8

Social and vocal learning

Learning ability per se can contribute little to a discussion on advanced cognition in birds because it takes place in most, if not all, organisms. Even unicellular organisms can learn, invertebrates learn (Dawson et al. 2013), fish learn (Brannas 2014; van Bergen et al. 2004) and so do all vertebrates. What is remarkable about this is not that organisms are able to learn but that, for a long time, humans staunchly refused to believe that animals could. A few notable exceptions aside, they held onto this belief well into the middle of the 20th century. This had some tragic consequences when the first serious attempts were made in conservation circles to hand raise and reintroduce species into the wild. Most of those early efforts failed dismally because the programs had failed to teach the animals even the most basic skills needed for survival, presumably believing that these would somehow be automatically accessible to the naive animals.

Ways of learning

Learning has been subdivided into many different types with varying cognitive demands. Social and song learning are types of learning with the potential of making special cognitive demands on the learner. Experienced adults or parent birds can guide an inexperienced one into what to eat and where to find it, where to live or from what to escape (Danchin et al. 2004). This is called social learning. Such social learning is widespread in the animal kingdom, from insects to mammals (Leadbeater and Chittka 2007). It can be defined as the use of social cues, often inadvertently left by other animals engaged in making choices between various options (Grüter et al. 2010).

Social learning in birds is not only concerned with song but with other types of learning. While it is important to know how song is learned, this emphasis ignores the importance of social learning in different (non-reproductive) contexts (Hausberger et al. 2005) and the possibility of different underlying mechanisms of brain activity. Indeed, many factors may determine how long learning takes and how strong a memory will be formed (Clayton and Soha 1999). Age is important and so is the outcome of the event, such as approval or punishment. Research on social learning in animals has been extensive but weighted very much in favour of mammals, and primates in particular (Box 1999). In birds, studies have demonstrated clearly that they have learned some specific behaviour or problem-solving task from a parent or a conspecific.

Observational learning is a more familiar category. For instance, common ravens will learn to pull on a flap to slide open the lid of a box containing food by watching another raven perform this task (Fritz and Kotrschal 1999) and New Zealand keas can perform equally well on the same task after observing a conspecific perform it (Huber et al. 2001). Budgerigars have no difficulties in learning by observation (Dawson and Foss 1965). The European blackbird passes on information about predators by social learning (Curio et al. 1978; Curio 1988). This has been demonstrated very effectively. If one bird sees a stuffed owl it will mob it and this will stimulate a naïve bird in a cage nearby to mob also, and so to learn to mob the owl or whatever else it sees at the time. These are now very familiar insights into bird behaviour.

image

Fig. 8.1. Brood parasite and host. (Left) A Horsfield’s bronze-cuckoo; (right) the smaller superb fairy-wren male in full breeding plumage (white indicates bright blue plumage). This cuckoo lays its eggs preferably in superb fairy-wren nests. (Image on the right by Peter Storer.)

Two of the rare studies of social learning in an Australian bird concerned the way in which the superb fairy-wren deals with brood parasitism (Langmore et al. 2012; Feeney et al. 2013). Superb fairy-wrens do not always recognise their brood parasites, and play primary hosts to the Horsfield's bronze-cuckoo (Fig. 8.1) and secondary hosts to the shining bronze cuckoo but naïve birds can be taught to recognise these parasitic cuckoos. The researchers used models of the shining bronze-cuckoo and, as a control, the white-plumed honeyeater. The honeyeater is a harmless species and lacks the dramatic markings of the shining bronze cuckoo with its barred front and belly feathers, but both species are of similar size and occur locally in the study sites. Young fairy-wrens showed no response to the honeyeater, as was predicted (functioning as a control) and at first no response to model cuckoos, but there were strong responses from experienced birds.

