Wordgames in birdsong
(appeared in July 2015)

(link to main website)

Features of human speech have been detected in the chatter of a boistrous Australian bird, says S.Ananthanarayanan.

That birdsong and some birdcalls have structure which conveys meaning has been demonstrated in a number of studies. It has been shown that combinations of tones, in patterns according to different rules, a sort of grammar,can be distinguished. But there is no evidence of whether bird calls contain units of just a few tones in a specific order, to form what may be considered as ‘words’, with individual meaning. If there are such units, they may be the rudiments from which more complex communication, as we seen in humans, could have evolved.

Sabrina Engesser, Jodie M. S. Crane, James L. Savage, Andrew F. Russell and Simon W. Townsend of the universities of Zurich, Sheffield, Cambridge, Exter and New South Wales describe in the journal, PLOS Biology, their finding that Pomatostomus ruficeps, the chestnut crowned babbler, a bird found in the arid south-east of Australia, uses a simple pair of sounds to convey specific meaning when sounded one way, and another when sounded differently. “This is the first time that the capacity to generate new meaning from rearranging meaningless elements has been shown to exist outside of humans,” says Dr Simon Townsend from the University of Zurich, one of the authors of the paper.

That sequences of a pair of tones that follow a rule of formation can be learnt and distinguished by the European starling has been shown by studies of habituation to a pattern when heard repeatedly. It has even been shown that it is a particular part of the brain, the frontal nidopallium, that is affected when a familiar pattern is changed. Injury to this portion of the brain has been found to affect song recognition in canaries and the zebra finch, which is akin to injury to some parts of the human brain leading to selective language impairment.

Specific patterns of whistles of the male white throated sparrow have been shown to appear at breeding time, both as a courting call addressed to females as well as a warning or a challenge, to males. These sounds are found to affect the female sparrow in parts of the brain that are associated with reward, at a time when she is ready to mate, and the response has been found to be similar to human response to music when the music is appreciated. But these are observations that birdsong is positively a form of communication, rather than evidence of ‘speech-like’organisation that we find in humans.

The PLOS Biology paper notes that human speech is characterised by just a few sounds, which have no meaning by themselves, being arranged in different ways to create the dictionary of words with specific meaning. They cite the example of the three meaningless sounds elements, which are called ‘phonemes’, in the word, ‘cat. These can be rearranged as ‘act’or as ‘tack’ and the difference is recognised as conveying different meaning, and the preffix ‘c’ to the phoneme ‘at’ can be seen as creating meaning not suspected before.Varying the phoneme pattern to create words and then the rules to combine different words leads to the versatility of human language, but there is little known of how the capacity may have evolved, the paper notes.

Animal studies so far have not shown instances of the use of phoneme structure to convey meaning, they observe. There is evidence of a basic layer of syntax in the communication of non-human primates. For instance, certain monkeys produce two calls that announce specific predators, and the meaning of the calls are changed to a general alarm if a particular suffix is added. But as the calls themselves have meaning, and so also the suffix can be considered to have, this is not a case of ‘phoneme’ structure, the authors note. Even in the case of behaviour changes in birds in response to variation in birdsong, the variations are of a higher order structure and there is still no evidence of rearranging the elements of birdsong leading to contextual change in meaning.

Chestnut crowned babbler

The PLOS Biology authors now report an instance of just this in the repertoire of the chestnut crowned babbler, a highly social and cooperatively breeding bird. In earlier studies, the bird has been observed to have at least fifteen context-specific bird calls. Out of this collection, the team has taken up a specific pair, which share constituent elements – a double element call used during flight and a three element call used when encouraging the newly hatched young to feed. The elements of the flight call are denoted as F1 and F2 and the elements of the call at feed time, or the prompt call, as P1, P2 and P3.

An important feature of the five elements involved was that each of them was individually meaningless and represented no context. This said, frequency and volume analysis of two calls established that they were equivalent to the use of two elements only – either as AB, in the flight call or as BAB in the prompt call.

Next, was analysis of the context in which the calls were sounded. It was seen that flight calls arose in 274 out of 450 natural flights and in 58 out of 90 capture induced flights. But there were no prompt calls. Even in flights to and from nests, 62% of the flights had flight calls and 70% of events of with nestlings had prompt calls, and prompt calls during flights or flight calls in nestling provisioning were rare. This establishes that the two-element, AB call was specifically for maintaining group cohesion during flight and the three-element, BAB call was specifically for help in food transfer to offspring birds in the nest.

Response to variations of the elements in the two and three element calls were then investigated with the help of playback of recorded bird calls while the birds were placed in compartments with perches, food, a window and a bird-nest. The calls sounded were two natural calls, two calls where equivalent elements, F1, F2, P2 and P3 were switched and two calls with P1 alone or a three element call with an external element as the first, in place of P1. If the bird recognised the call as a flight call, it would look out of the window, but if it were a prompt call, it may look towards the nest. The source of sound was in the neighbouring compartment and the bird had to look away from the sound source to look out of the compartment or in the opposite direction to look at the nest.

The results were that the birds responded to natural calls just as they did in field trials, which confirms that the two element and the three element calls are distinct. Again, when the equivalent elements in the calls were switched, to test for any accoustic quality of the equivalent elements being responsible for the differences in response, it was found that the response remained essentially the same. The difference in the response to the two element call and the three element call was hence seen to arise only because of the element P1 in the first position in the three element call. The third kind of trial was with the element P1 alone and then a three element call with P1 which was in first position replaced by ‘C’, an external element derived from general chatter of the bird. This third trial was found to increase the ‘looking out’ activity and general movement, which could be attributed to a ‘different sound’, but there was no increase in the nestward looks, which were otherwise stimulated by three element prompt calls.

The P1 element thus appears to be the differentiator, like the ‘c’ in ‘cat’, and this becomes an instance, the first, of an animal using phoneme-like structure to differentiate meaning. ”To our knowledge, this is the first demonstration that animals have the basic capacity to use phoneme-like contrasts to derive qualitatively new meaning,…”, the authors say. The discovery that this can happen in the animal world is significant in understanding the steps through which complex speech as in humans may have evolved. The authors says that “for vocally constrained, highly social species, such as chestnut-crowned babblers, evolving new meaning by rearranging existing sounds offers a faster route to increasing communicative output than evolving new sounds,” which may be the reason that this is the method that allows humans to communicate almost without limit but with an economical font of phonemes to play around with.

Game Changer

One question that arises is why were no trials conducted with other elements, like F1, F2, P2 or P3, in the first position in the three element call. Andrew Russel responded to a query: “The first thing to point is out is that strictly, this is not like “cat” and “at”, but more like “tat” and “at”. This changes the game a bit because there are now less options to try (i.e. tat or att or tta are the only possibilities). Additionally, while att and tta are options, the birds do not use these combinations.” The calls used hence reduce to only AB and BAB, the initial ‘B’ in the second call being the differentiator.