Chapter II

Formal Structure

PRELUDE

I am sitting in a coffee shop in a university town. Through the window and across a narrow street I can see one of the facades of an old brick trolley-barn, several stories high. It has been many years since the electric trolleys came here at the end of their run, inched their ways up the central ramp to the storage bays on the upper floors, and passed the night. Recently, the building has been transformed into a multi-level shopping mall known as “The Garage.”

In its external appearance, the facade of the building resembles an ancient Roman apartment block such as those known from the old port city of Ostia two millenia ago. There are resonances also, for the historically-minded, with Italian Renaissance palazzi, with several tiers of large, arched windows, articulated roof-mouldings, and the like.

This is most extraordinary.

Indeed, it is remarkably puzzling and delightful at the same time. Here is an object masquerading as an ancient block of flats, designed to serve as a barn for electric trolleys and now serving as a bazaar dispensing everything from cheesecake to “classic” science-fiction comic-books.

How can this be? What is being “communicated” (if anything) by this bizarre hybrid of forms and usages? What were the intentions of the structure’s designer(s)—and how many ways are its messages variously read by users and passers-by in the course of a day? Are the connotations of this object’s appearance variable from year to year, and across generations of students? What is its future? Indeed, is it at all possible to capture the intended and effective meanings of this remarkable object in any clear and coherent way, even at a random moment in time?

Not that the very place I am sitting in is any less remarkable. It is a large cubicle, a space-cell in a large building compound erected over the last decade, containing a large variety of spacecells, housing a bewildering variety of functions, some directly related to the administrative business of the university, others entirely independent.

The cubicle to the left is a hair salon. It is otherwise of the same general configuration and external facade-setting as my cafe: same glass doors, same brick flooring, same painted brick walls within a single bay of the poured-concrete structural frame. To the right is a shop selling overpriced clothing to unknowing Saturday shoppers, then a shoe store, and finally an art-supply shop.

In what might be fantasized as a relatively reasonable universe, would we expect that each of these “functions” might elicit different material or formal settings? And if so, why? To be sure, the infrastructural organization of these otherwise identical cubicles varies from shop to shop, “reflecting” in some (presently) conventional sense their variant purposes and internal routines. But why this or that infrastructure? If we move about more widely in the city, we will notice that there will be other stores selling otherwise identical wares amidst strikingly different settings.

It is by no means a trivial question to ask how and why this extraordinary state of affairs has come about. We take this situation so for granted that such a question hardly surfaces at all. Perhaps this is as it should be. Often, such a question emerges only when architectural rhetoricians claim some “innate” connection between the formal structures they are proposing and its intended programme or usage: Ledoux’s hall for the sexual initiation of adolescents whose plan resembles a phallos,² or the countless hot-dog stands shaped like hot-dogs. Is our shopping mall masquerading as a Roman apartment building simply a remarkably subtle and sophisticated transformation of the aforementioned (apparently) simple one-to-one mappings of “form and function?” The mind boggles at the task of having to reconstruct the latter out of the former.

Let us put these issues at rest for a moment and consider again the large building compound in which my cafe is situated. There are activities going on in all of its component space-cells which are unknown or unknowable to the users of any given cell. These activities are taking place “simultaneously,” unknown to each other. 2:10 PM in space-cell M will have different contextual connotations from 2:10 PM in space-cell N.

We are situated in the midst of a bewilderingly complex space-time framework for human action and interaction. This framework, moreover, is no simple passive stage-set for activity. Its component parts actively engage our perceptions, challenging us at every turn, offering us possibilities for and constraints upon action and experience. A given component may have certain connotations for me when experienced from one direction rather than another. Spaces are perceived in subtly different ways at different times.

We are continually and relentlessly presented with experiential choices in the built environment, and yet we would appear to respond in ways which can only be partly “determined.” In some bizarre fashion, we have built ourselves into a vast artifactual web, something which is both beehive and language at the same time, and yet somehow not quite either.

Let us look more closely at the environment I’ve been describing. The window of my cafe is set into a concrete pillar-bay or frame. The pillar to the left is painted blue, while that on the right is painted red. The window sill itself is painted white. In the clothing store to the right, all three components are painted white. In the shoe store beyond, the two pillars are painted red, and the sill off-white.

I am fully aware of these differences only some of the time, principally when navigating the street at a distance: the blue pillar which I see down the block serves to orient my groggy steps in the early morning when I come for my daily caffein dosage. I cross the street at an angle appropriate to depositing me at the cafe in a straight (diagonal) line.

So, it would appear that, for me at least, this coloring serves as a navigational guide—although only insofar as I myself have made such an association and duly employ it as such. I have not enquired of my fellow patrons what navigational aids they employ, which I assume to be anything from something equivalent to the above, to the aroma of warm croissants which can (sometimes) be sensed over the automobile exhaust fumes.

And yet I am certainly inconsistent in my usage of this color coding to tell me “where” I am: as I consider the question, I become aware that I employ an enormous number of visual, tactile and olfactory cues to guide my way about the town. I “know where I am” in the city by employing anything from certain consistent odors to minute changes in the patterns of pavement, not to speak of written signs. It would appear that almost anything palpable (street J is always windy in Winter) may be so employed. To make up a taxonomy of such cues might be as useful—or as useless—as classifying all the oval objects in the universe.

