[3] In 1996, Johnson co-authored, (with Jeffrey Elman, Annette Karmiloff-Smith, Elizabeth Bates, Domenico Parisi, and Kim Plunkett), the book Rethinking Innateness,[7] which examines neural network approaches to development.
[8] In the book, Elman et al. propose that genetic information might provide "constraints" on how a dynamic network responds to the environment during learning.
For example, they suggest that a learning system can be seen as being subject to architectural constraints during development, an idea that gave birth to the neural network field of constructivist modelling.
Rethinking Innateness has received more than 1,500 citations,[4] and was nominated as one of the "One hundred most influential works in cognitive science from the 20th Century" (Minnesota Millennium Project).
As such, it echoes contemporary work in other areas of development, such as probabilistic epigenesis and gene regulatory networks.