Naïve young birds were able to learn from watching the behavioural responses of experienced adults to the model and then proceeded to mob and alarm call when a model of the cuckoo was presented. These results showed that such recognition had to be learned. Langmore and colleagues had first studied the behavioural responses in sites that they knew were heavily parasitised and in others that were rarely parasitised by cuckoos (Langmore et al. 2012; Feeney and Langmore 2013; Feeney et al. 2013). So why did one group suffer from substantial invasion by cuckoos and the other did not? Because the sites were relatively close to each other geographically, genetic adaptation could be safely ruled out. A cuckoo might have succeeded at several points of infiltration. Some of the typical strategies are depositing an egg unseen, depositing an egg that was of similar size and shape, having a cuckoo hatch and dispose of the host chicks by pushing them out of the nest or even by killing them. By recognising the danger points and, preferably, by preventing an egg being laid in the first place, the fairy-wrens can be assured to raise their own brood. The researchers concluded that the ability of one group to successfully avoid being parasitised had to have been achieved by learning and, more broadly, by social transmission (Davies and Welbergen 2009). The studies on superb fairy-wrens demonstrate very nicely that social learning may provide knowledge to one group but not necessarily to the species as a whole.

Another recent and fascinating study, also using the superb fairy-wren, has been conducted by Kleindorfer and colleagues of Flinders University. They ran some cross-fostering experiments and discovered an incubation call – a call the hen actively directs at the eggs – and the newly hatched offspring had elements of that incubation call in their begging call (Colombelli-Négrel et al. 2012). The researchers called this a parent-specific ‘password’. Indeed, parents can distinguish between these calls and those from other broods. They can identify the calls that were learned when the young were still embryos and that shaped call similarity between the hen and her own young, thereby enabling detection of foreign cuckoo nestlings by their calls. It was found that the vocal signature of cuckoos was well known and recognised by superb fairy-wrens and calls were accurately assessed according to risk. Langmore (2013) was right when she argued that members of the Maluridae (fairy-wren) family exhibit arguably the most advanced portfolio of defences against cuckoos of any avian brood parasite host yet discovered. It is worth repeating here that fairy-wrens are cooperative breeders and one could speculate that the combined cognitive efforts of a close-knit group may have an advantage over that of pairs in finding solutions to intractable problems such as cuckoos.

A very nice, but unfortunately brief paper, also convincingly showed that being a helper at a nest even in cooperative species is not an ability that is inherited. It also has to be learned. Brown and Brown (1982) described in detail how yellow-rumped thornbill juveniles try to help, but fail, by making a number of mistakes in collection or delivery of the food item. When the adult was in the nest, she took it from the juvenile and over a period of weeks, the juvenile eventually became very proficient in the task of feeding nestlings by watching the adult perform.

Information about song learning is plentiful and we have a great deal of information about how birds learn song and how long they take to learn it. The zebra finch has been used as the model for song learning (more of this later) and, apart from research on the songs of butcherbirds and lyrebirds, extensive research exists on the Australian magpie and a smattering of other Australian species.

Much of what we know about vocal development concerns species that are dimorphic: that is, male and female of a species are visibly different in plumage and/or song is only sung by males only (both apply in zebra finches).

However, quite a few Australian birds are monomorphic, meaning that males and females look alike (their plumage looks very similar) and there may be no difference in their songs and song may not be used for reproductive reasons but only in a territorial context. Of course, there could be slight differences, such as eye colour or size or small differences in plumage patterns that, in some cases, may be so slight that we miss them. However, birds have evolved their own insurance system to avoid mistaken identities.

Aural discrimination

Indeed, some species spend considerable effort on vocal identification. Songbirds have special neurons and neuronal clusters in the forebrain that are designed to process various sounds and also auditory feedback mechanisms that ensure that sounds can be produced to exact specifications needed. Galah parents (very similar in plumage and size and different only in eye colour), for instance, devote considerable time in getting their youngsters to be vocally responsive. They simply do not feed their young until the young emit a special vocal signal first, training their young to do so on demand and making them listen to their parents’ low intensity calls (Rowley 1980). The reasons for such detailed preparation become obvious when the social context is considered. Galahs, as mentioned before, have a crèche system with an adult supervising. The fledglings get dropped off among a group of youngsters about the same age while the adults go off feeding. As also in colonially nesting species, making sure that one picks up one’s own offspring is achieved by voice check. It is so personal and so specific that, even in large and noisy groups, parents and offspring successfully recognise each other but such recognition needs to be learnt. Interestingly, in crows it was shown that group members were successfully recognised whether visually or by sound but such recognition did not extend to non-group members (Kondo et al. 2012) and such recognition thus plays a vital role in survival of offspring, as well as in group cohesion.

