and facade width must match. The second piece of data, center zone width, is calculated by Planmaker using Palladio’s method illustrated in figure 3.20. (It need not pass outer zone width to Facademaker because this simply equals [plan width – center zone width] / 2.) Third is a list of outer zone window locations that Planmaker compiles from the vertical door axes running through outer zone rooms on the facade edge of the plan. Fourth and last are the wall locations within the center zone where Facademaker may not place windows.
In addition, we link the scale of plan and facade. As with Palladio, our unit of measurement throughout the facade is the column width. Following his usage in the Quattro Libri, our scale equates one wall width in plan to one column width in facade.
After receiving Planmaker’s four nuggets of information and selecting one of the sixteen block combinations, Facademaker goes to work. It begins with body blocks, moving from bottom to top, and adds the base last. One of the five floor blocks—column, column with dado, arch, colossal column, or center composition—is laid out first. For all floor blocks except center composition blocks, Facademaker determines first how many columns, colossal columns, or arch piers it can draw, then the height of each, and finally the dimensions of the windows. Only after making these calculations does it draw out the block.
For column and colossal column blocks, Facademaker bases its calculations on the proportions of Palladio’s classical orders. He illustrates five orders in Book I—Tuscan, Doric, Ionic, Corinthian, and Composite—and for each enumerates the proper ratios for column height to column width and intercolumniation to column width. Facademaker cannot reproduce the orders, as our computer draws detail far less ably than a Renaissance engraver, but we can reflect the proportions of the various orders by fixing column width and varying column height and intercolumniation.
Colossal columns, which are twice the height of standard columns, and piers, which are the supporters of relatively heavy arches, ought to be wider than standard columns; we fix their widths at li column widths. We space columns more generously than Palladio to compensate for Facademaker’s thick lines, but we do adhere to the classical principle of having columns grow taller and proportionally thinner as the intercolumniation decreases. Our ratios for column and colossal column blocks are shown numerically and graphically in figure 3.23.
Although Palladio offers no written instructions regarding arch proportions, his built instructions are relatively clear. There are five arcuated Palladian villas—Caldogno, Gazzotti, Pisani-Bagnolo, Saraceno, and Zeno—although only Saraceno is published as an arcuated facade in the Quattro Libri. The ratio of intercolumniation to pier width in these five varies between 12:1 and 2 2:1, but the proportion of the “springing box” in all is about 2:1. In other words, as the arch widens in span, the piers must grow taller to maintain the constant 2:1 ratio. As shown in figure 3.24, the relationship between intercolumniation and pier height is direct, and not inverse as with columns and colossal columns.
Depending on the type of floor block, Facademaker calculates the number of columns, colossal columns, or piers, depending on the block, that can fit within the center zone. The intercolumniation must be somewhere between two and four widths for columns and colossal columns, or between 1½ and 2½ widths for piers. From the resulting intercolumniation, Facademaker calculates the proper height of the supports using the ratios from figures 3.23 and 3.24.
Next, Facademaker calculates window dimensions. Windows fall between columns. So, as intercolumniations narrow, window widths must do the same. To avoid a crush of numbers we will not list all the ratios of former to latter, but in general windows are roughly half as wide as intercolumniations. As the intercolumniation increases, however, this proportion diminishes. This is to prevent excessively wide windows. Palladio instructs us to make windows 2⅙ as high as wide, and he follows this prescription closely in all his facades. So, as we adjust window width to a given intercolumniation, we also adjust window height to maintain this ratio. We illustrate progressive examples for column, colossal column, and arch blocks in figure 3.25.
In contrast to other floor blocks, center composition blocks do not adjust to center zone width. We store the different entrances from Palladio’s center composition facades as fixed drawings. When laying out the block, Facademaker copies the chosen entrance from its library of drawings onto the facade.
In laying out floor blocks, then, we generate the facade from the plan. The width of the portico equals the width of the center zone as passed on from Planmaker. Indirectly, via Facademaker ratios, the center zone also determines the number and height of columns or arches, and the number and dimensions of windows. As to placement, Facademaker rules in the center zone, save for the matter of window-wall conflicts. But Planmaker rules in the outer zones. In sum, the horizontal as well as the vertical aspects of the facade do indeed spring from the plan.
