Social learning in animals

Assessment | Biopsychology | Comparative | Cognitive | Developmental | Language | Individual differences | Personality | Philosophy | Social |
Methods | Statistics | Clinical | Educational | Industrial | Professional items | World psychology |

Animals · Animal ethology · Comparative psychology · Animal models · Outline · Index

---

This article deals with animal populations. For humans see: Social learning

Example of the two-action task paradigm. An individual can solve the task either by lifting the lever and sliding the food out, or turning the lever at the top and stabbing the food out with a stick.

Adaptive behavior in animals is often explained by natural selection on genes, however natural selection on the ability to learn socially may also be an important adaptive strategy, for various reasons. Unlike behaviour driven by genes, social learning allows individuals to adapt to environmental changes at a far more rapid pace. Unlike indicidual learning, social learning can be far less dangerous, as is the case with learning about new predators.  Social learning has been quite broadly defined as: " ...any incidence in which individuals acquire new behaviour or information about their environment via observation of, or interaction with, other animals or their products". Clearly, social learning does not refer to a single ability or mechanism, but a wide range of ways by which information may be derived from the environment. Various types of social learning have been proposed.

Some forms of "social" learning may involve simple association learning. An example is the learning of new predators in fish through associating the alarm behaviour of other fish with a new predator's chemicals (see below). 

Another example of social learning which is often treated as a seperate phenomena is teaching. In some species, evolution has shaped the behavior of animals to directly aid inexperienced others. Of course, teaching as it is studied by animal researchers uses a broader definition that is used in human teaching. The best evidence for teaching in nonhuman animals actually comes from meerkats and ants, rather than primates, which may be unusual for those who presume that primates are cognitively superior. 

Term	Description
Copying	Individuals may copy different components of another's behaviour. "Imitative" copying involves copying the form of action, wheras "end-state/goal emulation" involves copying only the environmental outcome of an action. "Sequence-centred social learning" is defined as being able to copy the shape and order of a series of actions.
Observational conditioning	An increase in the likelihood of a stimulus eliciting a response, based on the subject's unconditioned reaction to the responce of another animal seeing the stimulus.
Affordance learning	Learning the about the operating chatacteristics of objects or the environment via social interactions
Enhancement	Focusing attention on part of the environment
Guided Learning	Following a more experienced animal, and consequently learning something individually from the environment
Social/Response facilitation	Social 'facilitation is when a conspecifics behaviour motivates the individual to engage in similar behavior, , or generally incrase the motivation of an animal to perform a task. Aditiionally, some researchers posit that the behavior facilitated must already be in the individual's repertoire. However, some researchers would rather classify this form of "learning" as social influence, since no new behaviour is learnt.  However, the increased motivation to perform a certain action may direclty lead to individual learning.

Experimental Methods

Many experiments have been designed to establish the existance of social learning and to differentiate between which mechanisms animals are using. 

Term	Social Learning methods	Experimental manipulation
Two-action task	Emulation vs. Imitation	Where a learning task can be solved two ways (often to get a reward). If observers copy the demonstrators method, then imitation is implied, however if the observers are equally likely to use both methods and are equally likely to solve the task, then goal emulation is implied, as only the outcome is copied.
Ghost control tasks	Emulation vs. Imitaton	The observer sees the environmental changes normally caused by a model, however without a model causing the effect. Similar to other observational learning tasks, however the model is removed.
Do-as-I-Do	Imitation	Animal (normally performed with primates) has to copy the bodily actions of a human, such as gestures, facial expressions or actions.

In some studies subjects are reared in isolation from conspecifics to try to establish how much of a behavior is innate and how much has to be learnt in a social context. In other studies cross-fostering is used to try to tease out the same issues.

Evidence of Social Learning

Some reseachers have assumed that vertebrates, especially primates, are the only animals which are especially adapted for social learning. However, forms of social learning have been identified in a wide range of taxa.

Mammals

Apes

Main article: Social learning in apes

The greater apes consist of: chimpanzees, bonobos, gorillas and orangutans. In the 1990s, research questions shifted from asking whether "do primates copy" to "how do primates copy "?  The do-as-I-do task (see above)  has established that chimpanzees, bonobos and orangutans imitate bodily actions, and gorillas have been shown to have similar abilities using a different task. 

Elephants

Despite their large brain sizes, there is little non-anecdotal evidence for any sort of social learning in the elephant family, although this can mainly be attributed to the difficulty in creating experimental tasks appropriate for this species. African bush elephants recently failed a test of bodily imitation using the two-action task paradigm (see imitation above).

Rodents

Evidence for bodiliy imitation through the two-action paradigm (see above) has been equivocal for many vertebrates. Hayes found that rodents allowed to observe a conspecific push a joystick to the left or the right (depending on which experimental group they were in) would then imitate the direction that they saw the observer push the joystick.  However, these original studies did not control whether odour cues left by demonstrator rats could have an effect, and a meta-analysis showed that the effect sizes are typically low.