Tagging individual identity via the vocal signals is a dynamic process and the original vocal identity can be altered as circumstance may demand. For instance, when pairs of budgerigars (and many other parrots) first get established, their own vocal expressions start to converge, creating unique vocal identification of the pair bond (Hile et al. 2000). Even same-sex groups of budgerigars adjust their vocal utterances to each other (Hile and Striedter 2000). Discrimination of sounds is thus of crucial importance for both songbirds and parrots.

This personal voice ID can be taken to sophisticated levels. In groups of cooperatively breeding bell miners, complex kin and non-kin membership, as well as individual identity, are established by vocal means. This has a downside for some nestlings in that helpers tend to bias their provisioning in favour of close kin (McDonald and Wright 2011). A close relative of the bell miners, the noisy miner, also breeds cooperatively and is at least equally proficient in aural discrimination. Noisy miners are able to distinguish between individuals belonging to the group at large by vocal cues alone, even those that are entirely unfamiliar to them (McDonald 2012), and may belong to various cooperative alliances. Hence, for feeding, care and general social alliance, it is of paramount importance for hatchlings to acquire the skill to recognise voices belonging to parents, immediate helpers or kin and to do so rapidly.

Vocal learning

Of course, recognising sounds is not the same as learning to produce them. The former is based on having the perceptual ability to hear, on appropriate auditory exposure, and then on memory formation. Budgerigars have been tested and found to have excellent hearing (Farabaugh et al. 1998), fine powers of discrimination (Dooling et al. 1995) and the ability to commit a mate’s call to long-term memory (Kanesada et al. 2006): in themselves remarkable cognitive abilities.

Learning new sounds is a rather rare gift in the animal kingdom. Apart from humans, only birds, cetaceans, bats (Pettigrew 1986; Janik and Slater 2000) and elephants (Poole et al. 2005; Stoeger et al. 2012) have the brain structures to do so. Among birds, the ability is limited to songbirds, parrots and some hummingbirds (Zeigler and Marler 2008): groups that are not closely related taxonomically (Sibley and Ahlquist 1990). The astounding part to early researchers was that birds are the only ones achieving such feats of vocal production without a neocortex, once thought to be an indispensable precondition for language and cognition. However, we now know that there are analogues between the neural structures in the avian and mammalian brain with many of the same function: in particular, those known to involve higher cognition (Jarvis et al. 2005). Songbirds and humans have specific vocal pathways (Fig. 8.2) not found in those birds and mammals that do not learn their vocalisations.

Song of Australian songbirds remains one of the seriously understudied areas. The song of the Australian brown thornbill, the grey shrike-thrush and its extraordinary purely tonal and varied song and those of the noisy friarbirds and the whistlers, for example, have received little research attention, even though these species produce magnificent songs. The only songs that have been described and analysed in great depth are those of the zebra finch: in this respect probably the most studied bird ever (Arnold et al. 1976; Konishi and Akutagawa 1985; Margoliash and Fortune 1992; Vicario and Yohay 1993). Song in zebra finch males is acquired via aggressive tutoring, usually by the father, and this has been shown to lead to better learning (Weary and Krebs 1987; Eales 1989). Because females actively select males with long repertoires, and higher frequencies (Balzer and Williams 1998), the parental investment in tutoring his offspring is well worth it. We also have careful analyses of song of the lyrebird (Powys 1995; Robinson and Curtis 1996; Kaplan 2000), the pied butcherbird (Johnson 2003; Taylor 2010) and the Australian magpie (Brown and Farabaugh 1991; Farabaugh et al. 1988; Kaplan 2005), and some detailed records exist of various vocalisations of honeyeaters and corvids (Jurisevic and Sanderson 1994b). In the lyrebird, as in the zebra finch, song is learned only by the male to attract a female. Unlike the zebra finch, the song of lyrebirds is one of the most elaborate song performances known and it is famous for its spectacular mimicry (Robinson and Curtis 1996), as was discussed in more detail in Chapter 7. The territorial song, performed by both males and females, may be based on a genetically encoded template, but the male breeding song develops in response to sexual competition and it involves learning.

