Notes � animal behaviour VIII, social learning
Greg Detre
Monday, June 18, 2001
Prof. Marian Dawkins
Animal Behaviour VIII
Notes � animal behaviour VIII, social learning
What do animals learn from each other?
did not photocopy anything
A. Whiten and R. Ham (1992) On the nature and evolution of imitation in the animal kingdom: reappraisal of a century of research. Advances in the Study of Behaviour 21:239-283.
T.M. Caro and M.D.Hauser (1992) Is there teaching in non-human animals? Q.Rev. Biol. 67: p. 151-174.
D.F.Sherry and B.G.Galef (1990) Social learning without imitation: more about milk-bottle opening by birds. Animal Behaviour 40: p.987-9.
E.Curio et al (1978) Cultural transmission of enemy recognition. Science, N.Y. 202: p.899-901.
C.J.Nicol & S.J.Pope (1996) The maternal feeding dislay of domestic chickens is sensitive to perceived chick error. Animal Behaviour 52: 767-774.
C.Boesch (1991) Teaching among wild chimpanzees. Animal Behaviour 41: p.530-532
T.M. Caro and M.D.Hauser (1992) Is there teaching in non-human animals? Q.Rev. Biol. 67: p. 151-174.
T.M.Freeberg (1998) The cultrual transmission of courstship patterns in cowbirds Molothus ater. Animal Behaviour 56: 1063-1073.
J.Fritz and K. Kotrschal (1999) Social learning in common ravens Corvus corax. Animal Behaviour 57: 785-793.
B. Voehl & L. Huber
(2000) True imitation in marmosets. Animal Behaviour
60: 195-202.
Galef, B & Giraldeau, L.A. (2001) Social infuences on foraging in vertebrates: causal mechanisms and adaptive functions. Animal Behaviour 61: 3-15.
Teaching, along with imitation, are major processes that
can accelerate the acquisition of novel behaviour in an individual and are
considered instrumental in the development of human culture. Recent critical
reviews of the known cases of imitation in primates challenge all such claims
and offer more parsimonious explanations based on social facilitation and
stimulus enhancement (Whiten 1989). Similarly, teaching has been restricted to
humans, although in one case a sign-language-trained chimpanzee, attempted to
mould and influence the signing performance of a younger companion (Fouts et
al. 1982)
Chimpanzee mothers influence their infants� attempts to
crack nuts in the tropical rainforest of Tai National Park, Ivory Coast. The
nut-cracking techniques have been described in detail previously (Boesch &
Boesch 1984). Chimpanzee mothers may influence the development of nut cracking
in three ways (excluding the very wide-spread nut sharing):
1.
stimulating
2.
facilitating nut-cracking
3.
active teaching
Mothers often leave hammers (as opposed to carrying out around when collecting, at the risk of it being stolen by another chimpanzee) or nuts near the anvil (so that the infants only need to place the readily accessible nut on the anvil to begin pounding it). Also, providing infants with good hammers (adequate weight, regular shape, open a nut with < 10 hits) or nuts helps a lot.
They observed two cases in which the mother, noticing the infants� difficulties, made a clear demonstration of how to solve them, e.g. Ricci�s 5-year old daughter, Nina, was trying to use an irregular hammer. Nina struggled unsuccessfully, changing posture, hammer grip and nut position. Eventually, Ricci very slowly and deliberately, with Nina watching, slowly rotated the hammer into the best position, then cracked 10 nuts, sharing them with Nina. When Nina resumed cracking, she still had difficulty, but always maintained the hammer in the same position her mother had, despite changing posture 18 times, whimpering to attract Ricci�s attention and throwing a temper-tantrum.
Conclusively, Tai chimpanzee mothers show a concern about their infants� apprenticeship in tool use and, in various ways, facilitate their attempts even though at some cost to their own performance. They seem to have the ability to compare their offspring�s behaviour to their own conception of how it should be performed and anticipate the possible effects of their actions on those of their offspring. All functions attributed to human tutoring by Wood et al. (1976) seem to be performed by chimpanzee mothers here. Recruitment, direction maintenance and marking critical features are apparent in the stimulations, frustration control and scaffolding in the facilitations as well as in teaching alongside with demonstration.
