We tested this learning hypothesis by manipulating the properties of precursor signals.
On the one hand this training may serve to increase the dogs' attention to the experimenter and thus may confirm the lifelong learning hypothesis, if we find that after such training, the dogs' gaze-following propensity increases.
Short-term training for initiating eye contact (depending on the details and the context of the training) may have a two-fold effect on gaze following: either facilitating it, as proposed by the lifelong learning hypothesis, or hindering it, according to the formal training hypothesis.
The structural learning hypothesis also predicts that, when learning a new transformation, the motor controller should, at least initially, explore preferentially along the structure (thick black line in Fig. 1A) and reduce deviations from the structure.
According to the version of this hypothesis called the language bioprogram hypothesis, which was later revised and became known as the lexical learning hypothesis, children who were exposed to a pidgin at an early age created a creole language by adopting only the vocabularies of the pidgin.
According to the reinforcement learning hypothesis of error processing [19], inspired by earlier work on animals [20], [21], error commission results in decreased activation of the mesencephalic dopamine system.
At this stage, results support the social learning hypothesis.
The configural learning hypothesis predicts that hippocampal damage will spare tasks solved by learning about single cues (elements), but impair tasks that demand the conjunction of cues for their solution (non-elemental or configural learning).