image

Fig. 8.2. Pathways for vocal learning. (A) Lateral view of the song control system of the avian brain; (B) the exterior view of the brain of an adult Australian magpie. The shaded area is the forebrain (called the pallium). Diagram A shows the nuclei involved in song (longitudinal section). This is a simplified schematic version of the song control system (after Doupe and Konishi 1991) in the brain of the adult male zebra finch. The ellipses represent nuclei and the arrows are descending motor pathways (HVC; RA, robust nucleus of the archistriatum; nXIIts, tracheosyringeal portion of the hypoglossal nucleus). The arrows also identify a second pathway to link HVC and RA: from HVC to Area X to DLM (the medial nucleus of the dorsolateral thalamus) that we also found in our research on the magpie brain (Deng et al. 2001) then to L-MAN (lateral magnocellular nucleus of the anterior neostriatum) and from there to RA. Field L is the primary avian auditory area that projects to RA and HVC. There is also a link from L-MAN to Area X showing the auditory feedback mechanisms from input (L) to output, via two pathways, involving direct links to RA and HVC.

There are many extraordinary aspects of song in Australian birds that seem noticeably different from the sounds that beautiful songbirds in Europe, such as the blackbird and the nightingale, produce. To my ears, European birds sound like the folksongs of western Europe – in western major scales (the melodic and harmonic diatonic and minor scales), consisting of whole and half notes – in other words, based on the scale on which compositions since the Renaissance have relied and, in birds, as in folksongs, they seem largely in C or F major with some occasional concession for minor insertions.

By contrast, so many songs of Australian species sound like technobabble products, synthesizer and computer renditions of percussion instruments or are very cultivated carefully protracted tunes in the chromatic scale or even pentatonic scale. Some of the finer points in butcherbird song have been studied in detail by Hollis Taylor (2008, 2010). The percussionist elements and rhythmic sequences in drongo and bowerbird vocalisations, or the songs in chromatic scale, perfected in noisy friarbirds, have not yet been described and barely even mentioned. Indeed, much of drongo, bowerbird or dollarbird vocalisation (HANZAB 1999) consists largely of rhythm and percussion, reminiscent of Balinese music, especially of something like the monkey or ‘Kecak’ dance (pronounced ‘ket-chuk’), that consists of very complex rhythmic features but has no discernible melody line. Magpie warble – the beautiful tuneful songs that may be sung for hours at a time – has at times been described as similar to synthesiser music and so is that of the black-faced cuckoo-shrike. If one adds the staccato ‘laughs’ of the kookaburra, the haunting crescendo of the common koel and the weird cat-like mews of the catbird (a close relative of the bowerbird but sounding like a cat), there is an almost haunting quality to them. Indeed, the 1942 Hollywood film of Rudyard Kipling’s Jungle Book, produced in 1942 by Zoltan and Alexander Kora had in its sound track the calls of kookaburras when a particularly eerie night scene was shown. Given that the film was set in India, this is taking some licence because the kookaburras are unique to Australia and they do not sing at night. But it was a good sound effect.

These sounds make the soundscape of Australian birds and, with closed eyes, one would instantly recognise them as distinctly Australian and one day we may be able to decipher and describe in detail not only what all these sounds may mean and in what context they are used, but how repertoire is related to the environment, to brain size and cognition. The latter is no longer entirely futuristic and speculative (Moore et al. 2011).

In Australian songbirds, there is a wealth of material yet to be discovered about learning of song, song size repertoire, memory and even function. At this moment, we only know that there are many species with extensive repertoires that have not been studied at all to date. Generally we do not know what their learning period and their learning style is, or in how many Australian songbird species song is shared by males and females alike and what function it might have. It would also be important to learn whether any of these magnificent songbirds are lifelong learners or have other vocal capabilities as yet untested. There are several generations of research work waiting to happen.

