Anyone whose pet dog has mastered the word “ball” will probably have noticed their pooch’s ears prick up at the sound of similar words, such as “hall” and “fall”. According to a new study in the journal Royal Society Open Science, this is because dogs are unable to process all of the phonetic details of our speech, and are therefore incapable of distinguishing highly similar words. This, say the researchers, could explain why most dogs only ever manage to learn a small number of commands.
Exactly how attuned dogs are to the various phonetic elements of human discourse has been the subject of much debate for a number of years. For instance, a study conducted in 2005 found that canines that had been trained to obey the commands “come” and “sit” were more likely to hesitate if their owner mispronounced these orders by changing either the first letter or the vowel.
To settle the matter, the team behind this latest study sought to determine how the different elements that make up a word are processed by dogs’ brains. They achieved this by taping electrodes to the heads of 17 pet dogs and monitoring their brain activity as they listened to recordings of their owners’ voice.
During the experiment – which was conducted at Eötvös Loránd University in Hungary – dogs listened to four familiar two-syllable commands (in Hungarian), as well as similar but nonsensical words in which one of the vowels or letters had been changed, and completely dissimilar nonsense words that didn’t resemble the original word in any way.
Using electroencephalography (EEG), the researchers found that the dogs were immediately able to distinguish between genuine commands and nonsense words. This was confirmed by the fact that these two types of utterances elicited highly different brain signals – known as event-related potentials (ERPs) – as quickly as 200 milliseconds after the onset of these audible stimuli.
This mirrors activity patterns that are typically seen in the adult human brain, which respond differently to regular and nonsense words within the first few hundred milliseconds.
More interestingly, however, when dogs heard nonsense words that differed from genuine words by only a single speech element, their brains responded in the same way as they did to the real commands. In other words, the dogs’ brains were incapable of recognising subtle differences between highly similar words.
According to the authors, this type of word processing capacity is on a par with that of human babies, which tend to become proficient at distinguishing between the phonetic elements of speech during their second year of life.
“Similarly to the case of human infants, we speculate that the similarity of dogs’ brain activity for instruction words they know and for similar nonsense words reflects not perceptual constraints but attentional and processing biases,” explained study author Attile Andics in a statement.
“Dogs might not attend to all details of speech sound when they listen to words. Further research could reveal whether this could be a reason that incapacitates dogs from acquiring a sizable vocabulary.”