Fabry, r. e. (2015). enriching the notion of enculturation: cognitive integration, predictive processing, and the case of reading acquisition - a commentary on richard menary.

Enriching the Notion of Enculturation:
Cognitive Integration, Predictive
Processing, and the Case of Reading
A Commentary on Richard Menary
Many human cognitive capacities are rendered possible by enculturation in com- bination with specific neuronal and bodily dispositions. Acknowledgment of this is of vital importance for a better understanding of the conditions under which soph- isticated cognitive processing routines could have emerged on both phylogenetic and ontogenetic timescales. Subscribing to enculturation as a guiding principle for   udents.uni-mainz.de the development of genuinely human cognitive capacities means providing a de- scription of the socio-culturally developed surrounding conditions and the pro- found neuronal and bodily changes occurring as a result of an individual's ongo- ing interaction with its cognitive niche. In this commentary, I suggest that the pre- dictive processing framework can refine and enrich important assumptions made by the theory of cognitive integration and the associated approach to enculturated cognition. I will justify this suggestion by considering several aspects that support the complementarity of these two frameworks on conceptual grounds. The result Macquarie University will be a new integrative framework which I call enculturated predictive pro- Sydney, Australia cessing. Further, I will supplement Richard Menary's enculturated approach to mathematical cognition with an account of reading acquisition from this new per- spective. In sum, I argue in this paper that the cognitive integrationist approach to enculturated cognition needs to be combined with a predictive processing style description in order to provide a full account of the neuronal, bodily, and environ-   u ni-mainz.de mental components giving rise to cognitive practices. In addition, I submit that the enculturated predictive processing approach arrives at a conceptually coherent and empirically plausible description of reading acquisition.
Cognitive integration Cognitive transformation Enculturation Neural plasti- city Neuronal reuse Predictive processing Reading acquisition Scaf olded Monash University Melbourne, Australia In his target paper Mathematical Cognition: A Resting on his theory of cognitive integration Case of Enculturation, Richard Menary invest- (CI; e.g., he starts from the igates the conditions under which phylogenetic- idea that these processes are fully continuous ally recent, socio-culturally shaped target phe- with phylogenetically older ones (evolutionary nomena within cognitive science such as math- continuity). This type of continuity is justified ematics, reading, and writing have emerged.
by the assumption that the evolution of neur- Fabry, R. E. (2015). Enriching the Notion of Enculturation: Cognitive Integration, Predictive Processing, and the Case of Reading Acquisition - A Commentary on Richard Menary. In T. Metzinger & J. M. Windt (Eds). Open MIND:C). Frankfurt am Main: MIND Group. doi: onal reuse mechanisms allows for the redeploy- lations of tokens of a representational writing ment of cortical circuits for phylogenetically re- system, and they serve to complete a cognitive task. In order to describe the transformational . Ontogenetically, neuronal reuse is processes by which cognitive practices are ac- a precondition of learning driven plasticity quired, Menary introduces the notion of encul- (LDP), which "can result in both structural and turation: "Enculturation rests on the acquisition functional changes in the brain" of cultural practices that are cognitive in p. 8). That is, the human brain is as- nature" . That is, enculturation refers to sumed to be neuronally plastic so that its pro- any cognitive transformation that is rendered cessing routines are altered as the individual ac- possible by LDP and the individual's ongoing quires new cognitive abilities .
interaction with its cognitive niche. As a proof However, the acquisition of new cognitive abilit- of concept, deals with ies takes place within mathematical cognition and describes the ways in which individuals acquire expertise in manip- […] a highly structured cognitive niche ulating a public, socio-culturally developed that contains not only physical artefacts, mathematical symbol system. Relying on a set but also: representational systems that em- of empirical results, he arrives at the conclusion body knowledge (writing systems, number that precise mathematical operations are systems, etc.); skills and methods for rendered possible by the recruitment of a neur- training and teaching new skills onal sub-system during ontogeny. In contrast to & ; practices for manipu- the evolved approximate number system (ANS), lating tools and representations. which allows for subitizing and is also present in other animals, the neuronal realization of the discrete number system (DNS) heavily depends It is this cognitive niche that provides the re- on LDP, the individual's immersion into its cog- sources for scaffolded learning, which allows the nitive niche, and its active participation in scaf- individual to acquire new cognitive abilities folded learning routines. Thus, the acquisition through its ongoing embodied interaction with of mathematical skills is an important example its socio-cultural environment. Together, LDP of enculturation.
and scaffolded learning lead to cognitive trans- The purpose of this commentary is to en- formations that augment the individual's cog- rich and refine the enculturated approach. First, nitive capacities through ontogenesis: "Cognit- I will propose that the predictive processing ive transformations result from our evolved framework provides conceptual and explanatory plasticity and scaffolded learning in the develop- tools for describing and explaining the neuronal mental niche" p. 8) and extracranial bodily mechanisms underlying The result of cognitive transformation is the ac- cognitive practices and enculturation. Thus, I quisition of a sufficient degree of expertise in will accept the challenge to combine "[…] the performing a certain cognitive practice. Cognit- dynamical nature of causal commerce between ive practices are normatively constrained to the world, body, and brain and the inferential free extent that socio-culturally shaped procedures energy principle that allows their unification in work in close interaction with the cognitive one account" p. 18). I niche: They "[…] are culturally endowed (bodily) will argue that a new integrative framework manipulations of informational structures" that views CI and predictive processing as com- p. 4), such as manipu- plementary is able to meet this challenge.
