News
Event Calendar
Title Mania®
Cynosport® World Games
Team USA
 
Generalize or Personalize: Do Dogs Transfer an Acquired Rule to Novel Situations and Persons?

Does your dog understand what you say no matter where you say it? By Claudia Bensimoun


Dogs have always been great at understanding communicative signals given by trainers, handlers and their owners. When we think about all the different ways that dogs work with us, and how they have become such an important part of our lives, understanding human hand signals shouldnt be a problem for our canines. 

As reported in Plos-One, Dr. Juliane Kaminski and Dr. Michael Tomasello compared information about the pathways through which information is internalized in both dogs and infants. The researchers described learning as a "generalization of the originally acquired information to novel situations, new objects, new contexts and new places." They compared how long infants took to begin learning how to copy instrumental acts, which was as young as nine months old. By the time a child is 2-to-3 years of age, they'll understand a new game by normative means. ("Normative" means "what every child understands or knows.")

New research by Dr.Kaminksi and Dr. Tomasello questioned how dogs understood rules and the manner they did so. Would this be in a similar manner to human infants (episodic information which only exists in the immediate situation now), or generally, or as normative knowledge? In this study, Dr.Kaminski and Dr. Tomasello researched whether dogs would disregard a cue not to take the treat (1) when the communicator of the ban was present, (2) after a brief absence of the ban communicator and (3) in the presence of a new person. Not surprisingly, it was found that our canine companions tended to retrieve the banned treat more often when the communicator left the room even for a brief moment, and even when a new person entered the room, than when the communicator stayed in the room. 

Summary

These results suggest that our dogs "forget" a rule when the immediate person that gives the cue goes away, and demonstrates the importance of the presence of a demonstrator in modulating a dog's response. These studies by Dr. Kaminski also demonstrate how a dog's behavior will change according to the attentional state or mood of the person. "Dogs apparently did not perceive what they witnessed during demonstrations as being universally applicable. The authors concluded that dogs associate a given piece of information with the person who communicated it," via Plos-One.

According to Kaminski and Tomasello, "Dogs and other animals learn new things by observation and association, nonetheless are able to apply a communicative transmission pathway to transfer a bit of episodic information, which is relevant and important to the current situation," via Plos-One. This new study by Dr. Kaminski and Dr. Tomasello examines how much our canine companions are capable of rule-mediated learning, which predicts a similar performance in different settings, permitting dogs to understand a certain piece of information as a usual norm, and therefore as everyday norms. Read a similar study done by Dr. Ashton and Dr. Lillo at the University of Leicester.

The Study
  
In Kaminski and Tomasello's research that was done in Germany, owners and their dogs attended as volunteers for the study. Thirty nine pet dogs of all different ages and breeds were among those studied, with ages ranging from 8 months to 13 years. Testing was done at the Max Planck Institute for Evolutionary Anthropology in Leipzig. The testing room had a video camera and necessary equipment, which included a Plexiglas wall of 120 cm in height. The Plexiglas wall had a door, which could be opened remotely.

Multi-titled American Bulldog, Payne, demonstrates refusing a treat. Photo courtesy of Karen Mora Wood.  

Experiment

As reported in Plos-One, two plastic dishes were placed diagonally from each on the floor. One was placed 120 cm and the other 350 cm from the Plexiglas door. After luring the dogs, the experimenter positioned herself behind the closest dish, while the dogs stayed behind the Plexiglass.

Procedure

Dogs were all pre-trained and knew how to walk through the Plexiglas door. They also were familiar with the four locations from where food could be obtained and were taught not to eat from the dish nearest to the experimenter. 

During the baiting or luring sessions, the dogs began this by starting on the left-hand side, and the experimenter stood in front of the dog showing the treat. She would then walk to the left hand side and stand behind another dish while calling out to the dog using the word "Watch" The experimenter would then move her gaze from the dog to the dish in which she placed the treat. She then would walk to the right hand side and do this all over again. 

The second experimenter changed tactics and baited dogs at both dishes before leaving the room without giving any feedback to the dogs. In this case the dog was free to choose any treat from either of the two plates, without having any human interference. The dog would then choose a plate, and was then brought back through the Plexiglass. There were also demonstration trials in which experimenters stood behind the dishes, thus claiming ownership. In this case the experimenter would ban the dog from taking the treat by verbal using communication . When the dog chose the other dish that had no experimenter hovering over it, there was no cueing or interference that took place.

Results

Testing results demonstrated that the dogs, in most cases, would be affected by the positioning of the experimenter. Dogs were also much more likely not to choose from any bowl when the experimenter stayed in the room, than when the experimenter switched. 

Age

While the sex of a dog showed no effect on whether it would disobey, the age of a dog had an effect. Older dogs were more prone to choosing the forbidden dish.

Placement of Dishes

Dogs also tended to go for the dishes on the right-hand side more often, than those placed on the left-hand side.

According to Plos-One, both Kaminski and Tomasello had given the dogs a command and wanted to test how prone the dogs were to disobeying that command, when they were near the person giving it. They tested this in two situations: after an interruption of the testing situation and in the presence of a new person.

They found that when a preferred choice was prohibited, dogs tended not to disobey that rule if the experimenter stayed in the room near the dog. They also found that when the experimenters switched rooms (so a new experimenter replaced the one that gave the command), that dogs then disobeyed so much more. 

"Dogs therefore do generalize what they have learned during demonstrations to some extent to the new situation or person." via Plos-One. "Our results show that dogs revoked a rule as soon as the communicator was temporarily absent and inattentive. They only anchored it to the communicator after extensive repetitions. It remains unclear whether an interruption made the dogs regard the rule about not taking the nearby dish as invalid, or whether they misinterpreted the posture of the returning person behind the dish as local enhancement, and therefore a new imperative upon which to act."

Kaminski and Tomasello discuss an earlier study which demonstrates that dogs regard a rule as being valid when in the presence of the person giving that rule. Nonetheless, dogs do not generalize that rule to new people that they meet. Kaminski and Tomasello disagree with these findings. "We conclude that dogs are not able to learn through communicating rules. This is different from young infants, who readily conceive communicated information as conventional and transferable to new contexts and persons, which can be derived from the fact that they correct others who apply an approach that differs from the convention." via Plos-One. They also added that a study by Aston and De Lillo suggested that dogs can understand communicated information and rule mediated learning in a spatial search task, but that this did not replace associatively learned knowledge.

References

www2.le.ac.uk/departments/psychology/ppl/rla6/pdf/JCP2011.pdf

www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0102666

Claudia Bensimoun is a freelance writer in West Palm Beach.



Back



Copyright © 2004-2017. United States Dog Agility Association, Inc. All rights reserved.