Dog owners (of which I am one) tend to develop a strong sense that their dogs understand what they are saying. Research actually supports this conclusion. In a 2016 study dogs were able to respond to praise words without intonation, to intonation with neutral words, but most strongly to praise words with intonation. Dogs seemed to understand the words themselves – even absent the intonation. This does not mean they understand words the way we do, or that they process language the way humans do. But they can recognize words and associate them with something at least good or bad.

There is also evidence that dogs have more neurons and a somewhat higher neuronal density for a carnivore (although nothing like racoons, which are the real outliers in terms of neuronal density). And there is evidence that dogs evolved to socialize with humans, showing clear differences even from wolves raised with humans. Dogs also evolved greater expressiveness, (like the eyebrow thing that they can do and no wolves), which further supports the notion that their domestication involved them gaining more smarts so that they can socialize and communicate with humans.

Such studies are always just partial slices of reality, and often raise even more questions. One question that a recent study seeks to address is whether or not a dog’s increased ability to parse words is innate or learned, and what are the mechanisms for this. The research paradigm they used was to look at human and dog brains with an EEG while they were listening to fake speech, to see how well their brains synchronized with different parts of speech. Synchronizing means that the EEG shows a repeating rhythm that is the same as a repeating rhythm in the speech. They were specifically interested in consonant bias – humans are the only animals tested so far whose brains respond more to consonants than vowels (not even our closest primate relatives do). This consonant bias likely results from the fact that consonants tend to be more distinctive in parsing words than vowels, so our brains pay attention more to the consonants when trying to discern words. Put another way, there is no reason for a brain to synchronize more with consonants than vowels unless it was trying to parse speech. So it is a very interesting question, whether dog brains show a consonant bias.

What they found was – they do. This adds further evidence that dog brains are actually parsing speech to some degree. Again – this does not mean they understand what words mean or have language like humans. It likely just means they can distinguish words enough to associate them with activities, things, or moods.

The researchers also varied the speech to see how well dog and human brains synchronize with different aspects – whole words, syllables, and phonemes. Humans were able to synchronize with every level, but dogs were only able to synchronize with words and syllables, but not phonemes. This could mean that the smallest unit of words that dog brains are able to parse is a syllable. Not surprising – that they are able to process speech, but not as well as humans.

Another question the researchers tried to address is the degree to which this ability to see patterns in speech is innate or evolved in domesticated dogs. Did they evolve this ability, or is it simply a mammalian brain training on patterns it is exposed to? So they compared dogs raised with humans with dogs that spent the first three months on the street or in a shelter. They found no difference, which they interpret as – there is no critical window of development for gaining this consonant bias, to perhaps it is learned and not innate. However, I honestly don’t think you can make any firm conclusion from this. They may not have hit the developmental window, or the consonant bias may be evolved and baked in and not dependent on such a window.

Keep in mind – the “hardware” vs “software” analogy does not really work with brains. Brains are “wetware” – neurons and their connections, anatomy, and physiology are both the substrate of computation and the information. Further, our genes do not contain a blueprint of a developed brain, only rules for how such a brain develops. Mammalian brains evolved to respond and adapt to the environment and to stimuli. So it is not so easy to separate out an evolved predisposition or greater talent for an ability and the development of that ability through exposure.

What I would be more interested in is a comparison between wolves raised since birth with humans and dogs. If the consonant bias is dominantly determined by exposure to speech, then wolves raised with humans should show the same bias. If it is dominantly an evolved ability, or something that requires both talent and exposure, then they should not.

The consonant bias paradigm is an interesting and useful one, and lends itself to lots of follow up research to further explore this special ability that dogs have to socialize with humans.