Second, I will illustrate this by presenting read- 1 More precisely, according to p. 293) it is scaffolded ing acquisition as a paradigmatic case of encul- learning that renders LDP possible in the course of cognitive devel- opment of individuals: "Both structural and functional plasticity can turated cognition. In particular, I will demon- result from both endogenous and exogenous sources, but here the fo- strate that a position that combines the encul- cus is on structural and functional changes driven by scaffolded turated approach with predictive processing, Fabry, R. E. (2015). Enriching the Notion of Enculturation: Cognitive Integration, Predictive Processing, and the Case of Reading Acquisition - A Commentary on Richard Menary. In T. Metzinger & J. M. Windt (Eds). Open MIND:C). Frankfurt am Main: MIND Group. doi: which I call enculturated predictive processing, different aspects that are crucial for an integra- leads to a parsimonious and conceptually coher- tionist approach to cognitive processing: 1. Hu- ent account of reading acquisition that helps in- man cognition is continuous with animal cogni- terpret and unify a vast array of recent empir- tion on both diachronic and synchronic scales.
However, it has a special status in that it is situated in a particular cognitive niche and 2 Towards a more complete approach to
heavily rests upon neural plasticity which is it- enculturation: Cognitive integration and self an adaptation (continuity thesis). 2. Certain
cognitive processes are hybrid because they are constituted by neuronal and extracranial bodily In order to appreciate the descriptive power of components (hybrid mind thesis). 3. In the the enculturated approach, it is necessary to spe- course of ontogenetic hybrid cognitive pro- cify the mechanistic underpinnings of the acquisi- cessing, both the constitutive neuronal and ex- tion of cognitive practices. In his summary of the tracranial bodily functions are transformed CI framework, p. 2) ar- (transformation thesis). 4. The bodily manipu- gues that "[a]lthough the framework is unified by lation of specific environmental resources plays a dynamical systems description of the evolution a crucial functional role in integrated cognitive of processing in the hybrid and multi-layered sys- processes (manipulation thesis). 5. These ma- tem, it recognises the novel contributions of the nipulations are constrained by cognitive norms, distinct processing profiles of the brain, body and which are acquired through learning, and which environment." However, the dynamical systems realize socio-culturally developed habits for the style approach to the acquisition and enactment interaction with cognitive resources (cognitive of cognitive practices in the version first intro- norms thesis). duced in pp. 42-48) does not ex- In addition to the continuity thesis and haustively specify the distinct, yet highly interact- the cognitive transformation thesis, which are ive neuronal and bodily components of cognitive given centre stage in Menary's target paper, the processing. Furthermore, it does not account for hybrid mind thesis is important in that it ac- LDP, simply because it remains neutral to the knowledges the close interaction of neuronal and concrete realization of its neuronal component extra-neuronal bodily sub-processes in the com- system. Final y, the dynamical systems approach, pletion of cognitive tasks. In other words, cer- on Menary's construal, helps il ustrate what the tain cognitive processes "involve the integration interactive contribution of neuronal and extracra- of neural manipulations of vehicles and bodily nial bodily components to human cognition might manipulations of environmental vehicles" amount to. Yet, it does not spel out the mutual p. 236; see also p.
influence that neuronal and extracranial bodily 627). The notion of bodily manipulation as it is components have over each other. used here goes back to This is where predictive processing (PP) pp. 23f) account of environmentalism, which enters the picture. In the remainder of this com- claims that "cognitive processes are, in part, mentary I will argue that the PP approach made up of manipulation of relevant structures provides the resources for a more detailed ac- in the cognizer's environment". In this context, count of how human cognitive systems become manipulation is defined as "any form of bodily enculturated and how they are subject to integ- interaction with the environment – manual or rated cognition. not, intrusive or otherwise – which makes use of the environment in order to accomplish a given 2.1 Cognitive integration: Five theses
task" p. 23). Thus, subscribing to the about human cognition
manipulation thesis amounts to the assumption that "[c]ognitive processing often involves these In its original version (cf., CI is online bodily manipulations of the cognitive constituted by five theses. They emphasize the niche, sometimes as individuals and sometimes Fabry, R. E. (2015). Enriching the Notion of Enculturation: Cognitive Integration, Predictive Processing, and the Case of Reading Acquisition - A Commentary on Richard Menary. In T. Metzinger & J. M. Windt (Eds). Open MIND:C). Frankfurt am Main: MIND Group. doi: in collaboration with others" are rendered possible only by the individual's p. 3). Importantly, it is assumed that ongoing interaction with its socio-culturally extracranial bodily manipulations causally in- shaped environment in normatively constrained teract with neural sub-processes, thereby stress- ways. This means to do justice to the broader ing the hybridity of cognitive processes (cf.
socio-cultural context of enculturated cognition, p. 138). In addition to highlight- while being interested in a precise description of ing the constitutive role of embodied engage- its neuronal and extracranial bodily sub-com- ments with "external" cognitive resources as ponents. In this commentary I will operate on proposed by, cognitive integra- the individual level of description without deny- tionists claim that the manipulation of these re- ing that it is important to develop a fine- sources is constrained by cognitive norms. In grained description on the social level by spe- this vein, p. 5; p. 233) ar- cifying the properties of a certain cognitive gues that "[o]ur abilities to manipulate the ex- niche and the conditions under which it could trabodily environment are normative and are have emerged.
largely dependent on our learning and training To this end, I will now proceed by sum- histories." The idea that certain cognitive abilit- marizing the most important features of the ies are normatively structured thus concerns the predictive processing (PP) approach that will individual's interaction with specific resources help specify the mechanistic underpinnings of provided by the cognitive niche. Importantly, enculturated cognition. the normatively constrained ways in which en- vironmental resources are integrated into cog- 2.2 An outline of predictive processing
nitive processes are shared by many individuals.
