What do artificial intelligence (AI) systems "understand"? This question arises not only in assessing a system's intelligence but also in evaluation practices to ensure the safe and responsible deployment of AI. Drawing on scholarship from philosophy and cognitive science, and informed by current practices in AI, we develop a framework for asking more precise questions and making more precise claims about machine understanding. We conceptualize understanding as a relation between a system (S) and a target of understanding (T), and we discuss how to specify the relation, the system, and the target, offering a landscape of options in each case. Our goal is not to defend a particular account of understanding, but to provide conceptual tools for those working to assess or advance machine understanding.

From petitionary prayers to pilgrimages, rituals are found in every known culture. Yet, the reason for their persistence is a matter of active debate. Some studies portray rituals as attempts to affect uncertain outcomes, whereas others emphasize their role in facilitating social cohesion. We review the cognitive processes underlying both perspectives and draw on advances in reinforcement learning to integrate them. Specifically, ritual participation is motivated by two processes: habitual reinforcement of affective and social rewards experienced during performance (model-free learning) and reinforcement of pragmatic and cooperative benefits derived from culturally shared world models (model-based learning). This framework synthesizes previous accounts and illuminates ritual's role in sustaining intersubjectively aligned world models in past and present societies.

Human brain architecture is guiding brain-inspired artificial intelligence (AI) and has been treated as an optimal template, whose deviations could mark different psychiatric and neurological conditions. We argue this premise is wrong: under any single goal (e.g., minimal wiring cost or maximal communication efficiency), the human connectome is suboptimal. Instead, its organization reflects multi-objective trade-offs navigated over evolution and development under biological and environmental constraints. For psychopathology, atypical trajectories are not distances from an ideal brain but reweighted compromises in the same trade-off space. For neuro-AI, directly duplicating the brain's connectivity risks copying its irrelevant compromises. Treating brains and models as products of multi-objective optimization and co-tuning relevant objectives offers a more powerful framework for interpreting clinical phenotypes and designing next-generation AI.

Recent work by Siems et al. shows that the brain rhythmically samples competing alternatives through covert spatial attention. This challenges continuous models of decision-making and suggests that evaluation is temporally structured by oscillatory dynamics, with attention determining when alternatives are accessed rather than reflecting changes in their representations.

Individuals high in Openness to Experience are more likely to enjoy imaginary worlds, visit museums, be vaccinated, support redistribution, endorse animal rights, and volunteer, alongside a wide array of other characteristic behaviors. Although these behaviors appear disparate, we propose that they form a coherent constellation grounded in a tendency to explore the unknown and tolerate uncertainty. This clarifies why levels of Openness vary across individuals, groups, and historical periods: exploration is modulated by ecological conditions such as safety and resource availability. When environments make exploration worthwhile, Openness expands; when conditions are harsh or precarious, it contracts. Consequently, ecological differences should generate systematic differences in cultural forms. This framework explains how ecological conditions shape culture through their effects on personality.

Behavioral tasks often show robust group effects yet unreliable person-level estimates. Many 'reliability fixes' yield mixed results because they rarely change the two determinants of person-level reliability: trials per person and per-trial informativeness. Recent evidence supports this two-lever view and motivates routine reliability auditing and measurement engineering for translational use.

To prioritize the visual processing of task-relevant objects in our surroundings, we rely on an attentional template-an internal representation of object features that guides attention toward potential targets. Decades of research have characterized attentional templates for simple targets in artificial arrays. How could templates function in real-world search, where target appearance is variable and objects are embedded in complex, dynamic scenes? We consider two possibilities: (i) flexible templates that are adapted to changing scene contexts and (ii) stable ('one-size-fits-all') templates that generalize across contexts. We review recent behavioral and neuroimaging evidence for both possibilities and discuss how optimal search depends on balancing the relative costs and benefits of template adaptation, enabling efficient attention 'in the wild'.

