TRIZ was introduced to America unsystematically. As a result even such literate TRIZniks as Larry Ball do not know the evolution of ARIZ in general and evolution of the order in which technical and physical contradictions are resolved in particular. So let's me fill this gap.
Until 1970 there was no concept of physical contradictions at all. ARIZ-68 and earlier versions were aimed at exposing technical contradictions behind a problem and their resolution. Then the concept of physical contradictions was introduced and technical contradictions (and contradiction matrix) became immediately obsolete. The part 2 (called "Analytical stage") of ARIZ-71 instructed problem solver to do the following:
These later versions of ARIZ were introduced to America first (and may be last ?). No wonder that Larry Ball had difficulties with formulating technical contradictions ahead of physical ones, although sometimes technical contradictions lie on the surface. Luckily, the early versions of Invention Machine software implemented ARIZ-71, and Larry Ball learned about resolving physical contradictions ahead of technical ones from them (as he confessed in one of his letters to me). But now he trumpets this inverse order as his own achievement:
Two things are false here. The first stage of ARIZ, which may never was translated into English and which Larry Ball probably does not know, suggests searching prototypes in patent databases. The patent descriptions of prototypes often contain technical contradictions formulated already but not resolved and traded off. Usually they are shallow contradictions, which cannot be resolved without discovering the underlying physical contradictions. Thus, moving from the Technical Contradictions to the Physical ones is sometimes quite logical and constitutes no problem.
Secondly, the "incorrect logic" of the later modifications of ARIZ by no means slowed the advance of TRIZ.