Convergent cognitive evolution: what can be learnt from comparisons with corvids and cephalopods? Piero Amodio Sidney Sussex College University of Cambridge February 2020 Supervisors: Nicola S. Clayton & Ljerka Ostojić This thesis is submitted for the degree of Doctor or Philosophy Preface This thesis is the result of my own work and includes nothing which is the outcome of work done in collaboration except as stated in the Declaration and specified in the text. This thesis is not substantially the same as any work that I have submitted before, or, is being concurrently submitted for any degree or other qualification at the University of Cambridge or any other University or similar institution except as declared in the Preface and specified in the text. I further state that no substantial part of my thesis has already been submitted, or, is being concurrently submitted for any such degree, diploma or other qualification at the University of Cambridge or any other University or similar institution except as declared in the Preface and specified in the text. This thesis does not exceed the prescribed word limit for the relevant Degree Committee. ii PhD Candidate Piero Amodio, Department of Psychology, University of Cambridge Thesis title Convergent cognitive evolution: what can be learnt from comparisons with corvids and cephalopods? Summary Emery and Clayton (2004) proposed that corvids (e.g. crows, ravens, jays) may have evolved – convergently with apes – flexible and domain general cognitive tool-kits. In a similar vein, others have suggested that coleoid cephalopods (octopus, cuttlefish, squid) may have developed complex cognition convergently with large-brained vertebrates but current evidence is not sufficient to fully evaluate these propositions. The aim of my thesis is to gain further insight into these issues. My first objective is to further our understanding of how deep the cognitive convergence between corvids and apes may be. To this end I report four empirical studies exploring cognitive complexity among different domains in a single species of corvids, the Eurasian jay (Garrulus glandarius). In Chapter 2, I investigate physical problem solving, finding that jays cannot spontaneously select functional tools according to their physical properties (i.e. size, shape) but can use novel tools – sticks – to solve a familiar task. In Chapter 3, I test whether future oriented caching in jays is underpinned by future planning abilities or by spontaneous predispositions; data do not support either hypotheses. In Chapters 4 and 5, I focus on social cognition. I find no indication that jays can integrate the visual perspective and current desire of competitors to protect their caches. Surprisingly, jays could also not respond to either of the two social cues independently, thus questioning the reliability of previously reported caching strategies. My second objective is to lay theoretical and methodological groundwork for studying convergent cognitive evolution between cephalopods and large-brained vertebrates. Therefore, in Chapter 6 I propose an evolutionary hypothesis for the emergence of large brains and behavioural flexibility in cephalopods and put forward novel paradigms that may allow researchers to explore the cognitive underpinning of octopus’ complex behaviours. I conclude by discussing the implication of my findings and future directions for the study of convergent cognitive evolution in corvids and cephalopods. iii Declaration The work in this thesis was carried out under the supervision of Professor Nicola S. Clayton and Doctor Ljerka Ostojić at the Department of Psychology and in Clayton’s Comparative Cognition Laboratory at the Sub-Department of Animal Behaviour at the University of Cambridge where the corvids were housed. In Chapter 2, the experiments were designed in collaboration with Sarah Jelbert. Markus Boeckle was involved in the analysis and interpretation of data. Ben Farrar acted as second observer for the inter-observer reliability. In Chapter 3, the experiments were designed by Johanni Brea (EPFL, Switzerland) and Ljerka Ostojić. In Experiment 3.1, Ben Farrar and Ljerka Ostojić were involved in data collection. The statistical analysis of the data of Experiment 3.1 and 3.2 (Appendix C) was conducted in collaboration with Johanni Brea. In Chapter 4, the experiments were designed by Christopher Krupenye (University of St. Andrews) and Ljerka Ostojić. In Experiment 4.1, Ben Farrar and Ljerka Ostojić were involved in data collection. In Chapter 5, the experiments were designed in collaboration with Ljerka Ostojić. iv Acknowledgments First and foremost, I would like to thank Nicky Clayton for giving me the opportunity to join her lab, for providing me with warm support and guidance throughout