These acts are adjusted to the level of skill attained by the infant, e.g. stimulation reaches its maximum at 3 when infants start to learn the basic skills of nut cracking, whereas facilitation starts at three with a maximum at 5 years. Such forms of teaching have not been observed in other wild chimpanzees and the fact that nut cracking requires 10 years to be fully acquired may be important. Other behaviour patterns in chimpanzees probably require less practice and since the mother�s fitness might suffer if the nut-cracking technique is not to some extent acquired by her infant before she invests in new offspring, she increases her involvement.
Common ravens have a reputation for being intelligent. They are ecologically highly adaptable, are both neophobic and neophilic (Heinrich, 1988, 1995), long-term monogamous and territorial for most of their life, but also flock.
They trained a �model� raven to open a box with food inside a certain way, then paired it up with an �observer� conspecific. They also had pairs of untrained control birds. The control birds always opened the box the same way, but the observer learnt the different way from the model bird.
observation of a model�s activity makes the location or object of the model�s behaviour attractive for an observer (Thrope 1963, Galef 1988, Campell & Heyes in press). Heyes (1994) defined stimulus enhancement as one-stimulus learning, involving no association between the location or object and the reward.
Defined as a kind of classical conditioning, where the observer associates the location or object with the reward obtained by the model (Cook et al. 1985, Heyes 1994).
Both stimulus enhancement + observational conditioning can increase the probability of the observer learning an operant task.
Motor imitation is defined as learning the operant task directly through the observation of the model�s behaviour (Heyes 1994, Zentall 1996). It is usually considered the cognitively most demanding category of social learning (for the observer), since it requires the translation of a visual input into a matching motor output which may involve more complex central processing than other mechanisms of social learning (Whiten & Byrne 1988, Heyes 1988).
Although a number of studies� have focused on the imitative ability of different primate species (e.g. Byrne 1995, Whiten et al 1996, Bugnyar & Huber 1997), a few nonprimate mammals, such as rats (Heyes et al 1992, Heyes 1996), and several bird species, e.g. budgerigars (Galef et al 1986), grey parrots (Moore 1992), pigeons (Zentall et al 1996) and Japanese quails (Akins & Zentall 1996), the conclusiveness of the evidence for imitation learning is still debated (Visalberghi & Fragaszy 1990, Byrne & Tomasello 1995, Heyes 1998, Gardner & Heyes in press). Studies on �lower� mechanisms of social learning, e.g. enhancement or observational condition, are rare.
Usually, imitation learning (to the exclusion of other mechanisms) is demonstrated by the two-action procedure (Dawson & Foss 1965), where there are two alternative possibilities for solving a mechanical task. If the observer shows a bias towards the model�s behaviour over the control birds, that qualifies as imitation learning.
In their model, they used �qualitative asymmetry�, i.e. one of the methods was harder (pulling the lid) than the other (levering), since they didn�t have enough animals for the full two-action procedure. They didn�t separate the observer from the model, but had many testing devices in the arena, no distinction between observation and testing phase, allowing physical interaction and the possibility of �food-scrounging�.
Enhancement coupled with individual learning is the prevalent cognitive mechanism to explain learning processes that may be mistaken for motor imitation (Zentall 1996). This mechanism directs the observers� attention towards a place (local enhancement or area copying) or towards an object (stimulus enhancement or object copying) where the model acts (Giraldeau 1997).
All observers clearly differed from the control individuals in their opening behaviour. Fritz & Kotrschal noted three different aspects: the time birds took to open lids; the mechanisms involved in learning to open the boxes; and the scrounging behaviour of the observers.
Observers showed no fear-motivated behaviour (jumping jacks)� as all control individuals did on their first contact with a box. Other examples of reducing fear in the observers and focusing their attention towards the stimuli or conditions where the model acts (promotes the acquisition of innovations and the formation of traditions) include: pine-cone stripping of black rats (Zohar & Terkel 1996) or the opening of milk bottles by great tits (Sherry & Galef 1990).
They conclude that stimulus enhancement seems to be a sufficient mechanism to explain their main results.
Palameta & Lefebvre�s (1985) study of pigeons showed that observers were only sufficiently motivated to peck through a paper lid on dish when they saw a conspecific doing it, but not when they only saw the model either pecking or eating. This suggests that the pairing of a demonstrator�s response with a secondary reinforcer is very important.
By having less food available, there was a closer balance between the profitability of being a scrounger or a producer, so in this case the opportunity for scrounging did not prevent the learning of the task. It might even help social learning, since it requires such careful observation.
Rats are highly social and intelligent animals, well-adapted to live in unpleasant, often disease-ridden areas. When warfarin, an anti-coagulant, was employed as a rat poison to produce internal haemorrhages, rats evolved to become resistant to it, and became less likely to ingest it by avoiding places where it might be.