Put differently, the normativity of cognitive Recently, the idea that human perception, ac- practices helps "[…] stabilise and govern inter- tion, and cognition can be described and ex- active thought across a population of similar plained in terms of hierarchically organized pre- phenotypes" p. 4). Fur- dictive processing mechanisms implemented in thermore, the acquisition of a certain cognitive the human brain has enjoyed widespread atten- practice is tightly connected with the acquisi- tion within cognitive neuroscience (e.g., tion of the relevant cognitive norms in the course of scaffolded learning. This is because mind, and philosophy of cognitive science (e.g., "we learn cognitive practices by learning the cognitive norms that govern the manipulation of vehicles" p. 628). . The overall epistemic goal of this emer- From these five theses defended by CI it ging approach is to describe perceptual, sensor- follows that there should be two distinct, yet in- imotor, and cognitive target phenomena within terdependent levels of description for cognitive a single framework by relying on unifying mech- practices. First, there is the social level of de- anistic principles. Accounts of PP generally as- scription. On this level, cognitive practices need sume that human perception, action, and cogni- to be approached by highlighting the interact- tion are realized by Bayesian probabilistic gen- ive, cooperative cognitive achievements of a erative models implemented in the human large group of individuals sharing the same cog- brain. Since the human brain does not have im- nitive niche. Second, cognitive practices can be mediate access to the environmental causes of investigated by approaching them on an indi- sensory effects, it has to infer the most probable vidual level of description. In this case, the ac- state of affairs in the environment giving rise to quisition and enactment of a certain cognitive sensory data (cf. pp. 4f).
practice is described with regards to a certain PP approaches solve this inverse problem by as- individual. However, any individual level de- suming that generative models in accordance scription needs to acknowledge that certain cog- with Bayes' rule are implemented in the human nitive capacities of an enculturated individual brain. On this construal, a generative model Fabry, R. E. (2015). Enriching the Notion of Enculturation: Cognitive Integration, Predictive Processing, and the Case of Reading Acquisition - A Commentary on Richard Menary. In T. Metzinger & J. M. Windt (Eds). Open MIND:C). Frankfurt am Main: MIND Group. doi: "[…] aims to capture the statistical structure of that is an effect of environmental (or bodily) some set of observed inputs by tracking […] the causes (cf. p. 187; p.
causal matrix responsible for that very struc- 269, p. 88). This is known as prediction ture" p. 182). In order to be able to infer the causes of sensory effects, generative Prediction error minimization is a special models encode probability distributions. Each way of minimizing free energy in accordance generative model provides several hypotheses with the principle "that any self-organizing sys- about the causes of a certain sensory input. The tem that is at equilibrium with its environment system has somehow to ‘decide' which hypo- must minimize its free energy" p.
thesis needs to be chosen in order to account for 127). Applied to human perception, cognition, the cause of the sensory effect. The descriptive and action, minimizing free energy means min- power of Bayes' rule lies in its capacity to cap- imizing the amount of unbound energy available ture the probabilistic estimations underlying to the perceiving, cognizing, and acting organ- these choices. Applied to the case of human per- ism. This is where prediction error enters the ception, action, and cognition, Bayesian gener- picture. As p. 186) puts it, ative models are assumed to be realized in hier- "[p]rediction error reports this information-the- archically organized structures comprising mul- oretic free energy, which is mathematically con- tiple, highly interactive low- and high-level cor- structed so as always to be greater than ‘sur- tical areas. This is referred to as the Bayesian prisal' (where this names the sub-personally brain hypothesis (cf. p. 129). The computed implausibility of some sensory state hierarchical organization of probabilistic gener- given a model of the world […])." The relation- ative models is combined with a specific version ship between free energy and surprisal then is of predictive coding, where predictive coding that "[…] free energy is an upper bound on sur- "depicts the top-down flow as attempting to prise, which means that if agents minimize free predict and fully ‘explain away' the driving energy, they implicitly minimize surprise" sensory signal, leaving only any residual ‘predic- p. 128). Suprisal, however, cannot be tion errors' to propagate forward within the sys- estimated directly by the system, because tem" p. 182). That is to say, selec- "there is an infinite number of ways in which ted hypotheses inform prior predictions about the organism could seek to minimize surprise the sensory input to be expected at each level of and it would be impossibly expensive to try the hierarchy. These predictions fulfil the func- them out" p. 3). The solution to tion of encoding knowledge about statistical this problem lies in implicitly minimizing sur- regularities of patterns in the observable (or any prisal (and its upper bound, i.e., free energy) by imaginable) world. This hypothesis selection minimizing prediction error (cf. p.
proceeds in accordance with Bayes' rule. The processing of sensory input gives rise to predic- p. 6). It is exactly here where prediction tion errors. Prediction errors carry neuronally 2 On a neuronal level of description, hierarchical generative models realized information about "[…] residual differ- are assumed to be neuronally realized by multiple connections ences, at every level and stage of processing, across low- and high-level cortical areas. Each level within the cortical hierarchy is connected to the next subordinate and between the actual current signal and the pre- supraordinate level, thereby ensuring effective inter-level message dicted one" p. 4). Import- passing (cf. pp. 67f). According to p.
antly, it is only prediction errors, and not sens- 187), predictive generative models are implemented in "a kind of duplex architecture". This means that there are distinct neuronal ory input per se, that are fed forward within units dedicated to the representation of predictions of environ- the hierarchy (cf. pp. 182f; mental (or bodily) causes, so-called representation units, on the one hand, and those dedicated to the encoding of prediction er- p. 3, p. 47, p. 4). The overall ror, so-called error units, on the other (cf. p.
aim of this multi-level processing mechanism is 829). To date, a detailed account of the concrete neuronal realiz- ation of these functionally distinct units of message-passing is to minimize prediction error, that is, to reduce still missing (cf. However, it is hypothesized that repres- or to ‘explain away' the discrepancy between entation units might correspond to superficial pyramidal cells, predictions and the actually given sensory input while error units might correspond to deep pyramidal cells (cf.
p. 8; see also pp. 187f).