My behavior is apparently inconsistent in this regard: I employ the “same” objects in different ways (but are they then the same?), and differently at different times. I employ a wide variety of cues even in the same street to help me navigate myself to the same place (is it then the “same” place?).

Moreover, what do I know of the “intentions” of those who may have structured these large or small components of the built environment? I am not so presumptuous as to assume that they were all put there for my benefit, and yet they do function on my behalf, even if they function differently for others. They are broadcast widely for all who would employ them in whatever ways they might conceivably be employed. This is not to say, however, that portions of this transmission are not privileged with regard to various receivers.

Let us consider the colored window frame of my coffeeshop. The red-white-blue pattern of the frame (for me at least) suggests associations with France, and I can dimly perceive a long line of associations—i.e., verbally linear, but somehow all visually simultaneous—with Parisian cafes, French cuisine, spirited conversation and/or silent people-watching, some indefinable “ambience.”

It comes as a slight jolt to recall the name of the cafe: “Piroschka,” recalling in fact Vienna and a legend about a little Hungarian girl. The proprietors are German-speaking, and inherited the place from a time when it was called (more appropriately?) “C’est si bon.” Perhaps, then, the red, white and blue window frame is left over from the days of Gallic hegemony. But why is it retained and what purpose does the color-coding serve? Why not Austrian or Hungarian colors? The mind reels; maybe little Piroschka wore a red, white and blue dress (or was it blue, red and white?). Maybe the colors are intended to stand for the verbal message “Attention! French cuisine inside. Enter with haste (or caution).”

The more one presses the object for a coherent or at least relatively homogeneous message, the further away an answer seems: even the “French coffee” listed on the blackboard menu is “Medaglia D’Oro” Italian coffee. Bear in mind also, that the complex web of associations which I have begun to spin is my own (apparently) idiosyncratic web: the person at the next table will undoubtedly have her own web to spin. And both of ours may be widely divergent from the “intentions” of the owners, assuming theirs were homogeneous in any way in the first place. After all, they didn’t build the place, they merely infilled an existing cubicle, and not all of that either, having inherited some furniture, appliances, and grocery purveyors: if memory serves, these are the same old croissants as in the old days.

It may be wise to put a halt to the peeling of this fruit before we reach what will undoubtedly be its hollow core, and attempt to regain some of our initial delight in the extraordinary complexity of the built environment. It may be useful to press the issue of our tricolor window frame in a somewhat more circumspect fashion.

We have noted that this color-coding exists in a wider context of window frames in the same building compound, each of which manifests certain contrasts with the others: ours is red, white and blue, that next door is all white, the one beyond is red, white and red, and so forth. We have a palpable, perceptible distinction in the coloring of the three window frames.

We will observe further that our tricolor window frame stands in a space-cell otherwise painted white (over brick). The neighboring space cells are composed of brick walls, one painted cream, the other white.

We would seem to have arrived at an obvious yet nonetheless remarkable observation: window frames (or space-cells, for that matter) need not be colored any one way. We can imagine (or observe) thousands of space-cell components around the city painted any which way, for no reason than anyone can figure out, except that in some cases a given color or color-combination may have certain conventional or symbolic associations: the crimson red so commonly found around the university is deliberately employed to make an association with the university and its name. But this association is completely arbitrary and conventional; moreover, there are objects painted crimson red in the city which are not thought of as associated with the university (even though some persons may conjure up an imaginary association if they so wish).

So, if objects need not be colored any one way, and may be colored variously, some of these colorings may be “symbolic” some of the time. And, equally, some will not be; some will be “meaningless” insofar as their presence serves merely to contrast their objects (or space-cells, or structures) from others. The established contrasts, moreover, may be connotative for any individual in a variety of ways (navigational or locational or directional aids, etc.), some or none of which may have been directly intended by those who painted or designed the objects.

We are left with an apparent paradox. It would seem that at least some of the component objects in the built environment can be both “meaningful” and “meaningless” at the same time. Or perhaps they can be “meaningful” in different ways (symbolic or merely contrastive). If we look more closely at the nature of this “contrastive” function, we would be hard put to assign any value to the context of the contrastive object (in this case coloring) apart from mere “otherness.” The red in this object may be there solely to contrast with the adjacent green; if the adjacent object were red, the former might just as well be green. Or blue, yellow or white.

Hence, we might affirm that at least some portion of the built environment has at least some of its function that of creating perceptible or palpable distinctions or contrasts (for whatever reason distinctions may be required). The built world is addressed to human perceptual mechanisms which themselves function through the apprehension of contrast (edges, changes of color-region, and other discontinuities).

Indeed, it would appear to be fundamental to the nature of the built environment to create disjunctions, contrasts and discontinuities: every object, every space-cell, every locus of action and interaction, may be distinguished through its boundedness, separation or enclosure (by whatever physical or material means).

