A New Approach To Predictive Processing Through Decolonial Anthropology
Arthur Nicolas
Contemporary neuroscience is the dominant framework for understanding how our brains perceive reality. However, some scholars have argued that it reflects historically Western-centered assumptions, which have led to a focus on Western scientific paradigms and practices and may underrepresent how epistemic variation relates to cognition. This paper brings together decolonial anthropology and predictive processing to propose a novel framework for understanding mental-behavioral relationships.
In Neuroscience, the theory of predictive processing has posed our understanding of the brain as a ‘prediction machine.’ It presents that the brain's main function is to ‘infer’ the causes of its sensory stimulation. In “Making up the Mind: How the Brain Creates Our Mental World”, Chris Frith (2007) proposed that the brain constructs our perception by combining incoming sensory information with prior knowledge to predict what will occur next. Within this framework, perception is understood not as a direct interface with reality but instead as a probabilistic construction generated by the brain. The most common notation of predictive processing is Karl Friston's formula, P(o|s) and P(s) from “The free-energy principle: a unified brain theory”(2010, p. 128).
Despite its explanatory power, predictive processing often assumes that the structure of priors and prediction schemas is broadly universal (Litwin & Miłkowski, 2020) and has been critiqued for its grounding in Western scientific paradigms (Henrich et al., 2010; Gordon, 2022). This grounding has been challenged by cultural research that demonstrates that cognitive processes and knowledge systems differ significantly across cultures (Muthukrishna et al., 2020). Ultimately, this suggests that predictive processing may overlook alternative models of prediction that could broaden its scope as a theory of cognition.
Decolonial anthropology is a critical orientation of the study that draws on decolonial theory to challenge Western assumptions about how knowledge is produced (Bolles, 2023). Anthropologist Philippe Descola’s work, Beyond Nature and Culture, provides a critical foundation for understanding how epistemological differences translate into varying prediction structures. Through analysis of a set of cultures, Descola argues that societies organize the space of possibility according to different modes of identification structured by interiority and physicality, which leads to cultures having unique relationships with the world around them. Effectively, these ontological frameworks shape constraints on perceptive experience. This implies that the set of variables, relationships, and expected patterns used for predictive comparison is different across cultures. As a result, predictive processes are not only shaped by beliefs flowing through a fixed system, but are instead structured by distinct generative schemas. Through analysis of Descola’s work, ontologies are understood as higher-order models of prediction, which not only shape the priors of the system but also the form that predictive reasoning takes.
Similarly, in learner psychology literature, research reflects this concept. In “Reconstructing constructivism: Causal models, Bayesian learning mechanisms and the theory theory”, Alison Gopnik and Henry M. Wellman (2012) argue that learners construct theoretical structures and revise them in response to new evidence, implying that the schema used to generate predictions is learned rather than operating from a single inferential structure. This suggests that different epistemic frameworks can correspond to structurally distinct models of prediction.
Ultimately, this challenges standard forms of predictive processing (Clark, 2015; Friston, 2010), which state that people's differences come from their learning history and priors, and operate within the same framework. In fact, the typical prediction processing formula reduces these variations into parameter differences, which limits the ability of predictive processing to account for vast differences in ontological structures. Although predictive processing doesn’t explicitly deny ontological plurality, it does not clearly specify how to represent it.
Proponents argue that the current model already accounts for ontological variation sufficiently through social expectations and weighted inference (Ramstead et al., 2016). From this perspective, people with different experiences generate distinct predictions while still operating within the model. This position may be compelling given its flexibility; however, it makes the assumption that all people share the same architecture. If, as decolonial anthropology and learner psychology suggest, ontological differences extend to the structural level, then they cannot be accounted for as differences in priors alone. Even within predictive processing literature, some authors have raised concern over the current model’s ability to adapt to differences in structure (Rutar et al., 2022), arguing that the canonical formulation is better suited to parameter-level variation as opposed to structural revision of generative models. Because predictive processing has such great explanatory power, it must make proper assumptions about the underlying structures of cognition. When predictive processing is operationalized in fields of psychology, it's used to model how internal cognitive structures shape visible behavior, for example, stress responses (Krupnik, 2020). It would be one thing if prediction were used solely to explain causal relationships or to make short-term model predictions, but when reconstructing the functions of the brain with relevance to behavior, it becomes necessary for predictive processing to explain actions outside of simple input-output mapping.
