This paper proposes how various disciplinary theories of cognition may be combined into a unifying, sub-symbolic, computational theory of cognition. The following theories are considered for integration: psychological theories, including the theory of event coding, event segmentation theory, the theory of anticipatory behavioral control, and concept development; artificial intelligence and machine learning theories, including reinforcement learning and generative artificial neural networks; and theories from theoretical and computational neuroscience, including predictive coding and free energy-based inference. In the light of such a (...) potential unification, it is discussed how abstract cognitive, conceptualized knowledge and understanding may be learned from actively gathered sensorimotor experiences. The unification rests on the free energy-based inference principle, which essentially implies that the brain builds a predictive, generative model of its environment. Neural activity-oriented inference causes the continuous adaptation of the currently active predictive encodings. Neural structure-oriented inference causes the longer term adaptation of the developing generative model as a whole. Finally, active inference strives for maintaining internal homeostasis, causing goal-directed motor behavior. To learn abstract, hierarchical encodings, however, it is proposed that free energy-based inference needs to be enhanced with structural priors, which bias cognitive development towards the formation of particular, behaviorally suitable encoding structures. As a result, it is hypothesized how abstract concepts can develop from, and thus how they are structured by and grounded in, sensorimotor experiences. Moreover, it is sketched-out how symbol-like thought can be generated by a temporarily active set of predictive encodings, which constitute a distributed neural attractor in the form of an interactive free-energy minimum. The activated, interactive network attractor essentially characterizes the semantics of a concept or a concept composition, such as an actual or imagined situation in our environment. Temporal successions of attractors then encode unfolding semantics, which may be generated by a behavioral or mental interaction with an actual or imagined situation in our environment. Implications, further predictions, possible verification and falsifications, as well as potential enhancements into a fully spelled-out unified theory of cognition are discussed at the end of the paper. (shrink)

Butz, Achimova, Bilkey, and Knott provide a topic overview and discuss whether the special issue contributions may imply that event‐predictive abilities constitute a root for conceptual human thought, because they enable complex, mutually beneficial, but also intricately competitive, social interactions and language communication.

Butz, Achimova, Bilkey, and Knott provide a topic overview and discuss whether the special issue contributions may imply that event‐predictive abilities constitute a root for conceptual human thought, because they enable complex, mutually beneficial, but also intricately competitive, social interactions and language communication.

Topics in Cognitive Science, Volume 13, Issue 1, Page 128-141, January 2021.

Events—the experiences we think we are having and recall having had—are constructed; they are not what actually occurs. What occurs is ongoing dynamic, multidimensional, sensory flow, which is somehow transformed via psychological processes into structured, describable, memorable units of experience. But what is the nature of the redescription processes that fluently render dynamic sensory streams as event representations? How do such processes cope with the ubiquitous novelty and variability that characterize sensory experience? How are event‐rendering skills acquired and how do (...) event representations change with development? This review considers emerging answers to these questions, beginning with evidence that an implicit tendency to monitor predictability structure via statistical learning is key to event rendering. That is, one way that the experience of bounded events (e.g., actions within behavior, words within speech) arises is with the detection of “troughs” in sensory predictability. Interestingly, such troughs in predictability are often predictable; these regions of predictable‐unpredictability provide articulation points to demarcate one event from another in representations derived from the actual streaming information. In our information‐optimization account, a fluent event‐processor predicts such troughs and selectively attends to them—while suppressing attention to other regions—as sensory streams unfold. In this way, usage of attentional resources is optimized for efficient sampling of the most relevant, information‐rich portions of the unfolding flow of sensation. Such findings point to the development of event‐processing fluency—whether in action, language, or other domains—depending crucially on rapid and continual cognitive reorganization. As knowledge of predictability grows, attention is adaptively redeployed. Accordingly, event experiences undergo continuous alteration. (shrink)

Warren and Dresang comment on the contributions from a psycholinguistic perspective, highlighting close relations between the respective research on events and proposing that, for example, verbs may indeed directly pre‐activate templates of the typically involved event participants.

Storchak, Ehlis, and Fallgatter provide an extensive literature review on electrophysiological measurements, which indicate that general predictive deficits in self‐monitoring are associated with various positive symptoms in patients with schizophrenia.

Stawarczyk, Bezdek, and Zacks offer neuroscience evidence for a midline default network core, which appears to coordinate internal, top‐down mentation with externally‐triggered, bottom‐up attention in a push‐pull relationship. The network may enable the flexible pursuance of thoughts tuned into or detached from the current environment.

Shin and DuBrow propose that a key principle driving event segmentation relates to causal analyses: specifically, that experiences that are attributed as having the same underlying cause are grouped together into an event. This offers an alternative to accounts of segmentation based on prediction error.

Paulin and Cahill‐Lane explore the origins of event processing and event prediction in animal evolution. They propose that the evolutionary benefit of being able to predict and thus to quickly react to anticipated events may have triggered the evolution of the earliest nervous systems.

A fundamental aspect of human cognition is the ability to parse our constantly unfolding experience into meaningful representations of dynamic events and to communicate about these events with others. How do we communicate about events we have experienced? Influential theories of language production assume that the formulation and articulation of a linguistic message is preceded by preverbal apprehension that captures core aspects of the event. Yet the nature of these preverbal event representations and the way they are mapped onto language (...) are currently not well understood. Here, we review recent evidence on the link between event conceptualization and language, focusing on two core aspects of event representation: event roles and event boundaries. Empirical evidence in both domains shows that the cognitive representation of events aligns with the way these aspects of events are encoded in language, providing support for the presence of deep homologies between linguistic and cognitive event structure. (shrink)

McRae, Brown and Elman argue against the view that events are structured as frequently‐occurring sequences of world stimuli. They underline the importance of temporal structure defining event types and advance a more complex temporal structure, which allows for some variance in the component elements.

