With the same mental constructs, by which a dog can come to understand the semantic register of certain commands, a dog can also begin to learn to formulate its own commands or requests by emulating or creating their own expressive constructions of logic patterns, i.e. having learned that before they are taken for a walk, their owner reaches for the leash, they begin going near the leash when they would like to be taken for a walk.
This would not be enough to suggest that dogs have a functional theory of mind. A parrot will learn to reproduce words on command, if doing so will guarantee it a treat. A dog is different in the regard that it can choose to express something rather than something else, and can likewise often choose not to express things.
An example: A dog has destroyed some papers while its owner was away and, knowing that the owner would be angry with this and would likely punish the dog, the dog hides the destroyed papers. The dog has chosen not to express, or reveal, something to its owner. Dog has inferred a certain mental process of its owner and acted according to those inferred mental constructs to behave manipulatively.
The primordial utterance of a basic theory of mind seems then to be a game of deception. Deception is a somewhat obvious marker of concretized interiority--and one might argue that it is overdetermining the machine to draw a conclusion on its nature from a single aspect of its activity--but this machine's activity, even in its most sophisticated constructs, retains the residual essence of this raw utterance of deception, duplicity, or expressed distance between one's logic and the logic behind one's logic.
Hermes Mercurius, messenger god, node of communicants, was equally patron to poets, liars, thieves and travelers. Geographic distance, as experienced by a traveler, entailed the result of being a stranger to others--like the bandit, the liar or the poet, one was always at a remove. Distance and dissimulation are the shared essence that brings these aspects into compartment. Some men, Villon or Leadbelly for example, contained all these compartments in a poetic equipoise.
Having recently concluded these things in regard to my dog, I have been interested in our shared and unsharable communication signals. There have been moments in which we are staring at each other, I trying to express something--to teach her to get the ball or the toy--, and she trying to understand but not being able to; moments that break into anxious frustration for both parties.
Knowing that she knows I have a knowledge system to impart to her, she wants to get it, to get the interior--to get behind and within my concept constructs--but the closedness of my communication does permit her to enter. This is very frustrating to her, because she wants to learn. When she learns something, she is happy to emulate that understanding to re-engage a communication point--goes to the leash, etc. When she has not learned something, she stares until she realizes she cannot get it and jitters and barks or breaks off.
In this regard, the dog is rather human. It is human to not understand, if we live with a communicative standard which stems from a theory of communication that is fundamentally deceptive, duplicitous, playfully hidden and hiding. We are invariably often the dog, staring stupidly, trying desperately to understand a foreign communicant, inscribed in the language of a stranger. Such is the difficult state, I have found, of reading and trying to 'understand' a poem. Most of the time we are either staring stupidly, trying to get it, if not completely distracted by the possibility of something else.