Not the point of the article at all, but in my experience when an author says "canonical map" it can typically be made precise as a unique natural transformatio

Not the point of the article at all, but in my experience when an author says "canonical map" it can typically be made precise as a unique natural transformation between whatever functors capture the objects of discussion. I think this is part of what MacLane meant by Category Theory being invented to talk about natural transformations.