Although we have excellent records, description and analysis of distress calls (Jurisevic and Sanderson 1998), and of alarm calls (Jurisevic and Sanderson 1994a; Kaplan et al. 2009) in Australian birds, we also know relatively little about their song development, whether it is fixed, how long it takes to learn, whether it changes throughout a lifetime or expands. It is therefore very difficult to judge how much learning needs to occur in various species to develop full song.

In the Australian zebra finch, there is a so-called ‘sensitive’ period for just some months: a brief window of opportunity in which the male juvenile can learn and remember the song he will have to sing on his own in the following breeding season, by then a fixed or ‘crystallised’ song (Marler and Peters 1982, 1988). The size of the HVC is larger in males that have larger repertoires (Airey et al. 2000). Once the breeding season is over, neurones are lost from the song nuclei and they shrink in size (Arnold et al. 1986). That learning is involved in the male’s song performance as a forebrain activity was very well demonstrated in a study by Simpson and Vicario (1990) in one specific long call that male and female zebra finches share. Their study showed that two paths for vocalisations may coexist: one activated in the midbrain (inherited) and another when acquiring additional vocalisation via learning, activated and achieved only by the involvement of the forebrain (learnt).

Gobes and Bolhuis (2007) showed that tutored-song memory and a motor program for the bird’s own song have separate neural representations in the songbird brain. Both in humans and songbirds, the cognitive systems of vocal production and the auditory memory are processed in different brain regions.

Other avian species have extended learning periods and some of them may be taught while others improvise. Finally, there are birds that are lifelong learners (also called ‘open’ learners), and these include parrots and magpies, meaning that learning is not restricted to the juvenile period and can occur at any time during adulthood (referred to as brain plasticity). There are no correlational studies yet to tell us whether the ability for lifelong learning is linked to brain size but one would suspect that there might be a link because all cockatoos and budgerigars (although not even songbirds), for instance, are lifelong learners and they feature among the birds with the largest brain to body weight ratios, although current knowledge suggests that magpies and butcherbirds have a rather more average brain size ratio, despite their versatility.

Repertoire size may also tell us something about the brain. Magpies have very large repertoires as have lyrebirds. This would require substantial memory formation and, in some cases, may reveal very complex behaviour, such as intentional communication as has been studied so far only in Australian magpies in the wild (Kaplan 2005) and, in the laboratory, only in domestic chickens (Evans 1997). However, as mentioned above, in 2011, Moore and colleagues published the results of a paper that examined the song control system (Fig. 8.2) and compared it to song repertoire size of 49 species from three continents. They found a clear relationship not in overall size of the song control nuclei but in relative size of the HVC to RA, and concluded that repertoire size is more accurately predicted by the number of neurons in higher motor areas relative to that in their downstream targets than by the overall number of neurons in the song motor pathway. This is an important finding because it argues that it is not necessarily an increase in overall brain size that allows a particular behaviour to occur (i.e. an increase in forebrain activity), but it depends on the distribution and allocation of neurons. Hence, a well-endowed HVC, supported by another area (RA), together make large song repertoire (and a memory of it) possible.

Song learning models and their inadequacy for some Australian species

Studies in sexually dimorphic avian species have shown a strong association between development, learning and function of song (and the manner in which it is transmitted) but how much of it applies to Australian species? What we know, and the literature is extensive, holds for sexually dimorphic and also migratory birds in high-latitude species of the northern hemisphere (Brown and Farabaugh 1991; Slater and Mann 2004) but it may be wise not to treat the theoretical models that we do have as being universal. As Kroodsma reminded us, this unitary way of viewing the high latitude as the standard for all bird behaviour may have led to the neglect of important topics (Kroodsma 1996).

In particular, rather little attention has been paid to the structure of the song control system in birds in which both the male and female sing at similar rates and share other tasks almost equally (Kroodsma et al. 1996).

Many of the Australian species that can learn their vocalisations and are lifelong learners also have monomorphic vocal habits and may have similar monomorphic physical traits and behaviour. Indeed, Australian species often seem rather ‘egalitarian’, to borrow a socio-political term. Although one needs to add that monomorphism may not always mean equality. Especially in groups (i.e. containing more individuals than in a pair), there is often a very strict hierarchy established. Yet there is often no difference in the song between male and female, and no marked difference in plumage or size, brooding and feeding of youngsters and defence. Many pair-bonded species brood, nest, feed and protect their offspring equally or, at least, even if some role division is apparent, share the overall workload in raising offspring. In tawny frogmouths, as I described elsewhere in detail (Kaplan 2007a), the pair has a strict roster system for sharing in brooding. The female broods during the night, the male during the day.