my time as a graduate student but especially in the most difficult times of my PhD, and for encouraging me to explore my own ideas. I have tried hard but couldn’t find the words to faithfully express how deep is my gratitude to my co-supervisor, Ljerka Ostojić. She has been a constant inspiration, a rigorous mentor and sympathetic friend. In the past four years, Ljerka has read every report, manuscript, research proposal (maybe even grocery list) I wrote, and tirelessly provided me with crucial feedback. If I have grown as scientist throughout my PhD, Ljerka has surely played a key role. A special thanks goes to my two advisors, Corina Logan and Steven Montgomery, for their inputs and support, and to Graziano Fiorito, for encouraging me to pursue a PhD in Nicky’s lab, for allowing me to conduct my octopus research at the Stazione Zoologica in Napoli, and for his continuous guidance. Thanks to Christopher Krupenye and Johanni Brea for sharing intriguing research questions. I wish to thank all members, past and present, of the Comparative Cognition Lab for discussions and funny chats: Luigi Baciadonna, Pauline Billard, Markus Boeckle, Katharina Brecht, Ning Ding, Anna Frohnwieser, Elias Garcia, Alison Greggor, Sarah Jelbert, Elsa Loissel, Rachael Miller, Steven Samuel, Alex Schell, and particularly Ed Legg and my PhD mates, Ben Farrar and Rachel Crosby. I have learnt something important from each and every one of you. Thanks also to the great team of technicians in Madingley, Sam Melvin, Jenny Bartley, Sarah Manley and Maggie Dinsdale, and to Ian Miller and Diane Pierce. Your help and patience have been invaluable. I am deeply grateful to my mum, dad, brother and to my friends, for allowing me to share excitements and disappointments, and for giving me the strength to go through the entire PhD. Thanks also to Gino Moruto nella Vanella and Ponta, for the fun we had while writing up a piece of our own theses together. Lastly, thank you Risca. v Contents Chapter 1: Introduction …………………………………………………………..... 1 1.1: What is intelligence? ……………………………………………………... 1 1.2: How can Intelligence be measured? ……………………………………... 3 1.3: When does Intelligence evolve? …………………………………………. 5 1.4: Convergent cognitive evolution …………………………………………. 9 1.5: Corvids’ cognitive evolution ………………...………………………….. 16 1.6: The Eurasian jay ………………………………………….……………. 24 1.7: Cephalopods’ cognitive evolution ……………………………………… 27 1.8: Thesis overview ……………………………………...………………… 29 Chapter 2: How flexible is tool use in the Eurasian jay? …………..………….... 33 2.1: Introduction ……………………………………………………………. 34 2.2: General methods ………...……………………………………...……... 38 2.3: Experiment 2.1: Size selectivity test ...………………………………....... 41 2.4: Experiment 2.2: Shape selectivity test ……...…………………………… 46 2.5: Experiment 2.3: Stick tool test …………………………………………. 50 2.6: General Discussion …………………………………………………….. 52 Chapter 3: Testing two competing hypotheses for Eurasian jays caching for the future: future planning versus compensatory caching …………………. 57 3.1: Introduction …………………………………………………………… 58 3.2: Experiment 3.1 ………………………………………………………… 63 3.3: Experiment 3.2 ………………………………………………………… 71 3.4: General discussion ……………………………………………………... 79 Chapter 4: Can Eurasian jays integrate cues about others’ desires and perspectives to protect their caches? …………………………………….. 83 4.1: Introduction ……………………………………………………………. 84 4.2: Experiment 4.1: Do jays protect their caches by choosing where to cache? …………………………………………………………….. 87 4.3: Experiment 4.2: Do jays protect their caches by choosing what to cache? ……………………………………………………………… 93 4.4: General discussion ………………………………...…………………… 98 vi Chapter 5: Caching tactics based on another’s current desire or visual perspective: how robust is the evidence in Eurasian jays? …………..... 103 5.1: Introduction ………………………………………………….……….. 104 5.2: Experiment 5.1 ……………………………………………...………… 106 5.3: Experiment 5.2 ……………………………………………...………… 114 5.4: General discussion ………………………………………………..…… 121 Chapter 6: Cephalopods: a new model group for studying convergent cognitive evolution? …………………………..…………………………………..… 125 6.1: Introduction ……………………………………………………...…… 126 6.2: How complex is cephalopods’ cognition? …………………………...… 128 6.3: Cephalopods’ route to cognitive sophistication …………………..…… 134 6.4: Behavioural paradigms to test complex cognition in the octopus ……… 139 6.5: Conclusions …………………………………………………………… 153 Chapter 7: General discussion ……………………………………………..…… 155 7.1: Corvids ………………………………………………………...……... 156 7.2: Cephalopods ……………………………………………………..…… 164 7.3: Concluding remarks ……………………………………………...…… 169 Bibliography ……………………………………………………………………... 171 Appendix A: Bird data …………………………………..………………………. 205 Appendix B: Test data ….………………………………….……...……………. 207 Appendix C: Statistical