Fabry, R. E. (2015). Enriching the Notion of Enculturation: Cognitive Integration, Predictive Processing, and the Case of Reading Acquisition - A Commentary on Richard Menary. In T. Metzinger & J. M. Windt (Eds). Open MIND:C). Frankfurt am Main: MIND Group. doi: error minimization avails itself as a tractable ex- intuitively correspond to increasingly ab- pression of more general life-sustaining mechan- stract conceptions of the world, and these tend to capture or depend upon regularit- Prediction error minimization can be ies at larger temporal and spatial scales.
achieved in two distinct, yet complementary Lower-level (more ‘perceptual') ones cap- ways. The first of these is perceptual inference, ture or depend upon the kinds of scale and which can be described as detail most strongly associated with spe- cific kinds of perceptual contact. […] an iterative step-wise procedure where a hypothesis is chosen, and predictions are made, and then the hypothesis is revised Consequently, processes typically associated in light of the prediction error, before new with perception or cognition can only be distin- and hopefully better predictions are made guished by considering the temporal and spatial on the basis of the revised hypothesis.
resolution of the instantiation of PP mechan- isms and the levels at which model revision en- sues, respectively. This relationship between That is, prediction errors are propagated up the perception and cognition becomes important hierarchy leading to an adjustment of the initial once we consider how enculturated cognition hypothesis, thereby achieving an approximation has been rendered possible on both phylogenetic of the hypothesis generating the predictions and and ontogenetic time scales. For it helps specify the actually given input. The adjustment of pre- how evolutionary continuity could have been dictions and hypotheses in the face of fed-for- rendered possible in the first place. The evolu- ward prediction error occurs at every level of tionary development of perception and cogni- the hierarchy until any prediction error is ac- tion (and, as we shall see, of action too) may commodated. This complex process comprising have proceeded from more perceptual generative multiple levels is known as perception: "Percep- models present in many other animals to more tion thus involves ‘explaining away' the driving cognitive generative models exclusively realized (incoming) sensory signal by matching it with a in humans. This is in line with s cascade of predictions pitched at a variety of p. 45) observation that "[t]he underlying spatial and temporal scales" p.
neural models are basically species-unspecific, 187; see als p. 762). and the empirical cases move back and forth On Andy Clark's account of PP, one im- between many different model systems." Refer- portant consequence of this is that the tradi- ring to this observation, tional distinction between perception and cogni- p. 14) emphasizes that "[t]he basic elements of tion becomes blurred. It is replaced by a recon- the predictive processing story, as Roepstorff ceptualization of perceptual and cognitive pro- (2013, p. 45) correctly notes, may be found in cesses as a continuous employment of the same many types of organism and model-system." prediction error minimizing mechanism on mul- Thus, while certain (lower-level) model para- meters and processing stages of prediction error minimization are shared by many organisms, All this makes the lines between percep- there certainly are specific (higher-level) pro- tion and cognition fuzzy, perhaps even cessing routines that are shared only by encul- vanishing. In place of any real distinction turated human organisms in a certain cognitive between perception and belief we now get variable differences in the mixture of top- Furthermore, the idea that perception and down and bottom-up influence, and differ- cognition are continuous is relevant for consider- ences of temporal and spatial scale in the ations of the ontogenetic development of encul- internal models that are making predic- turated cognitive functions. This is because it tions. Top-level (more ‘cognitive') models anchors higher-order cognitive operations in Fabry, R. E. (2015). Enriching the Notion of Enculturation: Cognitive Integration, Predictive Processing, and the Case of Reading Acquisition - A Commentary on Richard Menary. In T. Metzinger & J. M. Windt (Eds). Open MIND:C). Frankfurt am Main: MIND Group. doi: more basic perceptual processes and thus allows to reduce sensory prediction error" for a fine-grained description of a certain devel- opmental trajectory leading to cognitive trans- This suggests that perceptual and active formation. Bearing in mind the hierarchical inference, or perception and bodily action for structure of generative models, another interest- that matter, mutually influence each other, ing consequence of the PP style approach to thereby minimizing prediction errors and optim- perception and cognition is that lower (i.e., izing hypotheses generating ever new predic- more perceptual) levels of the generative model tions. However, perceptual and active inference influence higher (i.e., more cognitive) levels by have a "different direction of fit" means of fed-forward prediction error. Vice p. 178; see also p. 13; versa, higher levels of the hierarchical generative p. 7)This is because in model influence lower levels by means of fed- perceptual inference, predictions are aligned to backward predictions (cf. p. 73).
the sensory input, while active inference is a This will become more important when we ex- matter of aligning the sensory input to the pre- plore how reading acquisition can be described dictions. It follows "[…] that to optimally engage as an ongoing enculturating process of predic- in prediction error minimization, we need to en- tion error minimization.
gage in perceptual inference and active inference Perceptual inference is only one way of in a complementary manner" p.