As I have written elsewhere (Kaplan 2008c), the Australian magpie is a case in point that the high-latitude song functions simply do not fit well in all cases. Presumably this also means that vocal and social rules vary with different functions. In our laboratory, we started our 25-year research involvement with magpies in the early 1990s in order to learn what the species can do vocally and cognitively. We first began to refute one of the queries as to whether or not males and females both sing and whether they do so to the same extent. I recorded nearly 10 000 magpie vocalisation sequences from males and females, juveniles and nestlings and then began the arduous task of deciphering the species’ vocal development, song function and diversity, also according to sex and age (Kaplan 2005). The neuroanatomy of their brain was examined with the particular interest in identifying all the regions of the brain known to be vital for song production in other species. The most essential one, as mentioned before, is the HVC and there is a cluster of other nuclei located in the forebrain (Fig. 8.2). We were able to show (Deng et al. 2001) that, indeed, both males and females have the same regions and particularly a well-developed HVC.

We then proceeded to ask how sound is produced, and found that magpies have the most complex sound production system so far known by being able to produce sounds individually and separately from either side of the syrinx (the sound-producing organ in birds, located in the chest well below the larynx). The experiments we reported (Suthers et al. 2011) provided a first examination of the peripheral vocal mechanisms that underlie the exceptional vocal versatility of the Australian magpie and the bird’s ability to achieve precise motor coordination between sides in order to produce very rapid amplitude modulation (Fig. 8.3). The result demonstrated bilaterally independent control of both the timing of phonation and the frequency composition of sounds generated on each side of their syrinx. These mechanisms also partially explain why magpies (both male and female) can warble with little interruption and seemingly little energy expenditure for many hours at a time (Suthers et al. 2011). The study was exciting because it showed that the method of song production makes the magpie one of the technically most advanced songbirds giving it great virtuosity. Yet the elaborate song of magpies is not part of reproduction and it is also not part of territorial defence, the other major function of song identified in high altitude species. For this, magpies use a specific call type referred to as carolling, but the ‘song’ – those beautiful and tuneful warbles that can be heard in backyards and in the bush for hours on end – do not feature in territorial maintenance.

image

Fig. 8.3. The breathing and sound apparatus of an adult magpie male. The windpipe (trachea) bifurcates into the lungs (the large two lighter flaps in the middle of the chest). The white dot indicates the position of the syrinx: at the end of the trachea (windpipe) the main sound-producing organ in birds with the same function as the human larynx. The syrinx is located at the point where the windpipe splits into the two branches. The diagram on the top left represents details of the syrinx, providing an indication of the structure of the syrinx at the point of the bronchial bifurcation. Importantly, the syringeal muscles on either side of the syrinx (as indicated by arrows) define a songbird. A bird is classified as a true songbird (oscine) if it has four or more pairs of syringeal muscles. The membranes produce the sound when air is exhaled and the syringeal muscles contract. On the right is an image taken of the actual view of the trachea ending in the syrinx of a magpie – note that the heart shape at the bottom (white arrow) comprises the muscles packed around the cartilage and the syrinx.

In sexually dimorphic species, singing increases markedly during the breeding season. By contrast, song in both male and female magpies declines sharply during the breeding season (Kaplan 2008). In fact, many Australian bird species fall silent during the nesting period.

There are thus a few ‘irregularities’ here in terms of theories of function of song. Here is a bird exceptionally endowed for song and yet so much of what is produced seems to have no easily identifiable function. Moreover, magpie juveniles practise song only in the absence of parents, not in their presence (Kaplan 2005). They are improvisers. In terms of versatility of vocal communication, in magpies there is plenty to learn: some 27 variants of alarm calls alone, including those specific to birds of prey alone, apart from a vast repertoire (Kaplan et al. 2009).

