Hi all,
Thanks Christof for the long answer! I looked into this a bit and it appears that every time you create an [array define myname] object there is a canvas named pd-graphX (created by array_define_new). If you know the order of creation of the array, you can send [; pd-graph1 vis 1(, and (probably) any other message you send to canvases. You can even rename it: [; pd-graph1 rename dog(, and then just send [;pd-dog vis 1(; The name won't stick, tho, on next load.
I attached an example of this. You might need a fresh instance of pd for it, or just tinker with the graph indices... The thing is that the graphX list is generated from scratch every time pd loads, and keeps growing on every new instance of an array, no matter what you do.