However, by critiquing and integrating neuroscience with decolonial anthropology, we can alter the predictive processing model to allow for structurally distinct cognitive processes, thereby improving behavioral predictions. For example, the inclusion of ontological frameworks would let us model decisions that appear irrational under standard prediction processing models, such as prioritizing communal or environmental obligations, which are easily predicted within a relational framework like Indigenous psychologies (Lorencova & Trnka, 2023). By incorporating epistemological assumptions into the model, we can more accurately predict behaviors that would otherwise be misrepresented. In this way, decolonial anthropology increases the empirical accuracy of behavior prediction by aligning the model with the underlying structures that shape decision-making and cognition.
By integrating decolonial anthropology with predictive processing, this paper proposes a new framework for understanding mental-behavioral relationships across cultures. It poses that ontologies shape the schemas of prediction by shaping the space of possibilities available to the mind, and in turn shaping the structural function that each relationship serves. Predictive processing models the most accurate form of the mind when it accounts for epistemic and ontological variation.
About the Author
Arthur Nicolas is a student at Skiatook High School.
References
In Neuroscience, the theory of predictive processing has posed our understanding of the brain as a ‘prediction machine.’ It presents that the brain's main function is to ‘infer’ the causes of its sensory stimulation. In “Making up the Mind: How the Brain Creates Our Mental World”, Chris Frith (2007) proposed that the brain constructs our perception by combining incoming sensory information with prior knowledge to predict what will occur next. Within this framework, perception is understood not as a direct interface with reality but instead as a probabilistic construction generated by the brain. The most common notation of predictive processing is Karl Friston's formula, P(o|s) and P(s) from “The free-energy principle: a unified brain theory”(2010, p. 128).
Despite its explanatory power, predictive processing often assumes that the structure of priors and prediction schemas is broadly universal (Litwin & Miłkowski, 2020) and has been critiqued for its grounding in Western scientific paradigms (Henrich et al., 2010; Gordon, 2022). This grounding has been challenged by cultural research that demonstrates that cognitive processes and knowledge systems differ significantly across cultures (Muthukrishna et al., 2020). Ultimately, this suggests that predictive processing may overlook alternative models of prediction that could broaden its scope as a theory of cognition.
Decolonial anthropology is a critical orientation of the study that draws on decolonial theory to challenge Western assumptions about how knowledge is produced (Bolles, 2023). Anthropologist Philippe Descola’s work, Beyond Nature and Culture, provides a critical foundation for understanding how epistemological differences translate into varying prediction structures. Through analysis of a set of cultures, Descola argues that societies organize the space of possibility according to different modes of identification structured by interiority and physicality, which leads to cultures having unique relationships with the world around them. Effectively, these ontological frameworks shape constraints on perceptive experience. This implies that the set of variables, relationships, and expected patterns used for predictive comparison is different across cultures. As a result, predictive processes are not only shaped by beliefs flowing through a fixed system, but are instead structured by distinct generative schemas. Through analysis of Descola’s work, ontologies are understood as higher-order models of prediction, which not only shape the priors of the system but also the form that predictive reasoning takes.
Similarly, in learner psychology literature, research reflects this concept. In “Reconstructing constructivism: Causal models, Bayesian learning mechanisms and the theory theory”, Alison Gopnik and Henry M. Wellman (2012) argue that learners construct theoretical structures and revise them in response to new evidence, implying that the schema used to generate predictions is learned rather than operating from a single inferential structure. This suggests that different epistemic frameworks can correspond to structurally distinct models of prediction.
Ultimately, this challenges standard forms of predictive processing (Clark, 2015; Friston, 2010), which state that people's differences come from their learning history and priors, and operate within the same framework. In fact, the typical prediction processing formula reduces these variations into parameter differences, which limits the ability of predictive processing to account for vast differences in ontological structures. Although predictive processing doesn’t explicitly deny ontological plurality, it does not clearly specify how to represent it.
Proponents argue that the current model already accounts for ontological variation sufficiently through social expectations and weighted inference (Ramstead et al., 2016). From this perspective, people with different experiences generate distinct predictions while still operating within the model. This position may be compelling given its flexibility; however, it makes the assumption that all people share the same architecture. If, as decolonial anthropology and learner psychology suggest, ontological differences extend to the structural level, then they cannot be accounted for as differences in priors alone. Even within predictive processing literature, some authors have raised concern over the current model’s ability to adapt to differences in structure (Rutar et al., 2022), arguing that the canonical formulation is better suited to parameter-level variation as opposed to structural revision of generative models. Because predictive processing has such great explanatory power, it must make proper assumptions about the underlying structures of cognition. When predictive processing is operationalized in fields of psychology, it's used to model how internal cognitive structures shape visible behavior, for example, stress responses (Krupnik, 2020). It would be one thing if prediction were used solely to explain causal relationships or to make short-term model predictions, but when reconstructing the functions of the brain with relevance to behavior, it becomes necessary for predictive processing to explain actions outside of simple input-output mapping.