Importantly, magpies as lifelong learners display many of the identifiers in life history associated with complex cognition: brain plasticity and a long life, including a long period of dependency staying on as helpers not just for a year but for 4 to 5 years (Veltman and Carrick 1990). This very long gap between juvenile experience (under parental supervision) and reproduction is a feature in life histories that has been linked to complex cognition. Notably, similar credentials can be produced for the white-winged chough in terms of age of juvenile dependency and breeding. Juveniles take 4 years to mature sexually and always remain in the natal group for this length of time (Heinsohn 1991b). Hence, vocal development in both birds and mammals is more than merely a passive process of maturation of motor and memory abilities (Goldstein et al. 2003).

Monomorphism, social brains and complex cognition?

Perhaps it is not just a matter of social bonding and group formation that creates complex cognition, but it would be important to assess whether the partners may bring near equal stocks of knowledge and ability to their bond (i.e. such as similar or same range of vocalisations, sharing of raising of offspring). Some of the qualities come to the fore in Australian species that either occasionally or always cooperate (Cockburn 1996) and in many of the species that are pair bonded for life.

Here is perhaps a good time to reinforce a point first made in the preface relating to theories developed for high-latitudes birds being somehow too readily applied universally. One of the important topics in evolutionary biology is sexual selection (male competition and female choice). It is one of the important and unshakeable Darwinian principles of how evolution has occurred. However, female-only parenting is rare in Australian birds and the number of males that compete by looks or song for females is thus also rare. Yet as recently as 2014, in another landmark publication on comparative genomics, the discourse was almost unchanged from decades earlier and I quote:

We investigated the genomics of plumage color, a behaviourally important trait and longstanding example of sexual selection. Male birds have frequently evolved extravagant plumage color in response to both male-male competition and female choice resulting in remarkable sexual dichromatism (Zhang et al. 2014).

I do not believe that we have any overall figure worldwide showing that males have ‘frequently’ evolved extravagant plumage. At least in Australian birds the statement refers to a rather small proportion of birds and hence such theoretical positions sometimes do not ring true in the Australian context, yet they get reinforced and reiterated as if they required no further critical and sceptical review.

In sulphur-crested cockatoos, pink cockatoos and galahs, rainbow lorikeets and many cuckoos, white-winged choughs, apostlebirds, ravens, crows, kookaburras, magpies, currawongs, tawny frogmouths, owls, many pigeons and doves and many more, the differences in physical appearance between male and female are usually slight and nearly impossible to pick (at least for the human observer). In some cockatoos, only eye colour distinguishes the sexes; there are slight differences in markings or depth of colouration in magpies and kookaburras, or slight size differences as in tawny frogmouths or some owls. Added to this, females and males may jointly incubate eggs and feed the youngsters in turn, and it is therefore perhaps not surprising that equality is also often expressed vocally, in behaviour and even in similar or same plumage. In other words, overt competition and sexual dimorphism are largely absent in so many Australian species and it would be important to understand more fully why this may have occurred and what the implications may be.

In groups or long-bonded pairs, communication and some form of agreement of signals may be vital, especially in cooperatively breeding groups with such complex compositions as those of superb fairy-wrens, bell miners, noisy miners, white-winged choughs and many others. Group living does not just consist of cooperation, however. Conflicts can occur and there is often a very strict hierarchy that is established and such social complexities may need to be negotiated. Voigt and colleagues showed in cooperatively breeding white-browed sparrow weavers that the transition from subordinate helper to dominant breeder male induces the production of a new type of song (Voigt et al. 2007). This new song can only be produced if the HVC nucleus is enlarged. Expressed differently, changing social cues alone can lead to physiological changes as dramatic as actually generating enlargement of the HVC necessary for producing the new song (Voigt et al. 2007). Such differentiation and plasticity in a bird brain would have been considered impossible just 50 years ago. If a single change in dominance hierarchy can have such profound consequences on the brain within one lifetime then other group pressures could conceivably have profound effects too, especially over long periods of time.

Changes in song and repertoire may have ecological, social or biological reasons and not all of them have any effect on brain size, of course.