The assumption that attention and short-term memory combine to play a crucial role in cognition continues to influence cognitive modeling. We trace the development of the multicomponent model of working memory, initially consisting of a limited-capacity central executive controlling two domain-specific systems: the phonological loop and the visuospatial sketchpad. The later introduction of the episodic buffer addressed the need to explain how information from different sources is bound into unified episodes. Subsequent developments suggest that the buffer functions as a consciously accessible, multidimensional interface combining storage with executive manipulation and attentional control. These developments lead to a reformulation of the model in which the episodic buffer, renamed the 'awareness buffer', is placed at the focal point of working memory.

Human multimodal processing abilities have provided researchers with an invaluable set of methods for interrogating language understanding. Even young infants fixate on visual stimuli that match incoming auditory information. Experimental paradigms have harnessed this behavior to demonstrate early language comprehension abilities. Researchers have since adapted these paradigms to address new questions, such as studying individual differences in vocabulary size and structure, identifying which words are learned earlier or later, and assessing language in populations with disabilities. However, fundamental questions persist about the assumptions linking eye gaze with underlying linguistic competence. We aim to articulate these assumptions and outline what we know about whether they are met. By making these issues explicit, we highlight considerations for language development research across different populations.

Developmental perspectives are crucial to harnessing the human potential for altruism. We synthesise and highlight recent research on the motivations underlying children's (costly) prosocial behaviour. The order in which children develop intrinsic, extrinsic, and strategic motivations to benefit others has implications for efforts to positively impact young people's lives.

The syntax of human languages has long been argued to be complex and even unlearnable from the input alone. However, the success of large language models (LLMs) has challenged this idea. I argue for a simple view of syntax, where the syntax of a language is just the set of dependency rules, with no phrase structure or transformation rules-constructs central to Chomsky's transformational grammar. This approach accounts for diverse phenomena in human language processing and explains crosslinguistic word order universals. Moreover, it better explains human data for cases that differentiate these accounts and eliminates the syntax learnability problem. I speculate that LLMs, similar to children, learn the dependency grammar from linguistic patterns, leading to their impressive syntactic competence.

Some of the strongest evidence that human minds should be thought of in terms of symbolic systems has been the way they combine ideas, produce novelty, and learn quickly. We argue that modern neural networks-and the artificial intelligence systems built upon them-exhibit similar abilities. This potentially undermines the argument that the cognitive processes and representations used by human minds are symbolic. We consider possible interpretations of these results-that modern neural networks implement symbolic systems, or that they approximate them subsymbolically-and the theoretical consequences of these two possibilities for explanations of human cognition at different levels of analysis. This consideration leads us to offer a new agenda for research on the symbolic basis of the mind.

Perceptual multistability, observed across species and sensory modalities, offers valuable insights into numerous cognitive functions and dysfunctions. For instance, differences in temporal dynamics and information integration during percept formation often distinguish clinical from nonclinical populations. Computational psychiatry can elucidate these variations through two primary approaches: (i) Bayesian modeling, which treats perception as an unconscious inference, and (ii) an active, information-seeking perspective (e.g., reinforcement learning), which frames perceptual switches as internal actions. Our synthesis aims to leverage multistability to bridge these computational psychiatry subfields, linking human and animal studies as well as connecting behavior to underlying neural mechanisms. Perceptual multistability emerges as a promising noninvasive tool for clinical applications, facilitating translational research and enhancing our mechanistic understanding of cognitive processes and their impairments.

Contemporary consciousness science faces an impasse: competing theoretical frameworks-structuralist versus functionalist, universal versus local, intrinsic versus extrinsic-appear to be inducing philosophical deadlocks and conceptual standstills. While these debates have generated valuable insights, they have proceeded in parallel, without a systematic framework for understanding their relationships and implications. We contend that these parallel disputes reflect deeper, unresolved tensions in the conceptualization of consciousness. These debates can be addressed by recognizing three fundamental dimensions that encompass models of consciousness at a meta-theoretical level: explanatory medium, scope, and perspective. We show the interplay among different dimensions and how they mutually condition evidence, language, and concepts, reframing theoretical conflicts into tractable disagreements.