Week 13: Fabry (2017) Betwixt and between: the enculturated predictive processing approach to cognition

This week we were joined by Regina Fabry to discuss her work on enculturated cognition. This paper outlines a multilevel framework for accounting how cultural practices (referred to here, following Menary (2007), as “cognitive practices”) impact the neurocognitive profiles of human agents. Fabry does this by offering a synergistic combination of Richard Menary’sCognitive Integration” project with predictive processing – a recent popular position in the philosophy of mind and cognitive science which some have claimed has the potential to be ‘unifying theory of the mind/brain’ (e.g. Hohwy 2013 and Clark 2016). What I most enjoyed about this paper was the clarity with which Fabry outlines these multiple explanatory levels. But before discussing this, it is worth explaining the large number of technical terms. These fit into two main categories related to the two positions that Fabry is combining:

1. Enculturation: very briefly, this can be understood as the ontogenetic process by which human cognitive capacities are shaped by inhabiting specialised cultural-cognitive niches.

Cognitive practices: these are sets of cultural practices that guide how agents interact with the epistemic resources in their environment – the cultural cognitive niche. These are comparable to what we have discussed elsewhere as “patterned practices“.

Cognitive niche construction: this is act by which multiple agents collaborate over multiple generations (what Tomasello 1999 calls “virtual collaboration) to create a specialised physical and epistemic environment in which human agents act. Sterelny (2003, 2012) also refers to this “as downstream epistemic engineering”. As with my discussions in other posts I am using the term cultural-cognitive niche (this is to avoid confusion with how the term is used by evolutionary psychologists as I explain here)

The Manipulation thesis: Menary (2007) places an emphasis on the active embodied physical interactions by which human agents manipulate epistemic tools towards cognitive goals. Menary (2007, 2016) identifies a range of different coordination dynamics between agents and environmental cognitive resources:

  1. Biological coupling: sensory-motor loops and behavioural repertoires (e.g. saccadic eye movements associated with reading)
  2. Epistemic actions: the manipulation of an aspect of the environment which is not only a physical action but also has epistemic outcomes towards a cognitive goal (e.g. moving scrabble tiles to form a word)
  3. Self-correcting actions: the manipulation of the environment to constrain cognitive activity to legitimate moves in an abstract task space (e.g. phonological loops and verbal shadows that help to focus attention)
  4. Epistemic tool use: use of tools that don’t just have physical consequences but contribute to cognitive ends (e.g. pen and paper, calculators, or other measurement devices – in contrast to a simple physical tool such as a hammer)
  5. Use of public symbol systems: this is the use of representational systems towards a cognitive goal (e.g. Arabic-Hindu numerals used to count)

Cognitive norms: These physical manipulations are not only causal interactions but must also be seen as guided by norms that dictate the right and wrong ways by which an epistemic tool can be used to successfully complete a task. For instance, in the case of reading English one reads the text left to right and top to bottom.

Learning-driven plasticity: like Sterelny (2003), Menary (2014) stresses the importance of how plastic humans. In general humans are phenotyically plastic (this is what enables humans to live in an incredibly diverse range of habitats) but we are also especially plastic in regards to our nervous systems. As Downey & Lende (2012, p. 23) put it: “Our brain and nervous system are our most cultural organs”. Importantly, Menary emphasises that this neural plasticity can be driven by cultural learning (also see Heyes 2012)

Cognitive transformation: And it is this neural plasticity that enables humans to acquire a range of culturally derived cognitive capacities. This relates to Anderson’s (2010) notion of neural reuse. This is what Menary (2007, 2014) refers to as the “transformation thesis”: humans are able to perform “novel” skills and achieve cognitive tasks that would otherwise be highly infeasible and perhaps even impossible without these enculturated skills.

2. Predictive processing: the core point of predictive processing is that action, perception, and cognition can all be understood as involving a central process – namely, that the brain (or some part of the brain) is making predictions about incoming sensory signals, and attempting to make these predictions as accurate as possible by minimising prediction errors. This involves a functional hierarchy of levels (although the details of how this is physically instantiated in the brain are still being worked out) of both top-down predictions and bottom-up prediction errors (the latter are the discrepancy between the model generated by the brain and the actual state of the world). This functional system then attempts to minimise the prediction error whilst simultaneously balancing this against a number of other constraints to do with accuracy, cost, generality, etc (this is what is referred to as a precision estimation). This framework is shown in the diagram below (taken from from Hohwy 2013, p. 68).


Importantly, for the purposes of understanding Menary’s position on the importance of physically embodied manipulations of environmental resources towards cognitive tasks, predictive processing takes place in two ways (which are entwined): [1] Perceptual inference: the system makes predictions about the world and then adjusts and refines the prediction model to fit in the incoming sensory data. And [2] Active inference: the organism makes a prediction about the world and physically alters the world in order to reduce prediction error. Interestingly, Fabry proposes that physical manipulations of environmental resources can be understood (on the sub-personal level) as active inference.