Specific changes in the lives of Bengalese finches, however, led to remarkable changes. Bengalese finches – a strain derived originally from munia some 250 years ago – were bred for colour variation but, unexpectedly, they also developed far more complex song (Takahasi and Okanoya 2010). Why did they develop complex song from a species with simple song? The researchers discovered several differences. The tutoring of juvenile munias by the father and other males – very similar to the practice in zebra finches – had become relaxed among breeding Bengalese finches and the offspring developed more complex song without the paternal constraints.

Perhaps the most crucial finding was that the Bengalese finch showed substantially lower levels of corticosterone, the stress hormone than the munia (Suzuki et al. 2012). These studies found considerable evidence to suggest that song complexity is related to corticosterone levels.

Previously, the researchers had found that corticosteroid receptors were expressed in song-control areas of the brain of the Bengalese finch brain. They then cross fostered munia with Bengalese finches and found that munia could also learn the more complex song, even if not expressed. The learning style changed, the model changed and no strict rules were imposed. Munias were able to reach optimal song and memory conditions under a different regime.

Importantly, however, it was the low stress level that was pivotal in the thriving song development of these birds. Based on these results, they hypothesise that reduced stress hormone levels through domestication might be one factor allowing for the development of more complex songs in Bengalese finches (Suzuki et al. 2012).

From these data, the researchers then extrapolated somewhat daringly that such spontaneous change of song (from simple to complex) might also explain the evolution of human language – not arising through a long drawn-out process of natural selection as had been presumed, but through the innovation of stable communities and domestication. They argue that it was safety – that is, a reduction in stress hormones – that ‘liberated’ humans to speak (Douglas 2014). The cover page of The New Scientist (February 2014) carried the sensational title ‘Why this bird holds the key to human language’ (‘this bird’ is referring to the Bengalese finch). Perhaps that was a little overstated.

We do not need to pursue the efficacies of the relevance of song of the Bengalese finch to the evolution of human language. However, their discovery is still relevant here for several reasons. It was argued before that birds that play (i.e. juveniles engage in unsupervised activity with each other) show lower stress levels and develop larger brains (see Chapter 6). Here it was shown that release from the traditional constraints of controlled and reinforced learning by one species, the munia, led to remarkable and spontaneously improved and complex song and a marked reduction of corticosterone levels.

Whatever the conditions were that led Australian species to play or show lower stress levels – it is well known that lower stress levels also lead to a longer and healthier life and larger brains.

The results by Suzuki et al. (2012) confirm in a dramatic way that learning increases when stress is reduced. This is also entirely consistent with the findings by Keagy et al. (2009) that satin bowerbird females chose to mate with the smartest males, as will be further discussed in Chapter 11. Bengalese and munia females also prefer the males with the most complex song. And complex song demands a higher allocation of neurons to the HVC.

Hence, environmental conditions and natural selection produce conditions in which learning is suddenly catapulted into complexity, also demanding more from memory. Indeed, some writers have argued that social learning is also central to innovation (Logan and Pepper 2007). There could be no stronger argument why and how learning is highly relevant to cognition in birds but, perhaps, it is so only when the theoretical parameters are shifted away from the dominant discourse to some more unusual, but not uncommon, contexts as found also in Australian birds.

Importantly, the model of competition is obviously not the only sustainable model. Cooperation, safety and certainty may be equally important in the success of a species. Of course, the outcomes and parameters may be entirely different. In the competitive model, learning (in males) is for a well-rehearsed performance: an Eisteddfod of dance or song. In the cooperative model, learning is for communication and not all of that can be pre-rehearsed because it is interactive. The former may be done by rote learning and a behaviour may become fixed once learnt. The latter, cooperative model, may require considered and flexible responses not just on some occasions but on a regular basis, and even over a lifetime.

Hence, a social context that requires some form of negotiated living, be this in a pair or in a larger cooperative group seems to be more open-ended and thus conceivably could better foster vocal complexity and communication (Krams et al. 2012) and thus the evolution of a larger brain than the sexually dimorphic/competitive model, at least according to the ‘social brain hypothesis’. That will be the subject of the next chapter.

If you find an error or have any questions, please email us at admin@erenow.org. Thank you!