minimizing prediction error. The second is act- 91). Since both perceptual and active inference ive inference, where "[…] the agent will select- are aimed at minimizing prediction error and ively sample the sensory input it expects" optimizing generative models, "[p]erception and p. 129). The idea is that the system action […] emerge as two sides of a single com- can minimize prediction error by bringing about putational coin" p. 760). the states of affairs (i.e., the environmental hid- As emphasized earlier, perception and cog- den causes) that are predicted by a certain hy- nition are deeply related to the extent that both pothesis. This is achieved by performing any phenomena are the result of the same underly- type of bodily movements, including eye move- ing functional and neuronal mechanisms. By ex- ments, that make the selected prediction come tension, action is also deeply intertwined with true. The predictions at play in active inference cognition. This follows from the assumptions are counterfactual, because that 1. perception and cognition are continuous and 2. perception and action are subject to the […] they say how sensory input would same principles of prediction error minimiza- change if the system were to act in a cer- tion. As p. 5) puts it, both tain way. Given that things are not actu- ways of prediction error minimization "[…] un- ally that way, prediction error is induced, fold continuously and simultaneously, underlin- which can be minimized by acting in the ing a deep continuity between perception and prescribed way. p. 82; italics action […]." Yet, perceptual and active inference in original; see also fulfil distinct functional roles in their ongoing attempt to minimize prediction error. This be- comes even more obvious once we take the free Accordingly, in active inference the selected pre- energy principle into account: "The free energy diction is held constant and leads to bodily principle […] does not posit any fundamental activities that minimize prediction error by al- difference between perception and action. Both tering the sensory input such that it confirms fall out of different reorganizations of the prin- the prediction. Therefore, active inference is of ciple and come about mainly as different direc- crucial importance for prediction error minimiz- 3 The notion of two functions having "a different direction of fit" originates ation, "[…] since it provides the only way (once in s p. 234) speech act theory and in a good world model is in place and aptly activ- s p. 56) example il ustrating how words and states of ated) to actually alter the sensory signal so as affairs can relate to each other. I would like to thank Thomas Metzinger for pointing out the philosophical history of this notion.
Fabry, R. E. (2015). Enriching the Notion of Enculturation: Cognitive Integration, Predictive Processing, and the Case of Reading Acquisition - A Commentary on Richard Menary. In T. Metzinger & J. M. Windt (Eds). Open MIND:C). Frankfurt am Main: MIND Group. doi: tions of fit for prediction error minimization 2.3 Combining cognitive integration and
[…]" p. 13). Active infer- ence plays a crucial role in cognition (under- stood as prediction error minimization compris- To what extent is it feasible to describe the ing many higher-level predictions), for it helps mechanisms underlying cognitively integrated minimize prediction error throughout the cor- processes and enculturated cognition in terms of tical hierarchy by bringing about the states of prediction error minimization? After having affairs in the environment that are predicted on summarized CI and the core ideas of the PP higher levels. Therefore, on s p.
framework I will argue in this section that there 187) account, which he dubs action-oriented are many aspects of the CI approach that can predictive processing, prediction error minimiza- be enriched by making a crucial assumption, tion "[…] depicts perception, cognition and ac- namely that PP can account for many compon- tion as profoundly unified and, in important re- ents constituting cognitive practices on at least spects, continuous." functional and neuronal levels of description.
PP accounts of human perception, action, First, a major conceptual consequence of and cognition distinguish between first-order PP is that perception, action, and cognition are and second-order statistics. In contrast to first- both continuous and unified, if this approach order statistics, which amount to minimizing proves correct. This is because they follow the prediction error by means of perceptual and same principles of prediction error minimiza- active inference, second-order statistics are tion, yet are characterized by important func- concerned with estimating the precision of pre- tional differences. This kind of complementarity diction error. In second-order statistics, the in- fits neatly with the hybrid mind thesis defended fluence of fed-forward prediction error on by CI. Recall that the hybrid mind thesis claims higher levels of the hierarchical generative that cognitive processes are constituted by both model is dependent upon its estimated preci- neuronal and extracranial bodily components.
sion. Neuronally, the estimation of precision is By taking prediction error minimization into ac- captured in terms of increasing or decreasing count, this claim can be cashed out by assuming the synaptic gain of specific error units (cf.
that the neuronal components are equal to per- & p. 2). That is, "[t]he ceptual inferences at multiple levels of the cor- more precision that is expected the more the tical hierarchy, while the bodily components are gain on the prediction error in question, and mechanistically realized by active inferences.
the more it gets to influence hypothesis revi- The hybrid mind thesis emphasizes the indis- sion" p. 66; see also pensable, close and flexible coordination of p. 132). Conversely, if the precision is ex- neuronal and bodily components responsible for pected to be poor on the basis of second-order the completion of a cognitive task. The PP statistics, the synaptic gain on the error unit is framework, or so I shall argue, provides the re- inhibited such that the prediction on the sources for a careful description of the underly- supraordinate level is strengthened (cf. p.
ing mechanisms at play. It does so by depicting 123). It has been proposed that precision es- human organisms as being constantly engaged timation is equivalent to attention. This means in prediction error minimization by optimizing that "attention is nothing but optimization of hypotheses in the course of perceptual inference precision expectations in hierarchical predictive and by changing the stimulus array in the coding" p. 70; see also course of active inference.
