Language Instincts: Grammar in nature

From November 18, 2006

Many linguists would claim that grammar is what sets human language apart from anything else in the animal world. Some would disagree – bird song, for example, can be quite complex and it is thought that there might be some rules involved in its underlying structure. The question is, at what level of complexity does this ‘grammar’ occur?

A couple of recent studies have examined these claims about animal grammar with respect to communication in monkeys and birds. The interesting thing is that while the monkey researchers claim that their study animals cannot understand complex grammar, the bird researchers claim that their animals can.

First, some grammatical background: the kind of structure we are talking about here is called recursive grammar. This is the ability to insert phrases or clauses within other clauses. For example, we humans can say, “The bird sang from his perch”, or we can go further and say, “The bird, who had just caught a worm, sang from his perch”. We can go further still: “The bird, who had just caught a worm that was wriggling in the dirt, sang from his perch”. It is theoretically possible to keep on adding to a sentence like this forever, and come up with something that is infinitely long (but technically understandable).

In a recent paper in the journal Nature, researchers working with starlings claim to have demonstrated that, much like humans, birds can understand recursive grammar. Their methods involved creating a series of artificial songs following two different patterns: half of the songs had a novel element embedded into the middle of the song, while in the other half this element was added to the beginning or the end of the song. The results were that starlings could eventually learn to distinguish the two song-structure types.

While these results are definitely interesting, they don’t justify any sweeping conclusions about starling grammar (not yet, anyways). The ability to remember and distinguish different song patterns is surely different from the ability to use the patterns for the communication of specific information. The authors of the study have countered that even if the birds are simply using memory to distinguish the song-types, this behaviour is still “remarkable and previously thought beyond the realm of non-human abilities.”

Cotton-top tamarins

Interestingly, a similar study using cotton-top tamarins seems to demonstrate that recursive grammar is beyond the ability of these monkeys. This research involved teaching the monkeys an artificial grammar using recorded sounds, and testing whether or not certain deviations from the learned sound-order captured the monkeys’ attention. Apparently, the monkeys could recognize recordings that violated simple grammatical rules, but they did not respond to recordings that violated recursive grammar.

The monkey study was published in the journal Science, and in the same issue the psychologist David Premack provides several reasons why he thinks animals have not evolved language in the human sense. Premack believes that besides the lack of complex grammar, the lack of teaching, imitation, and voluntary control of sensory-motor systems is what sets animal communication apart from human language. But I’m not so sure that animals like primates and birds lack imitation and teaching. In any case, it would be interesting to know more about the patterns and structures underlying the whole spectrum of animal communication.

Here is a National Geographic article on the cotton-top tamarin study, and a Seed magazine article on starling grammar.

Language Instincts: Do animals lie?


From November 11, 2006

In my last few posts you may have noticed a theme: signals that are used to advertise sex in the animal world are generally thought to be honest ones. In fact, animal communication in general is pretty truthful. There may be different reasons for this: some signals may be impossible to fake (for instance, toad calls may contain honest information about the caller’s size simply because bigger bodies produce lower-frequency sounds). But even when a signal could be faked, the evolution of dishonest signaling is very unlikely. There is a simple reason for this: in the long run it would not benefit receivers to respond to a signal that could be cheated.

This is something that we might find surprising given the amount of deception that goes on in human interactions. Is deception really so rare in animal communication systems? Are there any animals liars?

We have some examples of deceptive communication between different species: for example, ground-nesting birds will fake an injury to draw a predator away from their nest, and some birds in mixed-species flocks will give false alarm calls to increase their own foraging success. Within species, however, the examples of deception are few. We know deceptive communication occurs within a number of primate species. Interestingly, some recent work using ravens has shown that, much like many primates, birds may also be capable of intentionally deceiving conspecifics.

This result came as a bit of an accident during an experimental study on social learning and scrounging in foraging ravens. The researchers provided their ravens with a series of covered plastic boxes that served as food caches (some containing pieces of cheese; some empty). The boxes were arranged in clusters and ravens were videotaped during their foraging explorations. Right from the start, the researchers noticed an interesting pattern between a pair of male ravens: rather than search for his own food, a dominant male relied on a subordinate male’s explorations, following the subordinate male around and eating the food that he discovered.

It eventually became apparent to the researchers that the subordinate raven wasn’t the only one being exploited in this situation. He had developed a strategy to trick his competitor. Whenever the subordinate male found a cluster of boxes containing food, he would quickly move on to a different cluster and start opening boxes there. The dominant male would soon follow, leaving the subordinate free to return to the other boxes and enjoy his snacks at leisure.

The parallels here to primate behaviour are interesting: chimpanzees have been known to walk away from a food site in order to induce other group members to do the same, and then return later to enjoy the food in privacy. Does the ability to communicate deceptively say something special about the cognitive evolution of a species?

You can read the raven study here.

Language Instincts: Human impacts on animal communication

Bird song

From November 5, 2006

As a change of pace from previous posts, I’d like to look into some of the ways that the human-altered environment may be impacting avian communication. Recently, a number of articles have been published that examine the impacts of human activities on bird song.

Some background on bird song

Bird song is a surprisingly complex communication system that we are only just beginning to understand. In most species, male birds use song for two reasons: to defend territory, or to attract a mate. In either case, song is thought to be an honest signal of individual quality, much like sexually-selected plumage.

When defending territory, males will often sing short, simple songs, and face off in vocal battles with rival males. Because singing is costly in terms of energy and time, only a male in good condition can sing loudly and continuously. When attracting a mate, however, males change their tune and sing longer, more complex songs. The function of song in sexual advertisement is very similar to elaborate plumage ornaments in that it signals individual quality to potential mates. For example, in many bird species, females prefer to mate with males who sing more complex songs or have larger repertoires. Experimental work with zebra finches has shown that males subjected to stress during development sing less complex songs later in life than males whose development was trouble-free. Thus, song complexity may be a good signal of genetic quality that females can use when picking a partner.

Bird song and background noise

The acoustic environment in a city is much different than in a natural setting. The background noise in human-altered environments is louder and generally lower in frequency than it is in natural habitat. This presents a problem for animals trying to communicate.

In the last few years, a number of studies have been published showing that birds in cities alter their song characteristics in response to urban environments, for example by singing at a higher frequency to communicate efficiently over low-frequency background noise. This is probably not the result of natural selection favouring birds with innately different songs in urban populations. Instead, it is probably due to birds born in cities learning to sing at a higher frequency in order to communicate effectively in their acoustic environment.

It has been suggested that local changes in bird song may lead to the divergence of communication systems between different populations. Because many female birds choose mates based on song characteristics, this divergence could lead to reproductive isolation between urban birds and surrounding populations. Cities therefore provide a unique opportunity for researches studying bird song in a number of different ways. Understanding how communication is altered in these environments may help us understand the evolution of signal diversity. At the same time, this work could help us understand the best ways to manage urban animal populations.