Electrical engineers have nice definitions for the foundations of their field:

We can define

and so forth.  But in robotics every term is subject to change without notice.  There is no common definition of robot, or autonomy, or intelligence.  The only common element is that researchers are still, after three or four decades, able to argue about what the definitions ought to be.

Thankfully, the discussions have died down from religious wars into cocktail party conversation, and the heated debates have largely transitioned from the definition of “robot” to the definition of “humanoid” or even whether topic X is “finished” or not.

It’s unclear whether the cessation of hostilities is due to weariness, to progress, or to maturity.

We could just be tired of going round in circular arguments, from “it has to do useful work” to “entertainment is a job” to “but it’s a toy, not a robot”.

We might have reached a common core definition, so the arguments are in the details around the parameters of the definition.  Instead of arguing about the definition of robot as “physically instantiated self-contained machine that performs a task for a human” we’re arguing about the definition of “task” and the definition of “self-contained”.

Or we might have concluded that what’s important is not that there be a single definition, but that individuals with something to say have the ability to say it clearly.  Instead of arguing when someone uses a non-standard definition, we simply accept it for the duration of the talk, or the paper, or the conversation, and try to understand the meat of what they’re trying to say.

The trouble is that whichever of these is the true answer, we are no closer to having an introductory textbook in robotics with a single definition that we can all agree is close enough to correct for it to be taught to beginning students.

Leave a Reply

Your email address will not be published.