Birds

Main article: Social learning in birds

Various studies of birds have positied social learning abilties. Bird song is a good example. Whilst bird song is affected by biological factors such as hormones and maturation, birds in different regions have different "dialects" which are explainable by social learning. Genetic explanations like differences caused by genetic drift or selection have been ruled out by studies which transfer young chicks to different regions, which then pick up the dialect of the environment. It has been shown that birds have a sensitive period (which varies across species) where the relocation of birds has an effect. Some have argued that this represents a "special" form of social learning which is functionally and evolutionarily constricted to learning songs.   Using the two-action paradigm, 

In Britian, a noticable trend was observed among blue tits, which began opening milk bottles left outside houses. The rapid spread of this behavior suggests a possible social learning explanation. However, naive chickadees (Parus atricapillus) have been shown to individual learn this ability relatively often (4/16 birds opened similar bottles without social learning) . However, in the aforementioned experiment, milk bottles are put into small cages, which makes the liklihood of an individual opening the bottle a lot higher- it is possible that individual discoveries led to a local enhancement effect of birds drinking from milk bottles drawing the attention of others, which then performed the behavior. 

Fish

Fish use social infomation in a variety of contexts, including: antipredator behavior, migration, foraging and mate choice. One form of social learning is the fright response of fish in response to predators. When one member of a group is startled, the rest of the shoal respond, and this reaction can be driven by both by visual infomation, and also chemical alarm substance in fishes of the superorder Ostariophysi. Naive fish can also learn to associate a new predator with the visual fright response of more experienced fish, clearly illustrating the adaptive value of such behavior.   These kinds of social learning, however, can be seen as little more than simple learning mechanisms in a social context. Some more complex forms of social learning, such as imitation, have not such been shown in fish.  

Social Learning in invertebrates

There is the least evidence of social learnining in this group, however this is largely to do with a lack of research. Cephalopods are relatively well studied, and we know that the common octopus (Octopus Vulgaris) shows "observational learning" when watching a demonstrator being rewarded for picking a specific ball, observers are much more likely to do the same. Social learning in the antipredator of Wood Crickets (Nemobius sylvestris) behavior has also been studied, showing that when crickets will copy the hiding behaviour of others, however it is questionable whether some unknown alarm chemical could also cause hiding in other crickets.

Eusocial insects have also been shown to learn socially from each other. 

Teaching in nonhuman animals

Tandem running (a) and carrying behaviors (b) in the ant species Temnothorax curvispinosus.

Caro & Hauser defined teaching according to four criteria:

1. A tutor modifies its behavior only in the presense of a observer
2. The tutor "teaches" at some cost to itself
3. The tutor can encourage, punish or facilitate a behaviour, as a result...
4. The observer gains knowledge or a skill more rapidly or efficiently than it would without teaching

Clearly, human teaching consists of much more than these criteria, however this definition has been useful in examining more clearly the adaptiveness of social learning and teaching in nonhuman animals.

One proposed example of teaching is "tandem-running" which is seen in some ant species. When "scout" ants have successfully found a new food-source or nest site, they use two methods of leading other ants to the site. Emigrations begin by ants leading conspecifics to the new site with tandem running (conducted by circa 10% of the group), which they then shift to using the carrying method to carry the majority of nest mates to the site. The latter method is more faster and used to carry the bulk of the group, however some have argued that ants who are carried cannot see where they are being moved to, so the tandem running method is used to "teach" recruiter ants the location of the new site so they can carry other ants to the site (the  "recruitment of recruiters" hypothesis ). Pratt and colleagues showed that 60% of followers in tandem runs later recruited ants from the old nest site to the new site, compared to 17% of those that were carried.

Experimental evidence that tandem running leads to ants learning the location at a faster or more reliable rate is lacking, however, so it is still obscure what is learned in this process. During tandem runs, the leading ant slows its pace to a 1/4 of the previous speed, so the follower can stay in close contact, pausing if the follower loses contact, which shows that the behavior is conducted as some cost to the "teacher" (see above for definition of teaching). One experimental test over whether tandem-running increases the liklihood of individuals discovering new nest sites found that over short distances tandem-running may not not be beneficial , which makes it questionable still whether we can really call this "teaching".

The clearest example of teaching in animals comes from meerkats. Adult meerkats tend to provision their youngest pups with poisonous scorpions which have been disabled or killed for them to eat, and when they get older they provide more and more intact pups. One hypothesis is that this behaviour is done so that the meerkats can "teach" their pups how to eat scorpions without being poisoned. The behavior is driven by the acoustic calls of the pups- experiments show that when calls from adult pups are played near young pups, that the adults will bring intact scorpions to the young pups (and the reverse is true). The presense of an adult also encourages young pups to tackle disabled scorpions (when compared to a control condition when pups where provisioned by experimenters with no adults around).  In addition, "teachers" will nudge prey to the pups when they do not attempt to eat the item (occurs 8% of occasions), and 99% of the time after this occurs young pups will bite the scorpions. Becasue the teaching behavior is driven by a mechanistic response to the pup's calls, one might presume that this rather inflexible teaching style might be genetic and adapted by natural selection. 

See also

• Animal cognition
• Animal culture
• Cooperation in animals
• Cultural transmission in animals
• Herd behavior
• Imitation in animals
• Imprinting
• Mimicry
• Social reinforcement