Learning the Innovation Journey: Order out of Chaos?

Abstract

This paper examines whether the developmental process of two biomedical innovations followed either (1) an orderly periodic progression of stages or phases, (2) a random sequence of chance “blind” events, or (3) a seemingly random process of chaotic events. Various diagnostics are applied to distinguish periodic, chaotic, and random patterns in time series data on innovation development events. We find that the actions and outcomes experienced by innovation teams exhibit a chaotic pattern during the initial period of innovation development, and an orderly periodic pattern during the ending development period: however, exogenous context events exhibit a random pattern during both the beginning and ending periods of innovation development. These research findings, if substantiated in other studies, significantly alter prior views of innovation and learning processes. The two most commonly-used explanations of the innovation process, that it follows either an orderly periodic sequence of stages or a random sequence of “blind” events, are not valid where chaos is found. Chaos tells us that the innovation process consists of a nonlinear dynamical system, which is neither orderly and predictable nor stochastic and random. The findings of chaos also expand existing definitions of organizational learning. Learning in chaotic conditions can be viewed as an expanding and diverging process of discovery. Learning during more stable and periodic conditions is viewed as a narrowing and converging process of testing. Timing of transitions from chaos to periodic patterns may explain whether decision or action rationality prevails. Hence, the difference between behavioral learning theorists and cognitive learning theorists may be just a matter of time.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 499 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.