& p. 2). For current purposes, it A second advantage of the prediction error is sufficient to focus in the main on first-order minimization framework is that it helps cash statistics. However, it is important to bear in out the manipulation thesis. This thesis, recall, mind the crucial modulatory role precision es- states that "the manipulation of external timation plays in prediction error minimiza- vehicles [is] a prerequisite for higher cognition and embodied engagement [is] a precondition Fabry, R. E. (2015). Enriching the Notion of Enculturation: Cognitive Integration, Predictive Processing, and the Case of Reading Acquisition - A Commentary on Richard Menary. In T. Metzinger & J. M. Windt (Eds). Open MIND:C). Frankfurt am Main: MIND Group. doi: for these manipulative abilities" calls "path-based idiosyncrasies". That is, one of p. 232). In terms of the PP framework, bodily the reasons why the coordination of neuronal manipulation can be understood as an instance and bodily components in the manipulation of of active inference occurring in specific contexts.
cultural resources is beneficial certainly is that That is, in order to complete a certain cognitive it takes place in a normatively constrained task, the system changes its sensory input by al- "multi-generational development of stacked, tering certain components of its cognitive niche.
complex ‘designer environments' for thinking This becomes even more obvious once we take such as mathematics, reading, writing, struc- into account that embodied activity is also a tured discussion, and schooling" . That is means of increasing confidence in sensory input to say that the performance of cognitive prac- by optimizing its precision. As suggested by tices in compliance with certain norms has the p. 6), "expected precision overall advantage of reducing cognitive effort, drives action such that sensory sampling is which can be captured as the minimization of guided by hypotheses that the system expects overall prediction error and the optimization of will generate precise prediction error." Applied precision on a sub-personal level of description.
to an organism's interaction with its socio-cul- At the same time, however, cognitive practices turally shaped environment, p.
themselves can be described, or so I shall argue, 238) argues "[…] that many of the ways we in- as having prediction error minimization as their teract with the world in technical and cultural underlying mechanism. This double role of cog- aspects can be characterized by attempts to nitive practices, described in terms of prediction make the link between the sensory input and error minimization, can be fully appreciated the causes more precise (or less uncertain)." once we consider the cognitive transformations However, bodily manipulation is more than just brought about by the ongoing interaction with a contributing factor to prediction error minim- cultural resources.
ization (and precision optimization). In order to Fourth, our cognitive capacities and the acknowledge this, we need to take into account various ways we complete cognitive tasks are that bodily manipulations are a crucial compon- profoundly augmented by our neuronal and ent of the performance of cognitive practices. In bodily engagements with the socio-culturally the performance of a cognitive practice, the structured environment through ontogenesis (cf.
minimization of prediction error and the optim- p. 341). Put differently, "cognitive ization of precision is not an end in itself.
transformations occur when the development of Rather, it serves to facilitate the completion of the cognitive capacities of an individual are a certain cognitive task. Furthermore, the con- sculpted by the cultural and social niche of that crete bodily manipulations given in terms of individual" p. 8). This active inference are subject to cognitive norms niche includes mathematical symbol systems, that constrain the ways in which human organ- representational writing systems, artifacts, and isms interact with cultural resources, such as so forth. It is this immersion and, importantly, tokens of a representational writing system.
the scaffolding provided by other inhabitants of That is to say that the performance of a cognit- the cognitive niche that ideally lead to the ive practice is not an individualistic enterprise.
transformation of neuronal and extracranial Rather, in completing a cognitive task, the indi- bodily components constituting cognitive pro- vidual is deeply immersed into a socio-cultural cesses, to enculturation that is. The PP frame- context which is shared by many human organ- work, or so I shall argue, offers a highly prom- ising account of learning that is most suitable Third, it is the normative constraints on for a sub-personal level description of cognitive cognitive practices that render their perform- transformation. On the construal of PP, learn- ance efficient and, in many cases at least, suc- ing flows naturally from the mechanism of pre- cessful. This is because compliance with these diction error minimization. For learning can norms induces what p. 195) generally be construed as a sub-personally real- Fabry, R. E. (2015). Enriching the Notion of Enculturation: Cognitive Integration, Predictive Processing, and the Case of Reading Acquisition - A Commentary on Richard Menary. In T. Metzinger & J. M. Windt (Eds). Open MIND:C). Frankfurt am Main: MIND Group. doi: ized strategy of optimizing models and hypo- highly structured, systematically ordered pat- theses in the face of ever new prediction error: terns of sensory input in the cognitive niche.
"Learning is then viewed as the continual up- This, however, needs to be complemented by a dating of internal model parameters on the fine-grained personal-level description of the basis of degree of predictive success: models are kind of interactions between experts and novices updated until they can predict enough of the that is needed in order to pass on the right set signal" p. 268). Broadly under- of cognitive norms. Furthermore, the kind of stood, ‘learning' thus figures as an umbrella cognitive transformation at play here requires a term referring to the ongoing activity of predic- description of the neuronal changes that are tion error minimization and model optimization correlated with the acquisition of a certain cog- throughout the lifetime of a human organism.
nitive practice. That is, we need a more fine- This is because potentially ever new and "sur- grained account of LDP and how it might be prisaling" sensory signals need to be "explained realized in the human cortex. From the per- away" by perceptual and active inference. For spective of the PP framework, one plausible current purposes, however, "learning" can also conjecture at this point is that LDP can be cap- be understood in a rather narrow sense as the tured in terms of effective connectivity. Effective acquisition of a certain skill, which is also sub- connectivity reports the causal interaction of ject to prediction error minimization through neuronal assemblies across multiple levels of the perception, action, cognition, and the modula- cortical hierarchy (and across different brain tion of attention. It is the individual's socio-cul- areas) as a result of attention in terms of preci- turally structured environment that delivers sion estimation. This line of reasoning is implied new sensory signals helping optimize parameters by p. 190) who argues that of the generative model: "[a]ttention […] is simply one means by which certain error-unit responses are given increased But those training signals are now de- weight, hence becoming more apt to drive learn- livered as part of a complex developmental ing and plasticity, and to engage in compensat- web that gradually comes to include all ory action." This last point is important, since the complex regularities embodied in the it stresses that it is not only perceptual infer- web of statistical relations among the sym- ence that drives learning and contributes to the bols and other forms of socio-cultural scaf- improvement of generative models, but also act- folding in which we are immersed. We thus ive inference. However, this approach to the ac- self-construct a kind of rolling ‘cognitive quisition of action patterns in concert with an niche' able to induce the acquisition of optimization of precision might raise the worry generative models whose reach and depth that learning is depicted here as being a rather far exceeds their apparent base in simple internalistic, brain-bound affair. But once we forms of sensory contact with the world.