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 build causal models of the world that support explanation and understanding, rather than merely solving pattern recognition problems; ground learning in intuitive theories of physics and psychology to support and enrich the knowledge that is learned; and 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. (shrink)

Inspired by, and in close relation with, the contributions of this special issue, Kuperberg elegantly links event comprehension, production, and learning. She proposes an overarching hierarchical generative framework of processing events enabling us to make sense of the world around us and to interact with it in a competent manner.

Topics in Cognitive Science, Volume 13, Issue 1, Page 187-205, January 2021.

Traditional approaches to human information processing tend to deal with perception and action planning in isolation, so that an adequate account of the perception-action interface is still missing. On the perceptual side, the dominant cognitive view largely underestimates, and thus fails to account for, the impact of action-related processes on both the processing of perceptual information and on perceptual learning. On the action side, most approaches conceive of action planning as a mere continuation of stimulus processing, thus failing to account (...) for the goal-directedness of even the simplest reaction in an experimental task. We propose a new framework for a more adequate theoretical treatment of perception and action planning, in which perceptual contents and action plans are coded in a common representational medium by feature codes with distal reference. Perceived events (perceptions) and to-be-produced events (actions) are equally represented by integrated, task-tuned networks of feature codes – cognitive structures we call event codes. We give an overview of evidence from a wide variety of empirical domains, such as spatial stimulus-response compatibility, sensorimotor synchronization, and ideomotor action, showing that our main assumptions are well supported by the data. Key Words: action planning; binding; common coding; event coding; feature integration; perception; perception-action interface. (shrink)

Events and event prediction are pivotal concepts across much of cognitive science, as demonstrated by the papers in this special issue. We first discuss how the study of events and the predictive processing framework may fruitfully inform each other. We then briefly point to some links to broader philosophical questions about events.

Topics in Cognitive Science, Volume 13, Issue 1, Page 243-247, January 2021.

Looking times and gaze behavior indicate that infants can predict the goal state of an observed simple action event (e.g., object‐directed grasping) already in the first year of life. The present paper mainly focuses on infants’ predictive gaze‐shifts toward the goal of an ongoing action. For this, infants need to generate a forward model of the to‐be‐obtained goal state and to disengage their gaze from the moving agent at a time when information about the action event is still incomplete. By (...) about 6 months of age, infants show goal‐predictive gaze‐shifts, but mainly for familiar actions that they can perform themselves (e.g., grasping) and for familiar agents (e.g., a human hand). Therefore, some theoretical models have highlighted close relations between infants’ ability for action‐goal prediction and their motor development and/or emerging action experience. Recent research indicates that infants can also predict action goals of familiar simple actions performed by non‐human agents (e.g., object‐directed grasping by a mechanical claw) when these agents display agency cues, such as self‐propelled movement, equifinality of goal approach, or production of a salient action effect. This paper provides a review on relevant findings and theoretical models, and proposes that the impacts of action experience and of agency cues can be explained from an action‐event perspective. In particular, infants’ goal‐predictive gaze‐shifts are seen as resulting from an interplay between bottom‐up processing of perceptual information and top‐down influences exerted by event schemata that store information about previously executed or observed actions. (shrink)

Time underlies many interesting human behaviors. Thus, the question of how to represent time in connectionist models is very important. One approach is to represent time implicitly by its effects on processing rather than explicitly (as in a spatial representation). The current report develops a proposal along these lines first described by Jordan (1986) which involves the use of recurrent links in order to provide networks with a dynamic memory. In this approach, hidden unit patterns are fed back to themselves: (...) the internal representations which develop thus reflect task demands in the context of prior internal states. A set of simulations is reported which range from relatively simple problems (temporal version of XOR) to discovering syntactic/semantic features for words. The networks are able to learn interesting internal representations which incorporate task demands with memory demands: indeed, in this approach the notion of memory is inextricably bound up with task processing. These representations reveal a rich structure, which allows them to be highly context‐dependent, while also expressing generalizations across classes of items. These representations suggest a method for representing lexical categories and the type/token distinction. (shrink)

Purpose: Constructivism postulates that the perceived reality is a complex construct formed during development. Depending on the particular school, these inner constructs take on different forms and structures and affect cognition in different ways. The purpose of this article is to address the questions of how and, even more importantly, why we form such inner constructs. Approach: This article proposes that brain development is controlled by an inherent anticipatory drive, which biases learning towards the formation of forward predictive structures and (...) inverse goal-oriented control structures. This drive, in combination with increasingly complex environmental interactions during cognitive development, enforces the structuring of our conscious self, which is embedded in a constructed inner reality. Essentially, the following questions are addressed: Which basic mechanisms lead us to the construction of inner realities? How are these emergent inner realities structured? How is the self represented within the inner realities? And consequently, which cognitive structures constitute the media for conscious thought and selfconsciousness? Findings: Due to the anticipatory drive, representations in the brain shape themselves predominantly purposefully or intentionally. Taking a developmental, evolutionary perspective, we show how the brain is forced to develop progressively complex and abstract representations of the self embedded in the constructed inner realities. These self representations can evoke different stages of self-consciousness. Implications: The anticipatory drive shapes brain structures and cognition during the development of progressively more complex, competent, and flexible goal-oriented bodyenvironment interactions. Self-consciousness develops because increasingly abstract, individualizing self representations are necessary to realize these progressively more challenging environmental interactions. (shrink)