Consider that there are (or were, as the case may be) trolley barns in the city which do not masquerade as Roman blocks of flats. Consider also that non-trolley barns may also masquerade as the same. In some cases, even blocks of flats may masquerade as (Roman) blocks of flats.

Consider that two otherwise identical space-cells may be painted different colors. Consider also two identical space-cells painted the same colors, the first having rough-textured walls, the second smooth. Consider two space cells of identical geometric configuration (cubes), one 10 meters on a side, the other 12 meters on a side. Or consider again two space-cells of identical configuration, size, color, and texture. One is a barber shop, one a chapel.

Such examples highlight the fact that the built environment is a highly complex system organized in hierarchical fashion, with contrastive levels of organization, each level apparently in some sense semi-autonomous of the others, yet connected to one another in at least partly predictable ways, given certain conventions held in common at a given place and time.

In attempting to understand the organization of an architectonic formation, it will be important to be clear about the position or perspective of analysis. In the previous paragraphs we have suggested a number of perspectives on formations. It is evidently the case that the architectonic system, like any system of signs, manifests a property of conditional elementarity. That is to say, in specifying what the component units of a formation are, that specification will be a function of the level of organization being addressed.

ARCHITECTONIC ANALYSIS

The principal problem in establishing the nature of architectonic units is the specification of what is invariant across the often bewilderingly complex variations in form in a given built environment. As we experience the complex network of formal relationships which go to make up a particular settlement or its component parts, palpable similarities will arise—recognizable patterns of repeated forms or patterns of composition of various forms in a three-dimensional array. Certain “underlying regularities” become manifest on nearly any level of sensory experience. Often such regularities do not cross the boundaries of specific neighborhoods, while at times whole settlements manifest a homogeneity in formal appearance. The highly complex cognitive mapping or internal imagery which we as viewers and users often manifest in our memory of given places is partly idiosyncratic and partly held in common by others in a population.¹ What is it that constitutes the data of our perceptual structuration of environments? Or, to turn the question around, how do built environments address our perceptual and cognitive mechanisms, which have surely in some sense evolved to interact with these remarkable artifacts which our species has built itself into?

At the same time that we are interested in establishing groups of regularities or broad patterns of organization in specific architectonic systems, we are concerned with what may be held in common by architectonic systems in general. It will be clear that any set of “architectonic universals” will be a subset of the former. Architectonic analysis has as one of its aims the establishment of what might be justly termed the “design features” of human architecture, namely those features or properties of organization which simultaneously describe the architectonic system per se and distinguish it from “animal architecture” on the one hand² (assuming there are significant and palapable distinctions) as well as from other semiotic systems.

It would appear that every architectonic system shares at least this one thing in common: that they are made up of what we might call ‘space-cells’ (of a wide variety of geometric configurations and sizes). That is to say, portions of the spatial continuum are bound off from each other in a great variety of ways: spaces are enclosed, or mass-forms are so composed as to suggest enclosure of some sort. Human activities are localized in portions of the environment in relatively habitual ways.

An entire settlement may be seen as a complex space-time framework for human action and interaction. It is a spatio-temporal framework in that activities are framed or situated not only spatially but also sequentially. An architectural object of whatever size and complexity is “addressed” by routines of behavior which unfold over time and which often are sequentially situated in groups of space-cells composed in three-dimensional aggregates. The linear or multilinear geometry of behavioral episodes are, as it were, “mapped into” the multidimensional geometries of architectonic objects.

Furthermore, this interactive mapping of activity and built environment is simultaneously specific and arbitrary: at a given moment, a behavioral episode is mapped onto a specific space cell or across a group of space cells, and yet this association may be temporary, arbitrary and conventional in nature—at the next instance, what we might recognize as a similar or equivalent behavioral routine or activity is differently mapped. We manifest a remarkable ability to employ a given space cell for a variety of apparently contradictory activities. Given certain palpable physical constraints, we appear to be able to “use” almost any bounded space for almost any activity.

It would seem that spatial behavior and built environments manifest a certain semi-autonomous relationship—much less rigidly determined than that between the behaviors of certain social insects and their hives or nests—and yet somehow not entirely or completely arbitrary, being to some extent grounded in the routines of convention which differ from one society to another. Furthermore, even within the conventions of “behavioral mapping” apparently canonical to a given society, we may observe wide variations in usage among its members—certain spaces are minimally used by one sex, or used in palpably different ways, or used by various social moieties in different ways on festive occasions, contrasting with “secular” usages at other times, and so forth.

In addition, if we are to say that a built environment suggests certain “readings,” then such “messages” may be decoded inside-out or upside-down, unilinearly or as a series of simultaneous embeddings, and so forth. In certain instances, a built form may suggest a “grammatical” procedure for its spatially-sequenced “unfolding,” but this may not necessarily mean that we cannot experience its organization “backwards.” The space-cells of a compound are rarely composed of moving walkways going in only one (or two) directions, and a given cell may participate in the intersection of a large number of experiential routines.

