In one respect the effect of these broken figures is not all that different from the open matrix of sliding planes. Space—whether conceived of as that universal spatial continuum of modernity, or the actual but open space of a palpable portion of the world (a site)—flows through these fragments. The projects propose a sort of Trojan horse of conventional architecture whose syntax, upon inspection, dissolves into a series of fragments. Space, or site, passes through them just as it does through the walls of the pool project. In the Alves Costa house, an emblematic moment occurs at both the front door and at the garage. At both these points fragmented figures overlap trailing walls, like stiff streamers, into the field of another figure. These trailing walls disrupt the sense of closure that the figure into which they penetrate might otherwise offer. Thus the virtual closure suggested by the figural fragment is conceptually undone and the reading of an open spatial syntax of walls—dividing up space—is forced upon it. In the garage in particular, a low extension of the east wall of the house slides under the open hanging corner of the structure, while the wall extends into the figural domain of the house. Both walls, each a part of a figural fragment, disrupt the spatial integrity of the other fragment into which they penetrate. The two figures' conceptual identity flip-flops as these walls are understood in one instant as boundaries of space, in the next as overlapping dividers of space. 

As with the pool project for Leca, the syntax of the Costa house is spatially porous. The conceptual transparency to the field of the site, the conceptual presence  of that field in the midst of the very figures enclosing the dwelling space of the house, presents to us the house as intervention "layered" into the site, an open sketch on the site—and thus the persistence in these projects of the archeological metaphor. 

The figurative expectations that the fragments set up—the expectation of closure that might have been absent in the more apparently modern and de Stijl syntax of the pool—in some respects amplifies the peculiarity of a conceptual intrusion of the site into the house, even in the absence of great rocks. 

The percourse into the house adds another peculiarity. With the apparent conventionality of the ceramic tiled roofs and the bounding of figures, the expectation that one might move through the building in a more conventional pattern also grows. Yet instead of, for instance, passage into a bounded room through a cut in the wall—a threshold, that is—at the front and back doors a person would, as the space described above did, move between the fragmentary figures as if they were a landscape of ruins. Here we begin to see a theme that will develop with more didactic clarity in the succeeding projects, but the notion of how the subject is placed in contrast to the weight of latent conventions of architectural figures begins to emerge. The split between how human movement and perception is orchestrated in contrast to certain conventionally apparent orders of the architecture begin to create an architectural corollary to the sketches we have described. 

From the 1970s Siza's work begins to exhibit more explicit uses of type. In projects for housing we see a pattern of siedlungen-like town houses (the SAAL housing at Bouca, 1973–1977; Sao Victor, 1974–1977, both in Oporto; and housing in Caxinas, 1970–1972). In several other projects we begin to see the repeated use of U-shaped courtyard schemes (the Pavilhao da Faculdade de Arquitectura, 1984, the Carlos Siza house, 1976–1978, and the Escola Superior de Educacao in Setubal, 1986–1992).  

Certainly, the concept of type is tricky and has changed over time. But let us say, for instance, that the "U" that appears many times in Siza's work is a configuration of form that wakes in us a chain of associations with other like configurations. It tends to be nameable, because it is that very characteristic—that it belongs to a category—that constitutes the being of types. What I have referred to as syntax in the case of the pool does not constitute a nameable configuration. It is more in the nature of a strategy or pattern of form than a nameable entity as a type must be. Thus although Siza was using such syntactical patterns, he was able to avoid a certain aspect of that initial anxiety about preestablished languages. Flexible spatial patterns appear to be more spontaneous and less burdened by history. 

Yet because the type has a certain integrity as a conceptual category, it also implies a kind of closed autonomy; its stable and independent conceptual existence is a form of aloofness. And it is here that it becomes susceptible to both the suspicions voiced by Tavora and Siza as well as Pessoa. It is not "style" but it has something of style's formulaicness. It is not language, but like language it seems public rather than intimate; like words, types seem to exist independent of us. Thus types were held in suspicion by Tavora and his colleagues because they suggested the possibility of a reified formalization of architecture. And even though the vernacular may have been susceptible to a typological survey and analysis, what was held to be appealing in the vernacular were its qualities of flux, its qualities of historicity—its layering of past and present—that seemed a palimpsest of its becoming. We should note that like the  language we speak, type's impersonality is susceptible to that endless reformulation that allows all learned languages to acquire clandestine and utterly unique qualities added by each speaker. The resonance of a word is created by the unique world of each mind, and diction and grammar are shifting sands that reflect the biologically infinite permutation of speakers and history. But types also never lose their fundamental correlation to the historical things by which they steal away from the actual and specific into a realm of remote concepts and categories. 

Types would seem to work against one complex and essential aspect of Siza's archeological metaphor. The manner of layering so far described has suggested a simultaneous intimacy and estrangement between the layers of new project and site. The transparency and conceptual incompletion of the formal language of the project that allowed the "intrusion" of the site's alienness into its midst is not obviously in the nature of the type. This is so because the type tends to be a closed or at least a finite world, which tends to conceptually close out or reorganize in its own manner what lies outside of it. It may rest archeologically on what precedes it, but it excludes those things through its own internal cohesion.