We continue up the facade to the entablature. In his intended insert to Book I of the Quattro Libri, which did not appear at the time it was written but has recently been published (cf. our appendix), Palladio states that the entablature should be two column-widths high. In practice, however, entablatures in the Quattro Libri are closer to one and two-thirds column-widths. We follow practice instead of theory and fix the height of both half- and full-entablature blocks at one and two-thirds column-widths, as shown in figure 3.26. Both are slightly wider than the plan, so as just to overhang the facade body proper. Their center sections (or, in three dimensions, those portions protruding from the facade), are slightly narrower than the center zone and just short of the edges of their supporting capitals.
Although it is a type of floor block, we use the column with dado only as a second story in combination with a first-story column block. In locating second-story windows, we follow Palladio’s instructions (which he states twice) to place the openings of doors and windows exactly over one another so that void may be over void and solid over solid. Similarly, Palladio tells us to place the upper columns directly above those underneath them (Quattro Libri, 1.6). Palladio’s instructions regarding the height of the second story (see appendix) are also unambiguous: “the lower will have to be one-fifth higher than the upper one,” to account for the effect of foreshortening. And, in fact, the first-story columns in the Quattro Libri are consistently 5 the height of the second-story columns; practice does follow theory in this case. Nonetheless, if we include the dado when measuring the height of the second story, we find that the two stories are nearly the same height in all facades. So Facademaker shortens the second-story columns and inserts the dado, as illustrated in figure 3.27.
Palladio also tells us to use a lighter order on the second story than on the first, and he does so without fail both in the Quattro Libri and in his built villas. If the first story is Doric then the second story is Ionic; if the first is Ionic then the second is Corinthian. Since Facademaker does not distinguish between orders, we will simply have to imagine these differences.
The height of an attic block is one-third the height of its supporting floor block. In the attic with pediment, illustrated in figure 3.28a, pediment width equals center zone width, and pediment height equals that of the attic. The lines of the attic with center lines, shown in figure 3.28b, also define the edges of the center zone. As with column with dado blocks, all windows in attic blocks align with the windows of the first story, but adjust their height so as to remain square.
The facade capstone is the roof block. Both those with and those without pediment contain a narrow cornice that, like an entablature, is slightly wider than the facade body (fig. 3.29). Like attic pediments, roof pediments are as wide as the center zone. As related earlier, roof heights and shapes and roof pediment heights vary within limits based on the width of the facade.
Our system of facade construction is now complete, but Facademaker still has one remaining duty. The last element in the previous chapter, the windows, was not incorporated into the plan at the time because Planmaker knew neither the width nor the locations of the windows on the plan’s facade edge. Both these decisions, we said, were to be up to Facademaker. Having now made these decisions, Facademaker can pass this information back to Planmaker.
So much for the villa’s main facade. The other facades are easy. Although Facademaker draws neither rear nor side elevations, we, like Palladio, locate their windows according to the plan. In plan, Palladio’s rear and side windows are the same width as the facade windows, and so Planmaker uses the same width for all windows.
Figure 3.30 illustrates how Planmaker cuts windows. On the facade edge of the plan, at each location indicated by Facademaker, Planmaker cuts a notch into the wall; the notch’s width is equal to the window width relayed by Facademaker. On the plan’s left and right edges it cuts the same notch at the end of each room’s horizontal door axis. It does the same in each room on the top edge, cutting the notch at the end of the vertical door axis. As is appropriate to its humbler status, the rear facade resulting from this method is simpler than the front facade.
We have come full circle. Planmaker spoke to Facademaker and Facademaker spoke back. The fruit of their labor, a facade-plan combination such as could appear in the Quattro Libri, is illustrated in figure 3.31. We have drawn the rear and side facades manually to illustrate that one can build them using the same blocks from the front facade. Palladio’s side facades are uniformly quite plain: he simply continues the cornices and stringcourses around from the front, and locates the windows according to plan. His rear facades, like that in figure 3.31, are more richly designed and often constitute a play on the main facade.3
Is our villa Palladian? We evaluate our method in the next chapter.