Hence, while existing as objects composed of parts simultaneously “given,” architectural artifacts may be experienced, read or decoded in ways which may be multidimensional and multisequential in the extreme, in conventional and idiosyncratic ways. But it is evident that not every corpus allows the same degree of interpretative freedom.³

Let us return to the question of the nature of architectonic units. We have asserted that what may be held in common by all architectonic systems is an entity we may term the space-cell or space-frame manifesting a topological property of boundedness. Hence in a very general manner, all space-cells are equivalent in a topological sense. It is simultaneously apparent that space-cells differ along what might be termed a geometric axis. In other words, from a topological perspective, two space cells will be equivalent though they may differ markedly in their formal or geometric realization, i.e., from a geometric perspective; X is a cube while Y is a cylinder.

It will be immediately apparent that two cubic cells, X and X’ will be equivalent geometrically while exhibiting variation in material or physical expression: X is constructed in ferroconcrete while X’ is realized in wood, or brick, or animal skins, or may be simply marked out by twelve logs arranged horizontally and vertically, without infilled “walls.”

We can carry out this variation even further: cubic cell X, realized in brick, may contrast with cubic cell X’ realized in brick which has been painted blue. Or, X may be a cubic cell 3 meters on a side, while X’ may be a cubic cell 4 meters on a side (whether or not both are brick, or even if they are brick and painted brick).

It will be clear that equivalent geometric forms may exhibit variation on a wide variety of realizational axes, and two geometrically equivalent cells may contrast as to media, texture, color, absolute size, and so forth. It would also seem clear, at least intuitively, that such variation is in some sense “meaningful”; i.e., that within the conventions suggested by a given social group, variations in form or material expression are intended to be matched by variations in meaning or function or association, however subtly.

The nature of this “matching” is of course highly complex, and resists being forced into a simplistic “form follows function” mold or into a completely arbitrary and idiosyncratic model. As we shall see in the course of the present study, the literature on “architectural meaning” is a mare’s nest of conflicting proposals and ad hoc suggestions on the part of social scientists, anthropologists, art historians, and architectural semioticians. (At the same time, there is so much that is intuitively insightful in the traditional literature that we would ignore it at great peril.)

The principal problem facing the analyst of built environments is “what is it that is constant (invariant) across a given array?” and, correlatively, “what are the parameters of variation which formative disjunctions undergo?”

In order to measure variation, the analyst will necessarily apply to an array some constant measure or frame within which variation can be highlighted and classified. Clearly, the choice of a reference frame will crucially determine the course of an analysis and the nature of its salient results. To be sure, at the outset of an analysis, the choice of reference frame may be more or less intuitively made, as it is evident that the process of measuring variation will simultaneously serve to clarify the conception and formation of the reference frame itself.

It should be maximally productive if the frame of reference is in some useful sense a simple abstraction of some recurrent formative component in arrays, a component which moreover appears to recur generally from one code to another.

As we have observed above, what is common to any built environment is some geometric transform of what we will call the space-cell, namely a co-occurrent mass-and-space formation revealing an overall property of boundedness. Such an entity comprises a contrastive opposition of simultaneously-occurrent and mutually-defining masses and volumes, defining a topological unity.

From the point of view of such a primitive entity, all the space-cells of an architectonic system will be topologically equivalent. In other words, behind the variations in geometric formation (and of course in material articulation or realization), all space-cells in a corpus will share a topological property of boundedness.

In order for the notion of the space-cell to be sufficiently general, it is necessary to distinguish it carefully from the geometric or formal concept expressed by (for example) the English lexical item “room,” which generically refers to a spatial or volumetric configuration delimited or bounded by one or more masses. The notion of the space-cell will also include the obverse of a “room,” or mass-bounded volume, namely space-bounded masses, or an object and its ‘surround.’ Moreover, a cell may result more generically from place-making constructions generated solely by bodies themselves.

Thus, formations as geometrically various as the following will be considered as equivalent in a cellular perspective:

Clearly, this is not to exclude other perspectives which will distinguish (a) and (b);or (a,b) and (c,d);or (a,c) and (b,d).

Furthermore, we may distinguish a series of space-cells (a_1, a₂, . . . a_n) wherein geometrically-equivalent cells contrast in terms of (1) absolute size, (2) constructional material, (3) color, (4) texture, and so forth. Each of the latter may provisionally be regarded as providing contrastive perspectives (or axes of contrast). Thus, (a₁) and (a₂) may share the same geometric formation as well as the same size (1), but may be widely divergent in terms of medium, color and texture (2,3,4);—i.e., materially.

In addition, two space-cells which are geometrically and materially identical may be associated with quite different functions or usages, and, furthermore, this contrast may or may not be correlated with contextual differences of relative placement in multicellular compound.

It will be evident that the analyst is dealing with objects which in various ways display a variety of component features or properties of formation, and that the task of analysis is concerned in part with establishing the diverse ways in which some properties of a given formation will contrast with some properties of another formation. In this way, we will come to understand the ways in which two isolated portions of an array may be both similar and different.

In addition, architectonic objects reveal component features of formation which are significant in different ways. A complex object such as a building may be constituted such that certain of its component elements are directly significative, while others may be significative only in concert with other elements.

