Bark to the future

: Responses of the domestic dog (Canis familiaris) to verbal cues issued by an artificial agent: Welfare and cognitive perspectives

Abstract

The twenty-first century pet domestic dog typically spends a large amount of time in the home environment. An audio/food dispensing automated device (an artificial agent, hereafter agent) was evaluated for enrichment potential. The agent issued verbal cuesto dogs and puppies, and food rewards for their correct behavioural responses. Dogs’ and puppies’ abilities to respond correctly to the cues and their welfare during interaction was measured. Three empirical studies are presented. Key findings of the cognitive tests (study one) include: 1. no significant difference in dogs’ correct behavioural responses to the agent’s cues between initial exposure and following a 24-month break (p = 0.119); 2. no significant difference in correct 1st choice behavioural responses to multiple, randomised cues issued by the agent and by owners directly (p = 0.526); 3. significantly longer response latencies to cuesin agent conditions vs owner condition (p = 0.026). Dogs were able to learn a novel sequence of known behaviours from the agent during a single testing session. Study two measured dogs’ welfare during the cognitive tests described using Qualitative Behavioural Assessment (QBA) and quantitative behaviour sampling. QBA PC1 characterised dogs as ranging from ‘attentive/interested/anticipating’ to ‘conflicted/apathetic’ in each experimental testing condition with the majority of dogs at the former end of the PC. Inter-observer agreement was high on all PCs (average Kendall’s W = 0.82). No significant differences in quantitative positive state behaviours between conditions were found (average p = 0.595); these behaviours correlated significantly with QBA PC1 in all conditions(average r = 0.791). The study represents the first application of QBA to a cognitive paradigm and suggests that positive reinforcement interaction with an owner or with an agent may promote an experience of pleasure in dogs. Study three examined puppies’ responses to randomised verbal cues issued by the agent and by owners directly for comparison. Attentional focus on the agent was rapidly achieved (average 18.8 sec from start of introduction to agent). No significant difference in puppies’ correct 1st choice behavioural responses to cues issued by the agent vs owners was found (p = 0.609); significantly longer response latencies to cues in the agent vs owner condition were revealed (p = 0.001). Overall, indicators of positive anticipation, motivation, competence and sustained engagement were identified in the majority of dogs and puppies throughout testing, suggesting that agent-interaction represents an appropriate cognitive challenge. Puppy owners reported overarchingly positive opinions of their puppies’ interactive experiences; the majority supporting the trial of an agent in the home.

Date of Award	2 Dec 2022
Original language	English
Awarding Institution	University of Winchester
Supervisor	Lisa Riley (Supervisor)