The phylogenetically ancient neuropeptide oxytocin has been linked to a plethora of social behaviors. Here, we argue that the action of oxytocin is not restricted to the downstream level of emotional responses, but substantially alters higher representations of attitudes and values by exerting a distant modulatory influence on cortical areas and their reciprocal interplay with subcortical regions and hormonal systems.

Contempt shares its features with other emotions, indicating that there is no justification for creating “sentiment” as a new category of feelings. Scientific categories must be created or updated on the basis of evidence. Building a new category on the currently limited contempt literature would be akin to building a house on sand – likely to fall at any moment.

Gervais & Fessler's defense of a sentiment construct for contempt captures features distinguishing the phenomenon from basic emotions and highlights the fact that it comprises a coordinated syndrome of responses. However, their conceptualization of contempt as the absence of respect equivocates. Consequently, a “dignity” culture that prescribes respect does not thereby limit legitimate contempt in the manner the authors claim.

The integration of the folk affect concept of “contempt” into the analysis of the complex institution known generally as charivari is mutually beneficial for both ethno-anthropology (which may thus access inner causes for disputed social and collective behaviors) and evolutionary psychology (which may thus study the length of tradition together with the width of the institution spread, serving the same social functions).

Beyond its implications for contempt, it remains to be determined whether the sentiment concept might be applied usefully to other domains of social affect. This commentary considers its applicability to the domain of parental caregiving. Characteristic features of sentiments are considered in conjunction with empirical research on the motivational psychology of parental care.

Gervais & Fessler argue that contempt is a natural kind and that its experience cannot be explained by a constructionist account of emotion. We dispute these claims and offer a positive constructionist model of contempt that accounts for the existing evidence and unifies conflicting findings in the literature on contempt.

The target article by Gervais & Fessler represents a comprehensive analysis of contempt but is not fully adequate in addressing how contempt produces hatred, how contempt is used manipulatively by aspiring leaders, and how contempt can be cured or at least mitigated. This commentary addresses these concerns.

Sentiment includes emotional and enduring attitudinal features of contempt, but explaining contempt as a mixture of basic emotion system affects does not adequately address the family resemblance structure of the concept. Adding forms of individual, group-based, and widely shared arrogance and contempt is necessary to capture the complex mixed feelings of proud superiority when “looking down upon” and acting harshly towards others.

Insensitive parental thoughts and affect, similar to contempt, may be mapped onto a network of basic emotions moderated by attitudinal representations of social-relational value. Brain mechanisms that reflect emotional valence of baby signals among parents vary according to individual differences and show plasticity over time. Furthermore, mental health problems and treatments for parents may affect these brain systems toward or away from contempt, respectively.

Gervais & Fessler argue that the perceived legitimacy of contempt has declined over time in the United States, citing evidence of a decrease in the frequency of its use in the American English corpus. We argue that this decline in contempt, as reflected in cultural products, is linked to shifts in key socioecological features previously associated with other forms of cultural change.

Gervais & Fessler reintroduce the concept of a sentiment as a framework for conceptualizing contempt, a construct with both attitudinal and emotional components. We propose that humility might also fit this mold. We review recent findings regarding the antecedents, phenomenology, and functional consequences of humility, and discuss why conceptualizing it as a sentiment may advance our understanding of this construct.

The target article argues that contempt is a sentiment, and that sentiments are the deep structure of social affect. The 26 commentaries meet these claims with a range of exciting extensions and applications, as well as critiques. Most significantly, we reply that construction and emergence are necessary for, not incompatible with, evolved design, while parsimony requires explanatory adequacy and predictive accuracy, not mere simplicity.

Recent progress in artificial intelligence has renewed interest in building systems that learn and think like people. Many advances have come from using deep neural networks trained end-to-end in tasks such as object recognition, video games, and board games, achieving performance that equals or even beats that of humans in some respects. Despite their biological inspiration and performance achievements, these systems differ from human intelligence in crucial ways. We review progress in cognitive science suggesting that truly human-like learning and thinking machines will have to reach beyond current engineering trends in both what they learn and how they learn it. Specifically, we argue that these machines should (1) build causal models of the world that support explanation and understanding, rather than merely solving pattern recognition problems; (2) ground learning in intuitive theories of physics and psychology to support and enrich the knowledge that is learned; and (3) harness compositionality and learning-to-learn to rapidly acquire and generalize knowledge to new tasks and situations. We suggest concrete challenges and promising routes toward these goals that can combine the strengths of recent neural network advances with more structured cognitive models.

In this commentary, we highlight a crucial challenge posed by the proposal of Lake et al. to introduce key elements of human cognition into deep neural networks and future artificial-intelligence systems: the need to design effective sophisticated architectures. We propose that looking at the brain is an important means of facing this great challenge.

We agree with Lake and colleagues on their list of “key ingredients” for building human-like intelligence, including the idea that model-based reasoning is essential. However, we favor an approach that centers on one additional ingredient: autonomy. In particular, we aim toward agents that can both build and exploit their own internal models, with minimal human hand engineering. We believe an approach centered on autonomous learning has the greatest chance of success as we scale toward real-world complexity, tackling domains for which ready-made formal models are not available. Here, we survey several important examples of the progress that has been made toward building autonomous agents with human-like abilities, and highlight some outstanding challenges.

We propose an alternative approach to “deep” learning that is based on computational ecologies of structurally diverse artificial neural networks, and on dynamic associative memory responses to stimuli. Rather than focusing on massive computation of many different examples of a single situation, we opt for model-based learning and adaptive flexibility. Cross-fertilization of learning processes across multiple domains is the fundamental feature of human intelligence that must inform “new” artificial intelligence.

The claims that learning systems must build causal models and provide explanations of their inferences are not new, and advocate a cognitive functionalism for artificial intelligence. This view conflates the relationships between implicit and explicit knowledge representation. We present recent evidence that neural networks do engage in model building, which is implicit, and cannot be dissociated from the learning process.

Lake et al. argue persuasively that modelling human-like intelligence requires flexible, compositional representations in order to embody world knowledge. But human knowledge is too sparse and self-contradictory to be embedded in “intuitive theories.” We argue, instead, that knowledge is grounded in exemplar-based learning and generalization, combined with high flexible generalization, a viewpoint compatible both with non-parametric Bayesian modelling and with sub-symbolic methods such as neural networks.

Technoscientific ambitions for perfecting human-like machines, by advancing state-of-the-art neuromorphic architectures and cognitive computing, may end in ironic regret without pondering the humanness of fallible artificial non-normative personalities. Self-organizing artificial personalities individualize machine performance and identity through fuzzy conscientiousness, emotionality, extraversion/introversion, and other traits, rendering insights into technology-assisted human evolution, robot ethology/pedagogy, and best practices against unwanted autonomous machine behavior.

We propose that early in ontogeny, children's core cognitive abilities are shaped by culturally dependent “software updates.” The role of sociocultural inputs in the development of children's learning is largely missing from Lake et al.'s discussion of the development of human-like artificial intelligence, but its inclusion would help move research even closer to machines that can learn and think like humans.