Periodic Reporting for period 4 - INQMINDS (The Evolutionary and Developmental Origins of Inquiring Minds: Studies of Causal Reasoning; Curiosity and Executive Control)
Berichtszeitraum: 2020-02-01 bis 2022-01-31
Human technology is vastly superior to that of other apes: human tools from telescopes to the Large Hadron Collider exploit causal relationships but also explore them. What change over the brief course of hominid evolution made for such a big difference? One theory implicates a single cognitive distinction: the ability to reinterpret the world in abstract terms. In contrast, in several studies, chimpanzees seemed to rely on surface appearances, as if for the apes there was no more to the situation than met the eye. The theory is seductively simple, but this project was motivated by the conjecture that it is too early to rule out deep evolutionary roots for inquiring minds. We pursued two broad lines of study: causal cognition and executive function. To disentangle causal reasoning from simpler mechanisms we took three approaches. The first looked at intervention: would subjects use exploration to seek hidden causes? The second focued on natural mechanics: would subjects reason about physical properties or rely on perceptual correlates? The third examineed causal learning from observation when the task’s mechanics are opaque. The second line of study explored executive functions. An inquiring primate mind could flourish if more information could be held in mind and manipulated. But we know little about how executive functions compare across primates. We developed 9 new tests, culminating in a new test battery explored levels of working memory, inhibition, and attention shifting. The project used cross-sectional testing of causal cognition and executive control in the same individuals to explore how these cognitive skills interact to produce inquiring minds. Our work shed light on the fundamental building blocks of cognition and how these might change over evolution, with implications for evolutionary biology and artificial intelligence.
Causality
In the first strand we made fascinating discoveries about similarities and differences between causal learning and curiosity in primates and children. Both non-human primates and human children were able to solve problems involving predictive causal inference – to avoid a trap or blockage (Jordan et al 2020; Yuniarto et al 2020). And excitingly, there was evidence that they could learn in the absence of reward, through exploration, suggesting that curiosity about the way the world works may have deep evolutionary roots (Jordan et al submitted; in prep). Yet we also found some differences in performance. In tests that required participants to infer an unseen cause in order to find a reward, we found that four- and five-year-old children outperformed monkeys and apes across two methodologies (Civelek et al 2020; 2021). We found that children consistently performed in line with predictions from computational models that are capable of learning not just specifics, but also overarching rules or hypotheses. However, in one study non-human primates showed no evidence of learning, neither at the specific nor the abstract level (Felsche et al 2019). However, in a second study we found tentative evidence that both chimpanzees and capuchin monkeys could learn the abstract rule governing the distribution of rewards in puzzle boxes (e.g. all boxes contain the same type of reward, or all boxes contain a mixture of rewards). Interestingly, when children outperformed primates their performance was related to explain their solution using language, suggesting that thinking in words plays a powerful role in enabling humans to go beyond their primate relatives in their causal reasoning, and harness their inquiring minds to make new discoveries.
Executive Functions
We developed new methods to allow us to compare executive functions across human and non-human primates, in three domains. We developed a new methodological approach based on human psychometrics for establishing validity (Voelter et al 2018). We used this approach to develop new tests of working memory (Voelter et al 2019; Reindl et al 2021); inhibition (Voelter et al 2022), and attentional flexibility (Reindl et al 2022; Reindl et al in revision). Interestingly in some of the experiments children performed better than chimpanzees by the age of 4-5, such as tests of attentional control (Reindl et al in revision) and working memory (Reindl et al in preparation).
Cross-sectional Test Battery
In a huge coordinated effort by our team, we tested 55 chimpanzees and 185 children on a test-battery, using the 12 new tasks developed in the previous 2 strands that yielded the appropriate signatures in both species (Voelter et al, 2022). Our results overall indicated some unity in the structure of individual differences, consistent with the evidence for adult humans. However, we found no strong evidence for an unequivocal structure in the EF tasks for either preschool children or chimpanzees. Task performance only correlated at low to moderate levels and there was no clear indication of a preference for any one of the models proposed in the literature. One explanation for the pattern of results is that performance in EF tasks in children and chimpanzees is not well explained by the presence of several dissociable underlying capacities. The exercise of executive control might best be understood as a dynamic and distributed process involving heavily overlapping networks. In other words, EF are a “moving target” rather than discrete, domain-independent skills. We found no evidence that these constructs are natural kinds that can be identified by tasks raising the same demand in different experimental contexts.
