Abstract: The ability of horses to learn through social observation may ease the implementation of new management systems, because the use of automatic feeders etc. by naive horses could be facilitated by observation of experienced horses. However, previous studies found no documentation for observational learning abilities in horses. This study aimed to investigate the ability of horses to learn an instrumental task from a familiar conspecific when social interaction was allowed during the demonstration. Two similar experiments were performed. In the first experiment, Observer horses (n=11) participated in ten successive demonstrations, where a trained Demonstrator opened an operant device by pushing a sliding lid aside with the muzzle in order to obtain a food reward. Immediately after the demonstrations the Observer horses were given the opportunity to operate the device alone. Control horses (n=11) were aware that the device contained food but were presented to the operant device without demonstration of the task. The learning criterion was at least two openings. Accomplishment of and latency to accomplish the learning criterion, and investigative behaviour towards the operant device were recorded. Five Observers and one Control, out of the eleven horses in each treatment group, accomplished the learning criterion. Even though this presents a high odds ratio (OR) in favour of the Observer treatment (OR=7.6), there was no significant difference between the treatment groups (P=0.15). Analysis of investigative behaviour showed, however, that the demonstrations increased the motivation of the Observer horses to investigate the device. Subsequently, a similar experiment was performed in a practical setting with 44 test horses (mixed age, gender and breed). We used the same operant device and the same number and type of demonstrations, although the horses were held on a loose rope to minimise aggression. In this second experiment, six of 23 Observer horses and five of 21 Control horses learned the instrumental task, representing no influence of the demonstration. Thus, this study did not demonstrate an ability of horses to learn an instrumental task through observation.

Abstract: Bird song learning is a major model system for the study of learning with many parallels to human language development. In this experiment we examined a critical but poorly understood aspect of song learning: its social context. We compared how much young song sparrows, Melospiza melodia, learned from two kinds of adult `song tutors': one with whom the subject interacted vocally, and one whom the subject only overheard singing with another young bird. We found that although subjects learned from both song models, they learned more than twice as many songs from the overheard tutor. These results provide the first evidence that young birds choose their songs by eavesdropping on interactions, and in some cases may learn more by eavesdropping than by direct interaction.

Abstract: There are numerous reports of novel learned behaviour patterns in animal populations, yet the factors influencing the invention and spread of these innovations remain poorly understood. Here we investigated to what extent the pattern of spread of innovations in captive groups of starlings, Sturnus vulgaris, could be predicted by knowledge of individual and social group variables, including association patterns, social rank orders, measures of neophobia and asocial learning performance. We presented small groups of starlings with a series of novel extractive foraging tasks and recorded the latency for each bird to contact and solve each task, as well as the orders of contacting and solving. We then explored which variables best predicted the observed diffusion patterns. Object neophobia and social rank measures characterized who was the first of the group to contact the novel foraging tasks, and the subsequent spread of contacting tasks was associated with latency to feed in a novel environment. Asocial learning performance, measured in isolation, predicted who was the first solver of the novel foraging tasks in each group. Association patterns did not predict the spread of solving. Contact latency and solving duration were negatively correlated, consistent with social learning underlying the spread of solving. Our findings indicate that we can improve our understanding of the diffusion dynamics of innovations in animal groups by investigating group-dependent and individual variables in combination. We introduce novel methods for exploring predictors of the origin and spread of behavioural innovations that could be widely applied.

Abstract: It is often argued that culture is adaptive because it allows people to acquire useful information without costly learning. In a recent paper Rogers (1989) analyzed a simple mathematical model that showed that this argument is wrong. Here we show that Rogers' result is robust. As long as the only benefit of social learning is that imitators avoid learning costs, social learning does not increase average fitness. However, we also show that social learning can be adaptive if it makes individual learning more accurate or less costly.

Abstract: Horses’ ability to adapt to new environments and to acquire new information plays an important role in handling and training. Social learning in particular would be very adaptive for horses as it enables them to flexibly adjust to new environments. In the context of horse handling, social learning from humans has been rarely investigated but could help to facilitate management practices. We assessed the impact of human demonstration on the spatial problem-solving abilities of horses during a detour task. In this task, a bucket with a food reward was placed behind a double-detour barrier and 16 horses were allocated to two test groups of 8 horses each. One group received a human demonstration of how to solve the spatial task while the other group received no demonstration. We found that horses did not solve the detour task more often or faster with human demonstration. However, both test groups improved rapidly over trials. Our results suggest that horses prefer to use individual rather than social information when solving a spatial problem-solving task