For example, the use of certain colors or materials may be directly associated with the identity of a given social class or with membership in a certain kinship group. In such a case, the presence of that material will be contrastively opposed to other materials which will be differently significative. Thus, the color crimson may be associated with an aristocratic clan, standing in opposition to (a) all other colors, which may or may not be specifically associated with other social groups, or (b) it may be successively opposed to other colors (as where crimson : blue : : aristocracy : merchant class, and crimson : green :: aristocracy : military, and so forth).

Furthermore, there may be cases where crimson-in-context X (limestone masonry) contrasts with crimson-in-context Y (timber frame construction; or where crimson + limestone masonry + curvilinear space-cells contrasts with crimson + timber frame construction + curvilinear and/or rectilinear space-cells.

In addition, other kinds of contextual conditions will modify the significative presence of a given material or geometric formation (such as location in an urban fabric, relationships with other buildings—contiguous vs. isolated, or directly on a street vs. removed from a street, etc.). The analyst will discover a highly complex hierarchy of significative associations wherein large numbers of elements in an array are ordered with respect to each other in many different ways, providing members of a society with an enormously varied set of associations and grades of significative modification.

In a similar fashion, but with quite different means, a verbal utterance will provide a listener with a great deal of information, e.g., about the referential context of a message or other orientations toward various components of the message in a functional sense,⁴ and many different formative components of the message may signal information about the social class, emotional state, etc., of a speaker and/or the speaker’s covert opinions regarding a given referent. Such aspects of the utterance as intonation, syntactic usage, phonological clarity, and so forth, may be directly significative in a wide variety of code-specific ways.

Architectonic objects are normally composed of various sign-types such that the overall significance of the object comprises a complex set of divergent associations. This may be easily demonstrated by a consideration of what constitutes “house” or domestic structure in one’s own culture. This may include (a) characteristic cellular formations (or the exclusion of certain types of formations); (b) characteristic sequences of cells, either (1) geometrically or (2) functionally; (c) the use of certain kinds of materials, colors, finishings, sizes, etc., or the exclusion of various kinds; (d) characteristic types of infrastructural fittings (furniture, with its own patterns of allowable interrelationships, mechanical support systems, etc.); (e) characteristic relationships to an urban fabric (e.g., if contiguous with other houses, minimal separation from a public street, or if isolated from other houses, set back behind walls, gardens, etc.); (f) characteristic sizes and positions of formative components such as windows, doors, stairwells, etc.; (g) the presence or exclusion of various kinds of symbolic motifs redundantly signifying domesticity.⁵

The following example may serve to illustrate the kinds of formative complexity typical in an architectonic code.

FORMATIVE UNITS

Plates I through VIII consist of isometric reconstructions of the groundplans of several contemporaneous domestic buildings erected on the island of Crete ca. 1500 B.C. Of interest here are the series of space-cells labelled in each diagram (a)(b)(c), known in the archaeological literature as “Minoan hall systems,” associated generally with the private domestic quarters of Minoan houses.⁶

The structures have been taken from several towns on the island, and are chosen at random to illustrate some of the variety of formations evident in the corpus. The buildings are referred to in the following analysis by their conventional names as fancifully assigned by their excavators.

They are:

In the first example (AKHL) we are concerned with cells (a,b,c) and their relationship to adjacent aggregates (m-n) and (g-h).

The hall system consists of (c-b-a): entrance to the system is from vestibule (h) via a double door opening both into the left-hand cell and the central cell (a,b). Cell (a) is formally realized with its right-hand side as a colonnade (/c/); cell (b) with its left-hand side that colonnade and right-hand side a pier-and-door partition (henceforth referred to as /PDP/); the right-hand cell (c) with its left-hand side a /PDP/.

Cell (c) may have been a dining room: it contains an L-shaped bench on the two far walls, and (c) abuts two smaller cells (m,n) which appear to have been a kitchen and pantry. All three cells of the system (a,b,c) connect with cells to the side flank of the system (h,h/g,g, respectively). The portion of cell (g) connected to system cell (b) is paved with flagstones (the rest of the floors in the house are beaten earth), and may have been a small light well. If cell (a) was similar to equivalent cells at TYL, it too may have been a light well.

The second house (KN HCS, Plate II) reveals a similar pattern (cells (a-b-c)), but here cell (c) is connected laterally to a curious cell (d) with a raised dias on a platform, of unknown usage. Entrance to the system is, similar to AKHL, laterally into the central cell (b). Here, however, the wall-position between cells (a) and (b) is not realized as a colonnade but as a /PDP/. There is a colonnade separating cells (c) and (d) however.

The third house (KN RV, Plate III.) has the arrangement familiar at AKHL A (cells (a-b-c)), plus a fourth cell (d) beyond (c) and behind a balustrade, of unknown usage.

The fourth (KN HF, Plate IV.) has a hall system of three components (a-b-c), where again cell (a) is a lightwell (as in the previous example), but its separation from (b) is not a colonnade but a /PDP/ plus a wall-extension, and the separation of (b) and (c), originally perhaps a /PDP/, is now a wall with two doors.

