Trained dogs can easily distinguish dozens of different words of human speech. It is always a temptation to believe that they actually understand what these words mean, yet given the nature of their own communication system, the odds seem strongly against it.
Dogs have come to associate certain sounds with certain actions, but those associations are often extremely dependent upon other contextual cues that we may not be aware of. One way to show this is to try giving a dog a familiar command over an intercom. Even a command that the dog is highly motivated to carry out is often ignored unless it is accompanied by some additional cues in our body language.
Indeed, for all of the many continuities that link humans with nonhuman animals, one of the great discontinuities is the way we use language. Human infants, almost as soon as they begin to learn the names of things, take a manifest pleasure in using the name for its own sake. They will point to an object and say what it is – not because they want it, but for no other reason than to share the pleasure of calling the attention of another human mind to it. Even language-trained animals, such as chimpanzees, that have been taught to create “sentences” with computer symbols or sign language expend something very close to 100 percent of their utterances on demands for food, toys, or attention. There is no evidence that they have an independent notion of the symbols as standing for concepts. They have, rather, learned to manipulate series of symbols to get results. Dogs have certainly learned to look at us, or come, when we speak their name, but there is not a scrap of evidence that they grasp the notion that their name is their name, in the sense that it stands for or represents them.
Given all that, however, it certainly seems odd that dogs can distinguish words in human language. Studies by Russian speech scientists found that dogs can readily be trained to distinguish the vowels a and i produced by an audio synthesizer; even when the base pitch of the vowels was changed, the dogs had no trouble telling the two apart.
Dogs may often be confused by substitutions of one consonant for another – try saying “Fly clown” instead of “Lie down,” and your dog will probably react exactly the same. But the ability to distinguish vowels depends on rather precise analysis or the higher-pitched resonances that accompany their base pitch. Dogs do not utter vowels themselves; why should they be able to tell them apart when we say them?
The simple and general explanation for this happy circumstance is that ears are older than speech. Mammalian ears have been around for tens of millions of years, and the ears of all mammals have much in common. Human speech, however, has been around for only 100,000 years or so, and the human vocal tract is a unique and late development. Only humans possess the vocal apparatus needed to generate the sounds of speech.