Metaphors in science help to understand a phenomenon and serve as tools for scientific development. The question I address is why researchers refuse to abandon the computer metaphor. By using the case within my field of expertise, movement sciences, I make the case that it is because there are no other metaphors at this level of simplicity and fruitfulness. If one considers the issue of pragmatism and productivism in science (despite their potential dangers for scientific development), the simplest must be considered the best. To replace the computer metaphor and avoid its pitfalls, we must reach the same level of simplicity and understand how to intervene for change. I offer another metaphor that has already demonstrated its power in other domains – the one of – and discuss potential ways to endorse it in science. Considering the brain as a landscape, one can identify the current state of the brain as a ball being pushed through mountains and valleys. Such a metaphor is immensely useful as it offers visualization for understanding various perception-action phenomena and can be implemented as a tool for scientific inquiry. Interestingly, as I present, there is room to believe that the metaphor (and its related concepts) might represent the natural point of convergence of the computationalists.

The computer metaphor derives from two parallel histories of thought, the first questioning the operation and activity of minds and brains, and the second aiming at finding a minimal model of the mind and brain. These histories informed two hypotheses of mind: mind as computation and brain as machine, respectively. We focus on the latter, termed the mechanistic hypothesis. First, we briefly review how the brain has historically been discussed with machine metaphors and identify five tenets that define a machine. We review findings in neuroscience that motivate that the brain demonstrates exceptions to these tenets and, thus, should not be considered a machine. We offer that an alternative classification may be found in far-from-equilibrium self-organizing systems known as dissipative structures. We review the properties of these systems and suggest that the brain is more like a dissipative structure than a machine. Ifbrains do not fit the mechanistic hypothesis underwriting the computer metaphor, then the cognitive sciences may need to seek alternative metaphors based on the assumption that minds and brains are some other kind of natural systems, namely dissipative structures.

The computer metaphor of mind and brain is unviable as an explanatory vehicle for the complex adaptive behavior ofliving systems. I present anomalies in the empirical record of the neurosciences that expose profound problems with the assumption that the brain performs machine-like computations. As an alternative, I suggest that evolved agents’ adaptive coordination of behavior is based on a massive redundancy of reality, instead of a massive modularity of mind. A research program of Radical Embodied Computation based on contemporary theories of physical information and natural computation takes the reproduction of similarity by analogy as a fundamental process in the generation of complex order. Finally, I provide empirical evidence for the ability of all evolved agents, including those without a nervous system, to exploit the massive redundancy of reality to coordinate their behavior.

To comprehend brains and minds as computers is to explain their activities and organization in terms of information processing, which is essentially the systematic manipulation of representations. This understanding has served as the fundamental guiding commitment of research in artificial intelligence, cognitive psychology, neuroscience, and related disciplines. While computational terminology has been applied metaphorically to brains and minds, it has certainly been applied in literal senses as well. The Human Brain Project’s SpiNNaker computer is presented as an example of the inability of research committed to the computer metaphor (and literalism) to facilitate understanding of brains and minds. Long-term progress requires shifting to investigative frameworks that approach brains and minds on their own terms, and not via inappropriate metaphors. Complexity science is such a framework, in which brains and minds are understood via core features of complex systems: emergence, nonlinearity, self-organization, and universality. Extended Haken–Kelso–Bunz models are presented as empirically supported research exhibiting a fruitful complexity science approach to brains and minds across spatial and temporal scales ofactivity and organization. The result is a clear case for the effectiveness of investigating brains and minds on their own terms as complex systems.

We argue that all computational processes require data that must be received, represented, and processed. The inherent ambiguity in these processes is incompatible with a lawful explanation of psychological phenomena, in particular the successful performance of everyday goal-directed behaviors in organisms all at levels of taxonomic scale. We argue that, as scientific devices, metaphors are best used a posteriori to generate testable hypotheses about a well-developed theory. Therefore, we develop a metaphor for mind and brain in the context of the ecological approach to perceiving, acting, and cognizing – an approach in which the successful performance of everyday behavior is a lawful process of detecting and exploiting lawful relations. We propose that in this context, the brain could be understood as a fractal antenna. That is, it could aid in the detecting and exploiting of lawful relations at multiple nested levels, without generating, modifying, or interpreting such lawful relations.

We see the computer metaphor as a holdover from premodern scientific traditions hoping to anchor the mind’s computational ability in a material anatomical part. The computer metaphor has likely outgrown its usefulness despite prompting decades of valuable empirical insights. Brains are context-sensitive and capable of adapting to novelty, eschewing the locality of meaning necessary to computation. Rather than proposing a new metaphor for the mind and brain, we think that Turing’s old idea of cascading instability is a metaphor worth reconsidering. It recommends a power law-driven geometrical framework that might knit together the mind, brain, and behavior in context.