The fifth example (ML DA, Plate V.) has the familiar arrangement (cells (a-b-c)), and entrance is gained through the right-hand cell (c) via a /PDP/.

The sixth (ML ZA, Plate VI.) is identical to the previous except that cell (a), a light well in ML DA, is here a garden; but entrance is similarly gained through right-hand cell (c) (a single door, however, and not a /PDP/.

The seventh and eighth (TYL AC, Plates VII. and VIII.) are also similar, the hall systems in TYL A being cells (a-b-c) and in TYL C being cells (a-b-c).

We will refer to the three components of all the hall systems as the “a” cell (or lightwell), the “b” cell (or middle room or “porch”), and the “c” cell (or “hall”), and the notations in our Plates are identical on this count.

Clearly, we are dealing with a constant or invariant cellular ‘syntax’ (a-b-c) across a number of contextual variants or alternative formal realizations. Note the following correspondences:

Table 1. Formal Cellular Realization

The internal syntactic aggregation of all examples is the same, but there are alternative formal realizations, as the Table above indicates:

(1) the (a-b) boundary may be either ICI or /PDP/;

(2) the (a) cell is invariantly an “outdoor” or unroofed cell (or a cell with a clerestorey roof), but may be either an enclosed open area (“garden,” as in ML ZA) or a relatively small lightwell. Hence there may be alternative formal realizations. In the hall systems of the two large “palaces” of Mallia and Phaistos,⁷ we find both a lightwell and a “garden.”

(3) The (a-b) boundary realized formally as ICI may be either L-shaped (TYL A) or unilinear (all others).

(4) In the example of KN HF, the (a-b) boundary is not only a /PDP/ and not a /C/, but it is only partly a /PDP/.

In our examination of TYL A and C we arrive at the tentative conclusion that there was an apparent invariance within the hall systems with respect to relative size of cells: at TYL, the (b) cell was the smallest, the (c) cell the largest.

But if we take all of the examples into consideration, it will be evident that this size-ratio is not invariant. In the next Table, “1” = largest cell, “2” = middle-sized, and “3” = smallest cell:

Table 2. Relative Size

Evidently, then, the relative size of the (a-b-c) cells is not invariant. Consequently, there will be alternative material realizations with respect to size (modularity), or what may be termed contextual variance.

Let us look at several other aspects or features which may be invariant across our examples.

With respect to the manner of access to the hall system—i.e., its syntactic position with respect to other aggregates—it will be seen that primary access (i.e., from cells leading to the building’s entrance) is as follows:

Table 3. Primary Access

Evidently, access from the outside may be gained into any of the three cells, and in two cases into more than one: at AKHL A, entrance-vestibule (h) has a door opening into (a) and one opening into (b); at KN HF, house vestibule (f) has two doors: one leads into a corridor (e-d) leading to (a), another into a series of cells (h-i) which lead into cell (b).

But what is invariant, however, is the fact that in all examples, access is lateral to the longitudinal axis of the system.

It will be noted that in all cases, the hall systems do not open directly onto a street or primary house entrance, but are at least one cell removed from that entrance: in the following Table, the access to the system is shown with respect to how many thresholds (doorways) separate it from the primary house entrance (E).

Table 4. Separation from Outside

In other words, there is invariantly a minimum of one cell between hall system and house door (or two doors), and the system may be as many as three cells removed, possibly more depending on the size of the house. But no matter how small the house, there is always at least one intermediary cell.

Thus far, we might specify the formative features of the hall system as consisting of

(1) an aggregate of 3 cells of type (a-b-c)

(2) in a linear sequence (a) + (b) + (c)

(3) lateral to direction of access

(4) forming a portion of a larger aggregate such that the system is 1 or 1+ cells removed from entry into the overall aggregate

It will be understood in (1) that the cells are contiguous, and share party walls (a/b and b/c). We have seen above (Table 1.) and (p. 18) the range of formal realizations of the cells regarding such party walls (/C/, /PDP/).

There are other features closely connected with the hall systems, notably the presence of stairwells to a second storey. These occur either in a cell adjacent to one of the cells in the system (but never (a)), or open directly into one of the cells:

Table 5. Associated Stairwells

In all cases, the stairwell is closely associated with the (c) or (b) cells of the system, and is, moreover, distinct from stair wells(s) which serve more “public” portions of a house. In other words, the hall system-stairwell is one of controlled access, behind doorways which are “beyond” (with respect to outside entrance into the house) the system itself. In several cases (TYL A C, KN RV) there are two private stairs nearby.

The houses also have several other types of cell, often closely associated with the hall systems, though not invariably so:

(1)“pillar crypts” or shrines:

    KN HCS (h), off (i) off (b)

    KN RV (e), off (c)

(2)pillar rooms, perhaps not shrines:

    AKHL A (i), off (h) off (b or a) (workshop?)

    TYL A (p), off (b)

    TYL C (p) (workshop?)

(3) washing/lustral cells and/or latrines:

    AKHL A (f) = latrine?

    KN HCS (f) = (sunken) bath

    KN RV (h) = latrine?