acknowledge that it is the performance and on- going improvement of embodied active infer- ences that play an indispensable functional role However, complex skills that are targeted at the in the completion of cognitive tasks, it becomes completion of cognitive tasks cannot be learned obvious that this worry is not warranted. For it simply by being exposed to the right kind of is the efficient interaction of neuronal and ex- "training signal" in the cognitive niche. What is tracranial bodily components (i.e., perceptual additionally needed is engagement in activities and active inferences in terms of PP) that res- that are scaffolded by inhabitants of that cog- ults from learning and the efficient engagement nitive niche who have already achieved a suffi- of human organisms with their environment.
cient degree of expertise. This is what Furthermore, LDP can now be considered in calls "scaffolded learning". From terms of the precision-weighted optimization of the perspective of PP, this amounts to the hypotheses throughout the cortical hierarchy strategy of exposing predictive systems to and the ever new patterns of effective con- Fabry, R. E. (2015). Enriching the Notion of Enculturation: Cognitive Integration, Predictive Processing, and the Case of Reading Acquisition - A Commentary on Richard Menary. In T. Metzinger & J. M. Windt (Eds). Open MIND:C). Frankfurt am Main: MIND Group. doi: nectivity, as new cognitive practices are ac- First, I do not assume that CI necessarily requires quired and successfully performed. The sub-per- PP. Hypothetical y, it is conceivable that another sonal description of cognitive transformation in theory of neuronal and bodily functioning might terms of prediction error minimization also does be more suited to cashing out cognitive practices justice to neuronal reuse as a guiding principle and enculturation more convincingly and more of the allocation of neuronal resources for phylo- extensively. To date, PP appears to be the best genetically recent cognitive functions such as unifying framework that helps specify exhaust- arithmetic or reading.
ively the functional and neuronal contributions of From this, the following question arises: bodily and neuronal sub-processes giving rise to What is the actual relationship between CI and cognitive practices and enculturation. This is be- PP supposed to be and what is the scope of this cause PP offers a fine-grained functional and theory synthesis? First of all, the position de- neuronal description of perception, action, cogni- veloped in this commentary is neutral with re- tion, attention, and learning that does justice to gards to metaphysical consequences that may or the complex interactions stipulated by CI and the may not result from the idea that CI and PP associated approach to enculturation. can be integrated into a unified theoretical Second, it could be assumed that CI and framework. Rather, this position has an instru- PP are merely compatible. This would mean that mentalist flavour to the extent that it tries to CI and PP were self-sufficient and co-existent the- answer the question by which means socio-cul- oretical frameworks whose claims and key as- turally shaped target phenomena can be best sumptions do not necessarily contradict each investigated both conceptually and empirically.
other. This compatibility assumption is too weak Thus, the combination of CI and PP is valid for various reasons thar have been presented in only to the extent that it displays great de- this commentary so far. For it is the purpose of scriptive as well as predictive power and is sup- the theory synthesis sketched here to enrich and ported by many results stemming from empir- refine the notion of enculturation and the associ- ical research. As such, the new approach on of- ated theses defended by CI. Furthermore, to the fer here is contingent upon the current state of extent that PP directly speaks to complex cognit- research in cognitive science. It is falsifiable by ive phenomena and learning, it benefits from the new empirical evidence or convincing conceptual effort of CI to do justice to the socio-cultural y considerations that directly speak against it.
shaped context in which these phenomena can be Furthermore, it sidesteps the concern that PP developed. This is to say that CI and PP can be and the underlying free energy principle might directly referred to each other in ways that I have be trivial because they can be applied to any started to il ustrate in this section. target phenomenon by telling a "just-so story".
Finally, from this it follows that both This is because the combination of CI and PP frameworks are more than just compatible – is applied to specific domains, namely to classes they are complementary. Taken together, they of cognitive processes that count as cognitive provide us with complex and far-reaching con- practices, with reading being the paradigm ex- ceptual tools for investigating complex cognitive ampleThus the approach advocated can be phenomena that are shaped by the individual's seen as a modest contribution to the project immersion in its cognitive niche. Thus, the com- aiming at a "[…] translation into more precise, plementarity of CI and PP leads to a new integ- constricted applications to various domains, rative framework that I dub enculturated pre- where predictions can be quantified and just-so dictive processing (EPP).
stories avoided" p. 14).
The idea that CI and PP can be combined 2.4 Defending enculturated predictive
can lead to different degrees of commitment 4 Thanks to Jennifer M. Windt for raising this point.