In the first strand we made fascinating discoveries about similarities and differences between causal learning and curiosity in primates and children. Both non-human primates and human children were able to solve problems involving predictive causal inference – to avoid a trap or blockage (Jordan et al 2020; Yuniarto et al 2020). And excitingly, there was evidence that they could learn in the absence of reward, through exploration, suggesting that curiosity about the way the world works may have deep evolutionary roots (Jordan et al submitted; in prep). Yet we also found some differences in performance. In tests that required participants to infer an unseen cause in order to find a reward, we found that four- and five-year-old children outperformed monkeys and apes across two methodologies (Civelek et al 2020; 2021). We found that children consistently performed in line with predictions from computational models that are capable of learning not just specifics, but also overarching rules or hypotheses. However, in one study non-human primates showed no evidence of learning, neither at the specific nor the abstract level (Felsche et al 2019). However, in a second study we found tentative evidence that both chimpanzees and capuchin monkeys could learn the abstract rule governing the distribution of rewards in puzzle boxes (e.g. all boxes contain the same type of reward, or all boxes contain a mixture of rewards). Interestingly, when children outperformed primates their performance was related to explain their solution using language, suggesting that thinking in words plays a powerful role in enabling humans to go beyond their primate relatives in their causal reasoning, and harness their inquiring minds to make new discoveries.
Executive Functions
We developed new methods to allow us to compare executive functions across human and non-human primates, in three domains. We developed a new methodological approach based on human psychometrics for establishing validity (Voelter et al 2018). We used this approach to develop new tests of working memory (Voelter et al 2019; Reindl et al 2021); inhibition (Voelter et al 2022), and attentional flexibility (Reindl et al 2022; Reindl et al in revision). Interestingly in some of the experiments children performed better than chimpanzees by the age of 4-5, such as tests of attentional control (Reindl et al in revision) and working memory (Reindl et al in preparation).
Cross-sectional Test Battery
In a huge coordinated effort by our team, we tested 55 chimpanzees and 185 children on a test-battery, using the 12 new tasks developed in the previous 2 strands that yielded the appropriate signatures in both species (Voelter et al, 2022). Our results overall indicated some unity in the structure of individual differences, consistent with the evidence for adult humans. However, we found no strong evidence for an unequivocal structure in the EF tasks for either preschool children or chimpanzees. Task performance only correlated at low to moderate levels and there was no clear indication of a preference for any one of the models proposed in the literature. One explanation for the pattern of results is that performance in EF tasks in children and chimpanzees is not well explained by the presence of several dissociable underlying capacities. The exercise of executive control might best be understood as a dynamic and distributed process involving heavily overlapping networks. In other words, EF are a “moving target” rather than discrete, domain-independent skills. We found no evidence that these constructs are natural kinds that can be identified by tasks raising the same demand in different experimental contexts.
In several of our experiments children performed better than chimpanzees by the age of 4-5, such as tests of inferring unseen causes (Civelek et al 2020; 2021), abstracting complex overhypotheses (Felsche et al 2019), attentional control (Reindl et al in revision) and working memory (Reindl et al in preparation). The results with children are especially interesting as our tasks have little to no verbal component. This might suggest that the advantage of language learning goes beyond making use of verbal instructions and also impacts performance on tasks in which children have to control their behaviour by setting their own goals in response to feedback.
Voelter et al (2022) was the first study to systematically compare the structure of executive functions in human preschoolers and chimpanzees. It was also the first to implement a 3 (constructs) × 3 (tasks) test battery for young children and the first to investigate all three main functions of EF (WM, inhibition, attentional set shifting) across two species. This goes significantly beyond the state of the art in which cognitive abilities are compared in isolation. Discovering that these supposedly disparate abilities may well not be natural kinds has significant implications for the study of cognitive evolution.
Voelter et al (2022) was the first study to systematically compare the structure of executive functions in human preschoolers and chimpanzees. It was also the first to implement a 3 (constructs) × 3 (tasks) test battery for young children and the first to investigate all three main functions of EF (WM, inhibition, attentional set shifting) across two species. This goes significantly beyond the state of the art in which cognitive abilities are compared in isolation. Discovering that these supposedly disparate abilities may well not be natural kinds has significant implications for the study of cognitive evolution.