    ML DA (f) = (sunken) bath

    ML ZA (g) = (sunken) bath; (h) = latrine

    TYL A (d) = latrine?

    TYL C (f) = latrine

In addition to the connections providing access into the hall systems, principally from the primary house entrance, some cells in various systems provide connections elsewhere in the house:

Table 6. Secondary Access

Consequently, several cells in the systems have at least a dual function: as a component in the system itself, and as a passageway to other cells outside the system (apart from connections to primary access (CP) above). In TYL A, for example, cell (b) serves simultaneously as a "b" cell and as a segment in a continuing corridor "bisecting" the system. In KN RV, cell (c) provides access to 3 cells, and serves as a segment in the passagefrom the left to the right side of the structure.

We have to add to our list of formative features of the hall system given above the following:

(5) the system is not self-contained with respect to other cellular aggregates, but may function as a passageway to other aggregates.

The following diagram (Figure 1.) portrays the invariant sequential order of the formative features of the hall system;

Figure 1

(understanding the range of formal and material alternative realizations of (a-b-c)). Hence the formative features of the Minoan hall system include

(A)the specified set of components (a), (b), (c), and their formal (and material) contextual variants;

(B)an ordering or structure of those components, as specified in Figure 1.

Included in both (A) and (B) are series of formal, material and syntactic constraints as specified in the discussions in this section. Thus, the specification of lateral access to the system in (B) implies the formal realization of at least one doorway in at least one wall of one (or two) cells, and so forth. The specification of (a) as lightwell implies the material realization of formal item —as here, viz., floor— as a cement or stone pavement, etc.

Let us now consider a comparative example and examine the manner in which other houses in another corpus are organized in terms of formative features.

Plates IX. and X. are plans of two Egyptian houses, at el-Lahun and Amarna. The former, dated to the reign of the Pharoah Sesostris II, in the Nineteenth Century B.C. is a row house in a planned village built in association with construction of that Pharoah’s pyramid. That at Amarna is a detached house in a yard, the residence of a bureaucrat known as the vizier Nakht.

The latter is considerably smaller than the former, and has many fewer rooms; it was constructed in the Fourteenth Century B.C., or some five or six centuries after the former.

Despite wide differences in size and layout, the two houses reveal a number of similarities-pillared halls of several types, for example, arranged in sequence. But the interrelations of the various cells appear very different in the two houses:

Figure 2

The networks of cells (their syntactic aggregation) is quite distinct; except for one connection (“j” in our diagram) which makes a looped “string,”^a el-Lahun is organized strictly as a series of terminal “strings.” On the other hand, Amarna reveals a mixed organization of terminal and looped “strings” (i.e., concatenated sequences of space-cells), including a loop within a loop, reminiscent of characteristic Minoan organizations.

What is constant across these aggregations are the following features (refer to the plans in Plates IX. and X.):

(1)Amarna (AMR):2 small square cells immediately abutting the primary entrances (a,b); el-Lahun (ELL) has two similar cells, but they are separated by a long north-south corridor with two halves (x);

(2)The second vestibule in both cases opens onto a large cell: a colonnaded court at ELL, a hypostyle hall with 8 columns at AMR (cells (c) in both plans);

(3)To the south of the latter is a 4-columned hall (d) in both houses; but between (c) and at AMR is another cell (cd);

(4)To the south of the latter is a one-columned hall (f) in both houses;

(5)To the side of cells (d) is, at ELL, a colonnaded court, and at AMR 4 6-columned hall:(e) in both plans;

(6)To the east of cell (f) is the master’s bedroom (mbd), to the west the mistresses’ bedroom (fbd);

(7)A connection with servant’s quarters is made at (j) in ELL, into the master’s area, and, at AMR, at (k), into the women’s area;

(8)To the inner hall of the (mbd/fbd) areas are appended a series of small cells.

It is evident that the two houses have in common a great deal—the identity of given cells and their relative sequencing, as well as their deployment on a north-south axis, corresponding to a progression from larger and more public spaces to smaller and more private spaces, diverging into two private sub-zones, a men’s quarters (on the east) and a women’s quarters (on the west).

It becomes apparent that the long north-south corridor at ELL is needed to bring the progress of entrance onto a “fixed” north-south progression. Because ELL stands on the north side of a street, the visitor must be brought around to the north side of the house before being allowed to penetrate into the private dwelling areas. Similar houses on the south side of the street at ELL don’t have the same problem, and the long corridor is absent.

The houses differ markedly in that where (c) or (e) are columned halls at AMR, they are courtyards at ELL. Perhaps because AMR stands in a large walled yard, outdoor courts within the structural frame of the building are unnecessary; ELL is a town house packed tightly alongside others.

But the relative positions of (c) and (e) in both houses are identical, and in the case of (c) its long axis is perpendicular to the long axis of the progression (c)-(d)-(f); in the case of (e), in both cases it approximates the size and proportions of (d) to its side, even though (e) in ELL is a court, and (e) in AMR is a hypostyle hall.

In addition, the relative proportions of (c), (d), (e), (f) are constant, and the sizes of (mbd) and (fbd) are similar in both houses.

