Can wildlife learn harmful and maladaptive behaviours from each other? If so, insights into social learning among animal populations in response to anthropogenic stimuli are of wide interest and applicability. Donaldson et al. (2012) address social learning in human–wildlife interactions involving food provision. ‘Food-conditioned’ animals are subject to operant conditioning in which learning about anthropogenic food arises from repeated exposure to human stimuli, behavioural responses to those stimuli and reinforcement of behavioural responses because of food reward (Whittaker & Knight, 1998; Samuels & Bejder, 2004). In this paper, Donaldson et al. carefully negotiate the asocial/social learning dichotomy. The former arises from individual responses to the availability of anthropogenic food (e.g. in species that receive limited maternal care and are solitary as juveniles and adults). However, responses to provisioning may also arise through social learning in cases where individual animals are repeatedly exposed to the feeding behaviours of conspecifics that exploit anthropogenic foods. Operant conditioning describes a learning process that may be acquired individually or socially (Sargeant & Mann, 2009). Social learning involving responses to anthropogenic food may then be facilitative or supplementary.
The specific focus of this paper is the long-term illegal provisioning of bottlenose dolphins (Tursiops aduncus) by members of the recreational fishing community in Cockburn Sound, south-western Australia. Social learning is available in this case, as individual bottlenose dolphins exist within complex social structures that involve extended juvenile dependence and enduring relationships between animals (Connor et al., 2000; Lusseau, 2003; Sargeant & Mann, 2009). Two variables emerge from the analysis as predictors of individual animals learning to accept food from recreationists engaged in fishing; the use of areas with high densities of recreational boats and association with previously conditioned dolphins. The harmfulness of such learned behaviours may include increased wildlife morbidity/mortality, interventions to address ‘problem’ wildlife (which may involve culling) and, therefore, compromised sustainability resulting in loss of economic opportunities.
These findings highlight the facilitative function of social learning in determining the development of individual animal behaviour responses to anthropogenic stimuli. They should motivate others to explore how these findings may extend beyond food provision (to other anthropogenic stimuli) and into a range of applied fields. One important application of social learning in human–wildlife interactions is sustainable tourism management. Knight (2009: p. 180) observes that we now live ‘… in an age when our visual appetite for wildlife has never been greater’. Wildlife viewing has rapidly moved into the mainstream of commercial tourism (Knight, 2009). More social species, including cetaceans, have become the focus of global interest, as tourists seek to observe critical behaviours such as maternal care, feeding and social interactions between conspecifics in the wild. The International Fund for Animal Welfare estimates that the whale watching industry now exceeds $2.1 billion per annum, catering for 13 million whale watchers, and generating 13 000 jobs (O'Connor et al., 2010). Neves (2010) describes this evolution as one form of periodic transformation in the global capitalist economy, and it is a transition that also raises the likelihood of social learning among individual animals that are subject to human stimuli through intensive and prolonged interactions with commercial tourism business and their clients.
The provisioning of wild dolphins in a tourism context is unusual but does occur (Orams, 1995). More broadly, the provisioning of wild animals, common in the past (e.g. bear feeding in Yellowstone National Park; Davis, Wellwood & Ciarniello, 2002) continues in some tourism contexts, both directly (e.g. use of food to attract pelagic birds, sharks) and indirectly (e.g. attraction to camp sites because of food storage and disposal, Knight, 2009). This raises questions of diminished behaviours and reduced ‘wildness’ associated with efforts (deliberate or otherwise) to make wildlife viewable (Knight, 2009). However, the relevance of social learning in human–wildlife interactions should extend beyond provisioning, to other forms of tourism (e.g. swimming with dolphins) and specific behaviours (e.g. bow riding – the behaviour of dolphins in which they swim or ‘ride’ the crests of waves formed from the front of moving boats) that may increase the likelihood of harmful outcomes (e.g. collision with vessels). This paper may also give reason to consider the possibility that absence of response behaviours may be socially acquired (Bejder et al., 2009). Such a scenario could arise through neonates being exposed to increasingly unresponsive mothers and/or peer modelling of indifference to human approach (Higham & Shelton, 2011). The absence of avoidance response, if socially acquired, may also compromise ‘wildness’ because of a reduced behavioural repertoire leading to potential harmful outcomes for some individual animals.
This paper highlights various avenues and approaches to continuing inquiry. They may include social learning of harmful behaviours arising from responses to a range of human stimuli (e.g. visibility/movement, noise, manoeuvre of vehicles), perhaps extending to a reduced behavioural repertoire (i.e. abandonment/absence of behaviours), which may also be maladaptive. While the importance of species-specific data and site-specific interventions is well established in the tourism management context (Higham, 1998), this paper underscores the importance of individual-based sampling methods to ensure that single animals are the base unit of analysis in longitudinal studies. The authors are clear in not dismissing the potential influence of propensities associated with age class and sex, and these attributes are available for further empirical investigation. Individual datasets with detailed longitudinal timelines are required to extend the current work to detect the effects of these attributes as predictors of individual behavioural responses to anthropogenic stimuli. This avenue of research should include individual-specific ranging data which, when analysed in association with spatial data on the locations where human–wildlife interactions occur, allows the identification of individual animals that may be at high risk of learning maladaptive behaviours. In such cases, management interventions are particularly critical where animals exhibit long-term fidelity to ecological zones where high levels of human use (tourism or otherwise) take place, although this also requires insights into the extent to which behavioural responses are spatially and temporally stable. It also raises questions as to the efficacy of management interventions intended to ‘uncondition’ animals in order to reverse maladaptive behaviours (Knight, 2009).
The relevance of this paper extends to all forms of human–wildlife interplay, adding weight to the case that individual animals are the base unit of analysis and interpretation. It highlights the complexity of conservation management interventions to mitigate the negative consequences of undesirable behaviours in wild animal populations. Although the provisioning of wild animals, once common in the tourism context, has become less prevalent over time (Tate, 1983), this paper offers broad applicability to the field of conversation management and will be of great relevance to those with interest in sustainable tourism management.