[bambax]

As described, the experience doesn't really show that dogs use humans as tools, but rather that they look to their face for hints as to how to behave.

This is the almost-universal behavior of low-status individuals to high-status ones (in a boardroom everyone looks at the boss to know how to react to a joke).

So it could mean, not that dogs "use humans" but rather that they recognize humans (and specifically their trainer) as the leader of their pack, which wolves do not.

It would be interesting to do this experiment again with dogs and a human that they absolutely don't recognize as a leader (a toddler that they have never seen before, for example).

My guess is, if there are two humans in the room, a "recognized leader" and a "recognized non-leader", and if the non-leader knows how to solve the riddle, and the leader does not, then the riddle will never be solved because the dog will never seek the help of the non-leader.

[brational]

>My guess is, if there are two humans in the room, a "recognized leader" and a "recognized non-leader", and if the non-leader knows how to solve the riddle, and the leader does not, then the riddle will never be solved because the dog will never seek the help of the non-leader.

Why do you think that? Why not the dog search each available human for cues?

[bambax]

Because (IMHO) the dog doesn't search "humans" but "the leader", and if the leader can't help him, that's it -- it's not interested (still IMHO) in humans that it doesn't see as its master.

This should be testable and I'm surprised it apparently wasn't.

[tessierashpool]

    So it could mean, not that dogs "use humans" but
    rather that they recognize humans (and specifically
    their trainer) as the leader of their pack, which
    wolves do not.

I think this is absolutely correct.