However, by critiquing and integrating neuroscience with decolonial anthropology, we can alter the predictive processing model to allow for structurally distinct cognitive processes, thereby improving behavioral predictions. For example, the inclusion of ontological frameworks would let us model decisions that appear irrational under standard prediction processing models, such as prioritizing communal or environmental obligations, which are easily predicted within a relational framework like Indigenous psychologies (Lorencova & Trnka, 2023). By incorporating epistemological assumptions into the model, we can more accurately predict behaviors that would otherwise be misrepresented. In this way, decolonial anthropology increases the empirical accuracy of behavior prediction by aligning the model with the underlying structures that shape decision-making and cognition.
By integrating decolonial anthropology with predictive processing, this paper proposes a new framework for understanding mental-behavioral relationships across cultures. It poses that ontologies shape the schemas of prediction by shaping the space of possibilities available to the mind, and in turn shaping the structural function that each relationship serves. Predictive processing models the most accurate form of the mind when it accounts for epistemic and ontological variation.
About the Author
Arthur Nicolas is a student at Skiatook High School.
References
- Bolles, A. L. (2023). Decolonizing anthropology. American Ethnologist, 50(3), 519–522. https://doi.org/10.1111/amet.13199
- Clark, A. (2015). Surfing Uncertainty. In Oxford University Press eBooks. https://doi.org/10.1093/acprof:oso/9780190217013.001.0001
- Descola, P. (2013, August 1). Beyond Nature and Culture. The University of Chicago Press. https://press.uchicago.edu/ucp/books/book/chicago/B/bo9826233.html
- Friston, K. (2010). The free-energy principle: a unified brain theory? In Nature Reviews Neuroscience (Vol. 11, Issue 2, p. 127). https://doi.org/10.1038/nrn2787
- Gopnik, A., & Wellman, H. M. (2012). Reconstructing constructivism: Causal models, Bayesian learning mechanisms, and the theory theory. Psychological Bulletin, 138(6), 1085–1108. https://doi.org/10.1037/a0028044
- Gordon, S. (2022). Philosophy of Western Science and the Mind-Brain Problem. In The Mind-Brain Continuum (pp. 3–16). https://doi.org/10.1007/978-3-031-10059-8_1
- Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The weirdest people in the world? Behavioral and Brain Sciences, 33(2–3), 61–83. https://doi.org/10.1017/s0140525x0999152x
- Krupnik, V. (2020). Trauma or Drama: A Predictive Processing Perspective on the Continuum of Stress. Frontiers in Psychology, 11, 1248. https://doi.org/10.3389/fpsyg.2020.01248
- Litwin, P., & Miłkowski, M. (2020). Unification by Fiat: Arrested Development of Predictive Processing. Cognitive Science, 44(7), e12867. https://doi.org/10.1111/cogs.12867
- Lorencova, R., & Trnka, R. (2023). Variability in Cultural Understandings of Consciousness: A Call for Dialogue with Native Psychologies. Journal of Consciousness Studies, 30(5), 232–254. https://doi.org/10.53765/20512201.30.5.232
- Frith. (2007, May 29). Making up the mind: How the brain creates our mental world. APA Psycnet. https://psycnet.apa.org/record/2007-00531-000
- Muthukrishna, M., Bell, A. V., Henrich, J., Curtin, C. M., Gedranovich, A., McInerney, J., & Thue, B. (2020). Beyond Western, Educated, Industrial, Rich, and Democratic (WEIRD) Psychology: Measuring and Mapping Scales of Cultural and Psychological Distance. Psychological Science, 31(6), 678–701. https://doi.org/10.1177/0956797620916782
- Ramstead, M. J. D., Veissière, S. P. L., & Kirmayer, L. J. (2016). Cultural Affordances: Scaffolding Local Worlds Through Shared Intentionality and Regimes of Attention. Frontiers in Psychology, 7, 1090. https://doi.org/10.3389/fpsyg.2016.01090
- Rutar, D., De Wolff, E., Van Rooij, I., & Kwisthout, J. (2022). Structure Learning in Predictive Processing Needs Revision. Computational Brain & Behavior, 5(2), 234–243. https://doi.org/10.1007/s42113-022-00131-8