Brunton and Macdonald used radiocollars to follow rats� movements on a farm where a major extermination attempt had failed to kill more than a third of the rat population. The rats had become incredibly cautious, avoiding any new food and travelling long distances to get to familiar food.
Galef paired rats in cages, then took a �demonstrator� rat out and fed it with a distinctive, novel food (e.g. cinnamon or cocoa). When he put it back in the cage, the �observer� rat groomed and sniffed the demonstrator rat, and when faced with a choice of novel foods, always picked the one that the demonstrator rat had just had. Next, the demonstrator rat was fed saccharine, which all the rats were used to, laced with lithium chloride. This made the demonstrator rat ill, and so the demonstrator rat was careful to avoid saccharine afterwards, having (mistakenly) connected it with its illness. The observer rat, who had had no contact with the lithium chloride, but had observed the illness of the demonstrator rat after eating saccharine, was also reluctant to eat saccharine after. Strictly speaking though, this is evidence of �social learning�, rather than �culture� (which implies the wisdom of one generation being passed non-genetically to the next).
Modern rat poison kills in very small doses, but if rats don�t try out unfamiliar food, they may well starve. Young rats show a marked preference for food that is surrounded by the scents, urine and faeces of other rats, and especially the food that they have watched their parents eat. This is evidence of �cultural learning�, at least on some definitions of �culture�. When a food X was laced with lithium chloride, the whole colony could be seen to avoid it after a few rats had fallen ill, and this cultural taboo about food X persisted down the generations.
Teaching is normally understood as directed instruction of one individual by another, which you would expect would by a faster and more efficient means of transmitting information than other less directed types of social learning. There can be vertical or horizontal transmission of information (Cavalli-Sforza and Fedlman 1981, Boyd and Richerson 1985, 1988).
Reasons for this review:
1. Comparative observations suggest there are rudimentary forms of teaching in a variety of species, so far ignored because teaching has been characterised by stringent and mechanistic operational definitions (e.g. Barnett 1969, Pearson 1989).
2. We don�t know how much teaching there is going on within the general social context.
3. Theoretical and experimental progress in studies of the dynamics of cultural transmission in animal populations.
4. Resurgence of interest in animal cognition and the possibility of the complex mental processes relevant to investigations of teaching
Caro & Hauser do not think that teaching depends on complex intentionality or attribution of mental states, although intentionality plays some role in some forms of teaching. Nor is modulation of the teacher�s behaviour for the benefit of the pupil, or recognition of the pupil�s ignorance, necessarily a part.
They want to provide a working definition unifying functional and mechanistic considerations, encompassing a diverse set of obesrvations but excluding other types of social learning:
An individual actor A can be said to teach if it modifies its behaviour only in the presence of a na� observer, B, at some cost or at least without obtaining an immediate benefit for itself. A�s behaviour thereby encourages or punishes B�s behaviour, or provides B with experience, or sets an example for B. As a result, B acquires knowledge or learns a skill earlier in life or more rapidly or efficiently than it might otherwise do, or that it would not learn at all.
This definition emphasises that the teacher�s behaviour will be (markedly) different from its normal background repertoire, rather than the observer simply imitating normal behaviour.
social learning takes the following forms:
stimulus enhancement
observational conditioning
motor imitation
facilitation???
teaching
would you expect the nut-cracking to become an innate technique after a while, or would you expect their brains to become bigger to learn it more easily and stay flexible???
why does evolution always lead to more complexity??? because there�s only certain types of food that simpler organisms can obtain/digest???
flocking in crows???
operant???
stimulus enhancement???
the 3 ways chimp mothers help their offspring learn to crack nuts???
what type of learning mechanism is the behaviour of rats smelling a new food on each and modifiying behaviour accordingly???
what about imitation learning in dolphins???
is social learning the same as cultural transmission???
what about linguistic cultural transmission???
how related are theory of mind and social learning, e.g. imitation???
if you take a wild animal and tame it, or let a hand-reared animal loose into the wild, are they ever able to adapt to both??? they can�t ever be fully adapted to both, e.g. affectionate and neophilic in captivity and aggressive and cautious/neophobic in the wild
what does the Fritz & Kotrschal study conclude/show about imitation or social learning??? � not imitation, but definitely stimulus enhancement soc learning, right???
Types of social learning
stimulus enhancement
observational conditioning
motor imitation
teaching (and facilitation)