The importance of explanatory pluralism for understanding human cognition

Having given a very brief overview of the key terms, I would now like to focus on what I take to be the most interesting aspect of this paper. [Here I take some interpretative licence by using slightly different terminology to those which she uses within the paper to match my own work based on Ed Hutchins – which I have discussed in previous weeks]

Fabry brings these two positions together in a parsimonious way – and demonstrates the strengths of this combination in a concrete manner by discussing reading acquisition. This example is chosen because Fabry claims that reading acquisition presents us with empirical evidence in favour of predictive processing (this is related to the relative, and ‘”U-shaped”, levels of activation patterns in the brain between non-readers, novices, and experts).

What I found of most interest was the emphasis on multilevel analysis for exploring how culture and cognition interact and intertwine. Fabry, following Kellert and colleagues (2006), refers to this position as explanatory pluralism. Explanatory pluralism is the claim that when investigating a specific phenomenon there is often more than one level  of analysis (a stronger version of this claim is to argue that it is always the case). And that we can attain a deeper understanding of the target phenomenon by bringing together these differing levels of analysis and differing fields with differing methods of interrogation. Although Fabry’s goal here is to gain a deeper understanding of the target phenomena, it is worth briefly noting that Wimsatt (2007) uses a similar position to argue for the reality of a putative phenomena – what he calls robustness. Viz. one can be relatively certain (although not completely) in the existence of a putative phenomena as a genuine feature of nature if multiple complementary but differing lines of inquiry point to the same conclusion. I recommend making this connection because then it creates another link between this topic and wider discussions in philosophy of science (something that I think Fabry has already done to a large extent).

In the case of enculturated cognition, Fabry nicely identifies two distinct scales – spatial and temporal (which she refers to, using the standard conventions as horizontal and vertical) – across which one can pick out at least three distinct levels (see diagram below). These are:

Spatial (horizontal):

  • sub-personal (e.g. neurotransmitters, neurons and neural regions)
  • personal (e.g. the psychological states of the individual)
  • supra-personal (e.g. the context  of wider cultural-cognitive niche in which the activity takes place)

Temporal (vertical):

  • ontogenetic (the developmental trajectories of the agent)
  • synchronic (activity at one particular time-slice)
  • phylogenetic (evolutionary and historical time scales)

Fabry, following Menary (2013), argues that when investigating human cognitive activity as it takes place at that moment we should also think about the developmental trajectories involved with how the agent has acquired sets of cognitive practices that orchestrate the activity. Additionally, we should also think about the phylogenetic time-scales in which sets of cognitive practices are transmitted and refined intergenerationally in the cultural-cognitive niche. Menary argues that if we want to properly understand cognition as it takes place at one moment it is often the case that we need to understand the diachronic time scales in which it takes place (this is also an argument made by Hutchins 2008, Sutton 2008 and others). But I still think that Fabry is right to identify this as a core point of difference between Menary’s approach and some other 4E positions (e.g. the extended mind).

blank graph 2

We can demonstrate the importance of diachronic temporal factors and multiple spatial scales by considering a simple example: a checklist. I have taken this case from Don Norman’s excellent discussion of cognitive artefacts (1991). Cognitive artefacts can be simply defined as devices that ‘enhance’ human agents cognitive capacities (as we will see below, this is a bit of a misnomer because it is often the case that the use of a cognitive tool does not straightforwardly ‘enhance’ an agent’s capacities, but instead mediates the activity and qualitatively alters either the task space or the agent’s capacities – also see Humphreys 2004, Menary & Gillett 2017 for similar points). When an agent uses this cognitive artefact this interaction is governed by cognitive practices which are the norm guided physical manipulations of this checklist aimed at completing the task successively – in this case a memory task. Doing so requires the agent having attained a degree of expertise through having learnt how to use the cognitive artefact (thus entailing that we pay attention to the ontogenetic time-scale). Additionally, in many many cases the particular cognitive artefacts used by modern humans were not invented by the user ex nihilo. Instead, as we saw in the discussion of Boyd and colleagues (2011), human intelligence and cognitive capacities – including the simple use of checklists – are mostly a cultural achievement (also see Hutchins 2008). And Norman notes that the use of cognitive artefacts has a big impact on our cognitive capacities. Importantly, how we understand this depends on the spatial level of analysis that we engage in. If one focuses on the supra-personal level of analysis one can see a hybrid system (of agent and checklist mediated by cognitive practices) that performs the memory task with a high degree of success. But if one then drops down to a personal level of analysis it becomes apparent that the agent is no longer straightforwardly engaged in a memory task. Instead, the task domain has been altered through being mediated by the checklist. And it now better understood as a perceptual task that  involves the manipulation of a public symbol system (what Hutchins 2005 calls a conceptual blend, and in which the checklist is an external representational device that “anchors” the task). Norman argues that if one does not consider the multiple levels of analysis here this point is overlooked and one makes conceptual errors (such as erroneously claiming that the checklist ‘enhances’ the agent’s memory capacity). Fabry’s aim is to supplement these other levels of analysis by providing a sub-personal level of explanation using predictive processing. Whilst I am personally ambivalent about the use of predictive processing, I agree with her wholeheartedly that we need a sub-personal level of analysis for explicating how the brain is enculturated. This a point that is well-recognised by Menary (2014) and other theorists who emphasise the importance of enculturation (e.g. Downey & Lende 2012). The flip side of Fabry’s synthesis is that the multiple levels of analysis covered by cognitive integration supplement and greatly inform the synchronic-individual level analysis that is the focus of predictive processing. I am much more in agreement here that a discussion of enculturation is useful and informative for the standard methodological individualism that pervades mainstream cognitive science (as Fabry’s example of reading and Norman’s example of the checklist both show).

