The Pattern of Knowledge Flows between Technology Fields: modularity, near-decomposability and autocatalytic sets. (Preliminary and incomplete draft)

We think of the disciplinary technological knowledge available in the economy as subdivided into di¤erent disciplinary …elds, or simply …elds, which only partly correspond to application domains. A …eld identi…es a set of funcional or ’phenotipic’ traits in the broad technology domain. The set of technology …elds existing at a given date and the relations between them are understood to be the outcome of the way in which the human society has historically explored and exploited the set of possibilities o¤ered by the physical world. Technological determinism is therefore inappropriate, in that there is a social element inherent to the architecture of technology. Innovations, that is, additions to the stock of technological knowledge, can be incremental, radical, or network. The scope of a disciplinary-…eld de…nition is su¢ ciently narrow that radical innovations are understood as exogenous events a¤ecting the number and quality of technology …elds. Network innovations are changes in the matrix describing the active cross-…eld learning interfaces and the strength of these knowledge connections. Incremental innovations are those additions to the knowledge stock originating in a speci…c …eld that do not a¤ect the set of available …elds or the design of cross …eld interactions. The average number of incremental innovations per unit of time in a given …eld depends on two main factors: in the …rst place, the set of innovation opportunities available in that …eld, in the second place the innovation e¤ort in the same …eld. In this paper we assume that innovation opportunities are primarily determined by the progressive local knowledge base. This consists of the subset of ideas that are known by R&D laboratories currently operating in the given technology …eld and that are potentially conducive to useful