Nearly a century of cognition research has relied on a metaphor of the mind as a computing machine, suggesting that the problem of cognition is calculations over various impoverished and imperfect inputs that are enriched by the brain and translated into output that will be sent to the body for action. Over the past several decades, the structure of this hypothesized biological “computing machine” has become more complicated as findings have repeatedly underscored the fundamental role of the body and environment in shaping cognition. In this chapter, we argue that the brain-as-computer metaphor has outlived its utility and point to the fundamental interconnectedness of social systems as a critical weakness. Rather than seeing physical and social dynamics as ancillary information that must be extracted from the environment and enriched by individual social actors, we argue that these richly nested structures are fundamental to cognition as a coconstruction of the social actors, their interactions among one another, and the social-physical-temporal setting in which they interact. We point to evidence in human and nonhuman groups to support our claims, focusing on eusocial colonies (the most cooperative kind of collective, most canonically associated with honey bees and ants). Although all metaphors are imperfect, we propose the eusocial colony metaphor as an alternative to the computer to emphasize better the interdependence, embeddedness, emergence, and adaptability of the mind and brain.

All variations of the computer metaphor are based on the premise that the principal function of the brain is to process information and, thus, that brain functions can be subdivided into information-processing functions. However, neurophysiological data strongly argue against such subdivisions, rejecting classical concepts of computational modules. Instead, I suggest the brain be viewed from the perspective of the process that produced it – evolution. This leads to the proposal, made many times over the past hundred years, that the brain is a control system. I briefly discuss evolutionary and neurobiological data that support this view and lead toward a functional architecture for the brain consisting of parallel and nested sensorimotor control loops. I suggest that this architecture is more compatible with the process of brain evolution and comparative data across diverse species, that it better explains a wide range of neurophysiological findings, and that it provides a better conceptual taxonomy for understanding behavior.

The computer metaphor of mind and brain states broadly that the brain is the control organ for the body. This implies that the brain (including the mind) and the physical body are separable from each other and the physical and social environment. Given the dominance of computing technology in daily life, many brain researchers and engineers are compelled to take brain-computer analogies not as metaphors but as literal descriptions of the brain function. These two fundamental assumptions manifest as overwhelming challenges when pursuing synthetic rather than analytic approaches, that is when we attempt to control artificial bodies such as robots computationally, especially when colocated with humans. I will discuss the computational brain metaphor from the perspective of bodies for whom computational control is a reality - robots and their creators, engineers. Rather than presenting new metaphors, I will use evidence from control engineering and human-robot interaction to argue for a shift of thought. If we can enrich how engineers approach robotic control, new robots could offer powerful momentum to shifting the scientific opinion toward embracing a less dualistic, more holistic view of the brain’s embedding in body and world.

In this chapter, we want to take a reflexive stance on the role of language in scientific practice, and we want to highlight some of the shortcomings of the computer metaphor of mind and brain by analyzing how an algorithmic conception of language fails to provide a reliable account of meaning in language. Our argument is premised on the conviction that cognitive science is a selfreferential endeavor in the sense that cognitive scientists make use of their cognitive capacities to come up with theories of how cognition may work (Sanches de Oliveira, 2023; Varela et al., 2017, pp. 3–14). In other words, and in line with a naturalist stance in the philosophy of science, we hold that any satisfactory account of cognition needs to be capable of explaining how human cognizers can gain scientific insights by virtue of using their cognitive capacities as scientists (see Giere, 1987; Sanches de Oliveira, 2023; Thagard, 2012). Language is central in this respect because scientific practice – here broadly understood as a systematic procedure of gaining insights by collective cognitive practices distributed across many scientists (see Giere and Moffatt, 2003; Nersessian, 2004; Sanches de Oliveira et al., 2023) – would not work without an effective means of communication, mutual understanding, and expressing agreement or disagreement. By drawing on Gödel’s (1931) incompleteness theorem and Tarski’s (1936) undefinability theorem, we show that a computational, algorithmic account of language does not provide a satisfactory explanation of meaning in language that fulfills our criteria to support and explain scientific practice effectively. Rather than proposing a single alternative metaphor that shall replace the computer metaphor, we outline some metatheoretical requirements for post-computational and post-cognitivist approaches in cognitive science. We conclude, perhaps counterintuitively, that alternative theoretical frameworks in cognitive science need to limit themselves systematically in their explanatory aspirations to avoid falling prey to the same pitfalls as the computational theory of mind does.