We might portray the sequential or syntagmatic ordering of the formative features of the two houses as follows:

Figure 3

Compare Figures 1. and 3. It will be evident how the formative features comprising the Minoan hall system differ from those comprising the Egyptian hall system, as well as how the orderings of those features contrast in both cases. Note with respect to the latter that whereas the Minoan system is invariant with respect to lateral access, the Egyptian system is invariant with respect to cardinal orientation (north-south progression from outer, public spaces to inner private spaces); the Minoan system lies perpendicular to the main direction of access from the exterior (E), while the Egyptian system lies along the same longitudinal, linear access. The Minoan system manifests variability with respect to connection of access to hall components, while the Egyptian system is invariably accessed at the largest, outermost cell. The Egyptian system, furthermore, bifurcates into two roughly identical aggregates, one associated with a male domain, the other with a female domain; no such bifurcation is apparent in the Minoan system (but note that we don’t know what went on on the second storey (absent in these Egyptian houses), in the organization of private sleeping quarters (or summer sleeping quarters?).

Two examples hardly constitute a corpus, and our intent has been principally to suggest the directions of an ongoing and more comprehensive analysis. Nor is it implied here that the formative (or significative) structure constant across these two examples will not be modified by further analysis, no matter how remarkable it may seem that these two houses, separated in time by nearly 500 years, are apparently so similar in organization. Our aim here has been merely to provide a contrastive example to what was discovered above in the examination of the Minoan houses, and, through this, to suggest some of the dimensions whereby architectonic codes will contrast with each other on many levels of formative organization.

Architectonic codes employ as significant distinctions in form only part of the total number of possible distinctions which could be made. Furthermore, what is significant in one code may be non-significative in another code. As the above examples demonstrate, the cardinal orientation of a matrix of cells in an idiomatic aggregate is significant in the Egyptian corpus but nonsignificant in the Minoan. In the former case, there is a correlation between the orientation of the cellular matrix north-: -south :: more public-: -more private. In the latter case, more private is signified by an assemblage of certain diverse features of formation and their alternative realizations. Thus, the hall system is either physically far-removed from a house entrance, or, if relatively close, is approachable through multiple doors and/or by abrupt 90-degree changes of direction, etc.

The comparative analysis of architectonic corpora will highlight the evident conventionality of signification and the code-specificity of meaningful formation. In the next section we will explore further the notion of code-specific significance, and propose a system for the classification of architectonic signs.

PLATE I

PLATE II

PLATE III

PLATE IV

PLATE V

PLATE VI

PLATE VII

PLATE VIII

PLATE IX

PLATE X

FOOTNOTES

CHAPTER II

¹A great deal of important research has been carried out in recent years on the subject of cognitive mapping. A keystone in this work is K. Lynch’s The Image of the City, Cambridge, Mass., 1960; see also the same writer’s What Time is This Place?, Cambridge, Mass., 1975. In our bibliography below we have listed a number of especially important studies in this area, one of the most important frontiers of architectonic research. Particularly insightful is the theoretical and experimental work reported by Moore, Downs, Stea, Rapoport and others in recent years.

²On the subject of comparative architectonics, see the discussions in Preziosi, 1978c.

³We are accustomed in the highly eclectic built environments of the Western world to a rather wide range of free variation in interpretative association, but this is not the case in traditional societies, or in the Western world in the past. This is an important question, but we must take care not to confuse the trees for the forest, for it may well be that our apparent freedom for ‘idiosyncratic’ interpretation is itself carefully circumscribed. This remains a salient issue for future study.

⁴See, on architectonic function, our Chapter on Meaning below, and Preziosi, 1978c, Chapter III, and id., ‘The Parameters of the Architectonic Code,’ 1978f, forthcoming, in the journal Ars Semeiotica, 1978.

⁵An interesting, albeit simplistically behavioristic, introduction to the subject of house-symbolism, may be found in C. C. Marcus, ‘The House as Symbol of Self,’ in Designing for Human Behavior, 1974, 130-146.

⁶This material has been chosen as an exemplary corpus of data in part because of field work done by the writer in this area, involving a detailed survey of some two hundred structures over several seasons. This work is reported in detail in the writer’s doctoral dissertation, Minoan Palace Planning and its Origins, Harvard University, 1968 (see Bibliography below). The Minoan system has also been chosen as an exemplar because of its relatively limited extent (in contrast to the highly complex environments of modern urban settlements), involving the handling of only a few thousand cellular items, plus the fact that in contrast to systems elsewhere at that time, it reveals a great deal of formative complexity. No two Minoan buildings are identical, and the amount of formal and material variation in the corpus is very great. Another reason for the choice is historically sentimental: Minoan settlements comprise the first large-scale complex townscapes in Europe. Since the time of these analyses, work has been carried out in a variety of other corpora, both contemporary and historical. Reports on these are currently in preparation.

⁷The large central megastructures in Minoan cities, commonly called ‘palaces’ in the archaeological literature, also incorporated residential apartments sharing formative features with the Minoan houses looked at here (however else these megastructures functioned).