Friday, 17th August - Res. seminar - Interspecific communication Flashcards
Interspecific Communication & Learning
Human-dog communication & learning – why are we interested?
First domesticated animal
– Interest in studying this from a comparative and evolutionary perspective.
Useful
– Farming, protection.
– Assistance.
– Scent detection (conservation, border security, explosives, etc.).
In ~30% of NZ households
– Pervasive in communities.
• Animal welfare
– Association between training and behaviour complaints.
Interspecific Communication
Communication is crucial for interspecific interactions, training.
Background on research on interspecific communication:
• Pointing signals, attentional cues, verbal signals, dog-given signals.
Interspecific Communication
Pointing is a frequently-tested signal.
Two-way object choice tests.
– Human signaller stands in between two items (e.g., bowls).
- Food hidden in one item.
– Dog faces person.
- Person signals towards one bowl.
– Dog’s response is recorded, reinforced for correct choice.
Question
Consider dogs and their natural behaviour, How might their natural behaviours limit the truth of the ‘point’ research.
Dogs have an excelent olfactory system. Having some hidden food in the bottom of the bowl, They may just be using their olfactory system to go for the bowl with the treat under it apposed to going to it because someone pointed at it.
Not all dogs respond to these pointing signals.
Exercise: what are your predictions here?
Who will respond to these human-given signals better/worse?
• ‘Ordinary’ domestic dogs vs. New Guinea singing dogs (Wobber, 2009).
1st: ‘Cooperative dogs’
2nd: ‘Ordinary’
Interspecific Communication
Phenotypic differences can also influence animals’ responses to pointing
Brachycephalic (‘short-nosed’) vs. dolichocephalic (‘long-nosed’) breeds (Gácsi et al., 2009).
– Both followed pointing signals better than by chance.
– Brachycephalic responded significantly better.
• Brachy were more attentive to human signaller.
– Distribution of retinal ganglion cells is more central in these dogs
(cf. dolicho.).
The pug (Brachyephalic) responded better
Define these terms
1. Brachycephalic
2. Dolichocephalic
Brachycephalic (‘short-nosed’)
vs.
Dolichocephalic (‘long-nosed’)
Interspecific Communication
Phenotypic differences can also influence animals’ responses to pointing
• Body size makes a difference.
– Larger (>22.7 kg) dogs perform better than smaller (<22.7 kg) dogs.
– Larger dogs have greater ‘inter-ocular’ distances, and thus depth perception.
Interspecific Communication
Attentional cues (rather than true signals) can also be used by dogs.
Many species recognise and react to eyes.
– E.g., eye spots and anti-predator behaviour.
Social species may have communicative role.
– Perceiving attention cues of others may be important in predicting future behaviour.
Interspecific Communication
Attentional cues (rather than true signals) can also be used by dogs.
- Humans’ attentional states influences dogs’ behaviour (Call et al., 2003).
- Dogs forbidden to eat food.
– Less likely to take it when a person was watching, compared to when the person was out of the room, turned their back, was distracted, or had their eyes closed.
Causation
Physical differences/Differences in morphology
Ontogeny
Animals get some advantage out of it (perhaps reinforcement)
Function
?
Evolution
Atrificial selection. Different types of dog breeds.
Interspecific Communication& Learning
- Because dogs are so receptive to human communicative signals, they may play a role in more-effective training.
- Relevant to Dog Learning Lab.
Three studies:
• Field study, experiment, training observations.
Positive reinforcement = ?
Unconditioned reinforcement = ?
Conditioned reinforcement = ?
Interspecific Communication & Learning
Many variables affect reinforcer effectiveness
– timing (temporal contiguity).
Delays to reinforcement can result in trained behaviour.
E.g., pigeons and rats.
Interspecific Communication & Learning
Field observations
Aim was to examine how/when reinforcement is delivered.
- Dog training clubs.
Times between training events measured via video analysis.
– Command > dog’s response > conditioned reinforcement > unconditioned reinforcement.
– Sit and down responses.
(44% of commands = no response)
(Only 53% of commands were responded to correctly)
– scope for improving efficacy.
- Variability in delays to R+.
– Temporal contiguity is important
– therefore some suboptimal learning conditions.
Interspecific Communication & Learning Experiment:
Aim was to examine the effects of experimentally delayed reinforcement when training dogs to perform a novel task.
Species-typical response is to investigate environment via olfaction.
– Task = putting their head inside yellow box.
Method in brief:
- Visual communicative signals/cues obscured by screen.
- Dogs trained to use a feeder.
- Boxes smelled like dog food.
- As dogs put head inside box, broke IR beam – connected to computer. Computer controlled R+ delivery.
– Conditioned R+ = beep.
– Unconditioned R+ = food treat.
• Three conditions:
A. 0 s delay (beep + food).
B. 1 s delay (beep + food).
C. 0 s to beep, 1 s delay to food.
Key Findings