In summary, I think Fabry’s clarification of the various levels of analysis that are crucial for understanding human cognitive activity is exemplary and something I am very much in agreement with (see Menary & Gillett 2017; and for other similar view see Hollan et al 2000). And I think that Fabry’s way of organising these various explanatory levels not only strengthens this point but also nicely shows how it is connected to wider debates in the philosophy of science (e.g. Kellert et al 2006).







Anderson 2010 Neural Reuse: A fundamental organizational principle of the brain. Behavioral and Brain Sciences, 33, 245-264.

Clark 2016 Surfing uncertainty: Prediction, action, and the embodied mind. Oxford, NY: Oxford University Press.

Downey & Lende 2012 Neuroanthropology and the Encultured Brain. (pp. 23-65) in D. Lende & G. Downey (Eds.) The Encultured Brain: An Introduction to Neuroanthropology. Cambridge, MA: MIT Press.

Fabry 2017 Betwixt and between: the enculturated predictive processing approach to cognition. Synthese DOI 10.1007/s11229-017-1334-y.

Heyes 2012 Grist and mills: on the cultural origins of cultural learning. Philosophical Transactions of the Royal Society: B, 367, 2181-2191.

Hohwy 2013 The predictive mind. Oxford: Oxford University Press.

Hollan et al 2000 Distributed Cognition: Toward a New Foundation for Human-Computer Interaction Research. ACM Transactions on Computer-Human Interaction, 7 (2), 174–196.

Humphreys 2004 Extending Ourselves: Computational Science, Empiricism, and the Scientific Method. Oxford: Oxford University Press.

Hutchins 2005 Material anchors for conceptual blends. Journal of Pragmatics, 37, 1555-1577

Hutchins 2008 The role of cultural practices in the emergence of modern human intelligence. Philosophical Transactions of the Royal Society B: Biological Sciences, 363, 2011-2019.

Kellert et al (eds.) 2006 Minnesota studies in philosophy of science, vol. 19: Scientific
pluralism. Minneapolis: University of Minnesota Press.

Menary 2007 Cognitive integration: Mind and cognition unbounded. Basingstoke, NY: Palgrave Macmillan.

Menary 2013 Cognitive integration, enculturated cognition and the socially extended mind. Cognitive Systems Research 25–26, 26–34.

Menary 2014 Neural Plasticity, Neuronal Recycling and Niche Construction. Mind and Language 29 (3), 286-303.

Menary 2016 Pragmatism and the Pragmatic Turn in Cognitive Science. (pp. 217-237) in Engel et al (Eds.) The Pragmatic Turn: Toward Action-Oriented Views in Cognitive Science. Strüngmann Forum Reports (Vol. 18). Cambridge, MA: MIT Press.

Menary & Gillett 2017 Embodying Culture: Integrated Cognitive Systems and Cultural Evolution. (pp. 72-87) in Kiverstein (ed.) The Routledge Handbook of Philosophy of the Social Mind. New York: Routledge.

Norman 1991 Cognitive Artifacts. (pp. 17-38) in Carroll (Ed.) Designing Interaction. Cambridge: Cambridge University Press.

Sterelny 2003 Thought in a Hostile World: The Evolution of Human Cognition. Oxford: Blackwell Publishing.

Sterelny 2012 The Evolved Apprentice: How evolution made humans unique. Cambridge, MA: MIT Press.

Sutton 2008 Material Agency, Skills and History: Distributed Cognition and the Archaeology of Memory. (pp. 37-55) in Knappett & Malafouris (eds.) Material Agency: Towards a Non-Anthropocentric Approach. London: Springer.

Tomasello 1999 The Cultural Origins of Human Cognition. London: Harvard University Press.

Wimsatt 2007 Re-Engineering Philosophy for Limited Beings: piecewise approximations to reality. Cambridge, MA: Harvard University Press.


Additional links:

Here is a video of Menary discussing enculturation in relation to mathematics and reading/writing.

Fabry also draws on the work of Daniel Ansari who also discusses enculturation and the development of mathematical cognitive abilities – here is a video of him discussing plasticity, education, and the acquisition of basic mathematical practices (e.g. enculturation).



Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s