John Bradshaw Ph.D.

Pets and Their People

“Lexical Processing” – by Dogs?

Contrary to reports, a new study shows no evidence that dogs comprehend words.

Posted Sep 05, 2016

Language matters.  When we speak of “a word,” we should mean “a separate meaningful element of speech...used with others to form sentences.” (According to the Oxford Dictionary.) Spoken words are, of course, a subset of “sounds” – anything that can be heard.

A research paper currently in press in the August issue of Science purports to demonstrate which parts of their brains dogs use for “lexical processing”; but it does nothing of the kind.  What it does show is that dogs can discriminate certain sounds that they’ve heard before, using one part of their brains, while another part reacts emotionally to the tone of voice used – both of which are such an integral part of routine dog training that neither finding can be considered remarkable.  But that’s all: there’s no evidence that dogs understand “language.” 

If dogs had a concept of “words” they should be able to extract words out of sentences, irrespective of who had uttered them. (The study used only the voice of the dogs’ trainer, and only single words or short phrases such as “Well done!”)  Humans do this remarkably well, and without conscious thought.  On a recent vacation in Devon, I was queuing with a few of my family at a beachside coffee stall when a voice behind us suddenly said “Is John there?”  We all, even I, reflexly half-turned to see who had spoken my name – and then quickly turned back as we realised that it was a complete stranger, initiating a cellphone call to another complete stranger.  Nevertheless, we had all reacted to the word “John,” even though it had been spoken by a voice none of us had ever heard before.  That’s “lexical processing.”  Would a dog whose name was “John” have reacted the same way?  I doubt it. What he (or she) would have heard is “Izjontheyah”, a complex sound (not, for the dog, a sentence) made by an unfamiliar voice, that would have had little or no relevance and would thus have been ignored.

Claims that dogs understand “words” are not new, even to Science, where the standard of refereeing ought to be highest, although on at least one previous occasion the journal has published commentary proposing alternative explanations.  But would Science have been interested in a manuscript entitled “Neural mechanisms for sound processing in dogs”?  It’s hard to escape the conclusion that both canine scientists and journals like to inflate their claims – the scientists get a paper into a highly prestigious journal, the journal gets an excuse for a cute cover picture, and plenty of media coverage (dogs sell papers). 

 used with permission
Source: Alan Peters, from "Dog Sense": used with permission

Is such collusion harmless?  I suggest not.  Dog owners who have read any of the extensive newspaper reports of this study (for example, here, here and here) will have found support for their already entrenched view that their dogs understand "every word they say."  They may then feel more justified in chastising their dog the next time he or she is “disobedient” – which will have come about not because the dog wilfully acted against their instructions (which is how they will see it), but because they had used “words” – sounds – that meant little to the dog.  The well-being of pet dogs depends on their owners understanding what they are capable of, and equally, when our expectations of them are too high.