Frances to William with thanks... It seems reasonable that the historical identity of architecture would be an initial starting point in defining architecture in theory, even if much of its history might now be a speculative interpretation. It may be that the first mobile case of quasi architecture outside of a cave was the holding up of a cape as a tent to protect the wearer. If a cape as a tent is in fact and in deed made to be like a cave, then the historic criteria would be based mainly on iconicity and its similarity, rather than on indexicity and its contiguity or on symbolicity and its arbitrarity. My instinctive tendency in considering theory is clearly exposed to lean in the direction of biotics and semiotics and linguistics and logics, rather than historics and anthropics and praxics and epistemics, because it seems to me that the one thing that all architecture may be found as would be as a sign, whether mainly iconic or indexic or symbolic, or mainly in combination of this tern. To support this leaning or bent of mine, the most technically and scientifically advanced architecture after all seems to have occurred in those societies that were highly literate, but further whose linguistic literal language system lent itself easily to typesetting and printing, so that scientific information about the tech and techne and technics of building edifices could be communicated to other persons in other places at other times. This is not to imply that out of literacy will necessarily flow intelligence, but it does suggest that records of fact and law can be kept.
You wrote... Our ideas of architecture, what it is, stem from Vitruvius' 10 books of architecture. I suggest the Loeb Classical Library editions. Another source to look at would be Louis Mumford's The City in History. He says architecture began with the notion of a citadel and buildings expressing power, and evoking "respectful terror". However architecture is defined in theory, its historical definition is a starting point.
