Researchers, led by Emile van der Zee, Ph.D.,of the University of Lincoln School of Psychology in Lincoln, England, have learned that dogs do not associate words to objects in the same way that people do.
Van der Zee, a cognitive psychology specialist, and his colleagues studied a single Border Collie named Gable, coming to the conclusion that “shape bias for word generalization in humans is due to the distinct evolutionary history of the human sensory system for object identification,” according to the study’s abstract, published in the peer-reviewed online journal PLOS ONE in November 2012. In other words, our evolutionary history defines how we use our senses to identify various objects.
From the onset of word learning, young children generalize names to new objects on the basis of shape and continue to do so as adults – a tendency known as “shape bias.” This is crucial to language development because it enables children to assign new objects to pre-established classes, according to the university. For example, kids know that a tennis ball and a football both belong to the category, “ball,” even though they have very different shapes.
The Lincoln researchers found that when dogs are introduced to new words to refer to new objects, they first generalize based on object size, then on object texture. Unlike humans, they do not appear to naturally discriminate based on shape.
“A number of recent studies have suggested that the domestic dog’s word comprehension is human-like,” van der Zee says. “Arguments have been made to refute this claim, but until now there has been no clear empirical evidence to resolve the debate. Our findings bring a fundamental new insight into this discussion and add to our understanding of the cognitive equipment necessary for true human word learning.”
For this research, Gable was challenged with four different tests.
The first was to confirm that he actually knew the names of 54 objects, as claimed by his owner. According to the paper written by the researchers about the project, the 54 toys were presented to Gable 10 at a time, 26 times. He was asked to retrieve two toys from each set, except with two sets when the request was for just one item. No more than 10 sets were presented on a single day, and he had a break after being tested with five sets. The 10 toys were arranged in a semi-circle about 12 inches from one another, in an area where Gable could not see his owner or the researchers.
Gable correctly retrieved 43 of the 54 objects, demonstrating to the researchers that he “reliably knew the words for 43 objects.” When requested to retrieve the other 11 objects whose names he supposedly knew, he retrieved the wrong object, or hesitated, whined or seemed to wait for further instructions.
The second experiment was designed to replicate a 1988 study on the importance of shape in the early learning of words, which found that word generalization is based on object shape in both young children and adults.
Variously shaped and sized objects, made of foam and covered in three different fabrics, were used. After being taught the word “dax” for one of the objects, Gable was able to link the word to other DAX-sized objects in 10 out of 10 cases in which he was given the choice between a DAX-sized object and a larger object. This confirmed that he “generalized the meaning of the word ‘dax’ to other objects that were of the same size as the DAX object, irrespective of their shape or texture,” according to the researchers. “In those cases in which Gable was able to make a choice between DAX-textured objects and objects with a different texture, and DAX-shaped objects and objects with a different shape, he did not appear to have a preference for objects that had the same texture as the standard DAX object or objects that had the same shape as the DAX object as is the case in humans.”
The third experiment tested whether the size bias found in experiment 2 was based on Gable preferring the smallest object, possibly because smaller objects are easier to pick up, or whether the size bias was based on generalizing word meaning to objects of the same size. Using the word “gnark” for this experiment, researchers determined that the dog had no preference for smaller shapes.
The purpose of the final experiment, undertaken four months after experiment 3, was to determine whether longer exposure to the “dax” object would affect his word associations. The researchers sent him home with the object for 39 days. His owner taught Gable the link between the object and the word “dax” every day in different contexts by putting the object with different sets of toys, then saying, “Get the dax.”
During subsequent testing, Gable chose the DAX object in six out of six trials when it was placed with nine or 10 of his toys. This demonstrated that he had learned to distinguish that object from his toysover the 39 days. However, in this experiment, Gable tended to generalize his word knowledge to objects of the same texture, but not to objects of the same size or shape. He could not “reliably distinguish between a DAX and other artificial objects that only differed in size, texture or shape,” as 2- and 3-year-old children can. “These results show that after a long ‘dax’ word-DAX object familiarization period, word generalization was qualitatively different for Gable compared to a short familiarization period, but also compared to word generalization in humans,” according to the paper.
The researchers concluded that the mental lexicon – the long-term mental store containing sound-to-meaning mappings – appears to be fundamentally different in dogs and humans, both in terms of how it is built and in how it operates.
Van der Zee added: “This would suggest that an important factor in the natural structuring of the mental lexicon may be the way in which sensory information is organized in a particular species. The human visual system is tuned to detect object shape for the purpose of object recognition. In our experiments, we excluded Gable using scent cues. It seems that his visual system and sensory cues linked to his mouth region are focused not on shape, but on size and texture. Only future experiments will reveal what role scent plays for the dog in generalizing words. It is only by comparing other species with humans that we can find out more about the neural and genetic foundations of word reference in language.”
The researchers’ findingsmay help to advance understanding of the foundations of language in humans and the critical differences with other species, according to the university. In addition, awareness of the absence of shape bias in dogs may help improve training programs for pets, working dogs and assistance animals.
This article was prepared with materials provided by the University of Lincoln and study findings published by PLOS ONE.