Engineering as lifeworld

Much ink has been spilled about the proverbial “physics envy” allegedly felt by the practitioners of various disciplines that are even the slightest bit quantitative. In machine learning or AI, for example, we constantly hear about how the field needs more science and less engineering. The latter word is, by and large, used pejoratively to connote hacking, tinkering, all those messy activities that no respectable scientist would be caught dead doing. This is deeply misguided on multiple levels, not the least of which is that, in fact, machine learning is inherently an engineering, not a scientific, endeavor, concerned with using empirical observations, domain knowledge, and computational methods to create artifacts in the sense described, for example, by Herbert Simon in The Sciences of the Artificial, or by Sanjoy Mitter in “Toward the definition of a new engineering education”:

What distinguishes engineering from the pure sciences (read physics) is that engineering is concerned with synthesis of new systems (invention, synthesis of new laws) and not discovering the fundamental laws of nature which exist and have always existed. This aspect of synthesis, which is captured in a deep way by the “black box” approach, requires the understanding of synthesis of “elements” of a “system” to follow a prescribed behaviour. Problems of approximation of models, dealing with uncertainty in models and in interaction with the environment, understanding complexity are all fundamental aspects of this synthesis question. It requires looking at systems both in terms of its “parts” and as a “whole.”

Sanjoy then goes on to state that engineering education should not be viewed in opposition to its humanistic counterpart, but should instead be integrated with it. Good engineers should be just as steeped in the humanistic view of the world as they are in the technological one. This echoes the point made by Reza Negarestani in a wide-ranging interview with NERO Editions:

Neither philosophers nor political theorists are able to design proper methods adequate to actualize possibilities, imagined or not. We need politically and philosophically informed engineers and designers. Engineers are indeed not mindless technicians, they are people who have one foot in the domain of thinking and one in the realm of an external reality or worldly affairs. They do not see action as a form of hubristic mastery to the extent that they know whatever we do at any level of reality—be it natural, social or cultural—will meet the resistance of that reality. To use a Sellarsian metaphor, reality in the broadest possible sense is not a block of wax ready to be imprinted. Engineers truly know that. They also never see reality in any sense as a flat universe, they see it as vast and deeply multi-scaled structure. In order to concretely intervene at any level of reality we must not only have a multi-level view of the reality but also know which methods, models or tools should be implemented, and at which level. To cut at the joints without splintering the bones is a description of what engineers—as Plato’s good butchers—do.

Over the last several years, I have become deeply interested in the philosophy, history, theory, and practice of engineering. Perhaps this sort of soul-searching is what every senior(ish) faculty member goes through; be that as it may, my aim is to write about this sort of thing here.