At first glance, the EPP framework might ap- 5 Thanks to an anonymous reviewer for helpful suggestions on this is- pear to be unwarranted. For prediction error Fabry, R. E. (2015). Enriching the Notion of Enculturation: Cognitive Integration, Predictive Processing, and the Case of Reading Acquisition - A Commentary on Richard Menary. In T. Metzinger & J. M. Windt (Eds). Open MIND:C). Frankfurt am Main: MIND Group. doi: minimization could be construed as being a environment only insofar as it tries to make the purely internalistic, brain-bound affair that does effects of hidden causes fit the predictions. This not leave any room for the idea that cognitive precludes the theoretical possibility of depicting processes are constituted both by neuronal and prediction error minimizing systems as being extracranial bodily components that are norm- situated, scaffolded, integrated, or extended. atively constrained, socially scaffolded, and However, this line of reasoning fails to ac- deeply anchored in a socio-culturally structured knowledge the conceptual necessity of emphasiz- ing the functional role of embodied active infer- First, consider a position that takes for ence in terms of its contribution to the minimiza- granted that cognitive processes can be coher- tion of prediction error and the optimization of ently described in terms of prediction error min- predictions. For even if the causal relations hold- imization, but which denies that cognitive pro- ing between a predictive, generatively organized cesses are co-constituted by neuronal and bodily system and environmental causes are mediated by sub-processes operating on socio-cultural re- hypotheses, predictions, prediction errors and pre- sources. Such a position is defended by cision estimation as encoded in the cortical hier- p. 240) who argues that "[…] archy, it does not fol ow that this system is just a many cases of situated and extended cognition passive receiver of sensory input that informs it begin to make sense as merely cases of the brain about remote states in the environment. Similarly, attempting to optimize its sensory input so it, it does not necessarily fol ow from the prediction as positioned over against the world, can better error minimization framework that it "[…] creates minimize error." In particular, according to his a sensory blanket – the evidentiary boundary – interpretation of the prediction error minimiza- that is permeable only in the sense that inferences tion framework, "[…] the mind remains secluded can be made about the causes of sensory input from the hidden causes of the world, even hidden beyond the boundary", as though we are ingenious in using culture and p. 7) claims. Rather, the predictive system is part technology to allow us to bring these causes into of its socio-cultural y structured environment and sharper focus and thus facilitate how we infer to has many possibilities for bodily acting in that environment in order to facilitate its own cognit- For Hohwy, this directly follows from the ive processing routines. Considering embodied causal relations holding between the predictive active inference, it turns out that the causal rela- system and the environmental causes it con- tion holding between embodied action (in terms stantly tries to infer. According to him p.
of bodily manipulation) and changes of the set of 228), this relation needs to be characterized as available stimuli in the environment is as direct as "direct" and "indirect" at the same time: any causal relation could be. This is because these changes are an immediate effect of these very pre- […] the intuition that perception is indirect diction error-minimizing and precision-optimizing is captured by its reliance on priors and actions, which in turn contribute to the perform- generative models to infer the hidden ance of cognitive tasks. Furthermore, we need to states of the world, and the intuition that take into account that genuinely human cognitive perception is direct is captured by the way processes occur in a cultural y sculpted cognitive perceptual inference queries and is sub- niche, which is characterized by mathematical sequently guided by the sensory input symbol systems, representational writing systems, causally impinging on it.
artifacts, and the like, and other human organ- isms with whom we interact. These cognitive re- Since the causal relation that holds between a sources have unique properties that render them predictive system comprised of inverted generat- particularly useful for the completion of cognitive ive models and the world is partly indirect, so tasksor example, consider the regularity of line the argument goes, the system is in constant embodied interaction and direct contact with its 6 Thanks to Richard Menary for raising this important point in per- Fabry, R. E. (2015). Enriching the Notion of Enculturation: Cognitive Integration, Predictive Processing, and the Case of Reading Acquisition - A Commentary on Richard Menary. In T. Metzinger & J. M. Windt (Eds). Open MIND:C). Frankfurt am Main: MIND Group. doi: arrangements and the orderliness of succeeding practices, such as the hybridity, the transform- letters in an alphabetic writing system. Once ative efficacy, and the enculturated nature of learned and automatized, fol owing these normat- cognitive processes, can be supplemented and ive principles facilitates several types of cognitive refined by taking the prediction error minimiza- processing routines. That is to say that it is the tion framework into account.
socio-cultural y shaped sensory input itself that The arguments in favour of the EPP has an important impact on the concrete realiza- framework directly speak to the current debate tion of prediction error minimization. This cannot within philosophy of mind and philosophy of be accounted for if we assume that the predictive cognitive science about the relationship between processing of cognitive resources is an internal- the prediction error minimization framework istic, secluded endeavour.
and approaches to situated, distributed, integ- Second, consider a line of reasoning that rated, or extended cognition. On the one hand, goes against the compatibility of CI with the prediction error minimization framework, that methodological and metaphysical grounds that might be put forward by an integrationist. She there is anything like these types of cognition might agree that we need a mechanistic descrip- from the perspective of prediction error minim- tion of the neuronal and bodily components ization. According to him, this is because pre- which jointly constitute cognitive processes in dictive systems have only indirect access to the the close interaction with socio-cultural re- world. Furthermore, there is "the sensory sources. But she might continue to argue that boundary between the brain and the world" the performance of cognitive practices is more which prohibits predictive systems from enga- than just the minimization of prediction error ging in any variant of situated, distributed, in- and the optimization of precisionFrom the tegrated, or extended cognition including CI perspective of PP, it needs neither to be denied p. 240). On the other hand, that human cognitive systems as a whole aim to p. 195) argues that the PP frame- fulfil cognitive purposes by completing cognitive work at least "[…] offers a standing invitation to tasks and that they do so by engaging in cognit- evolutionary, situated, embodied, and distrib- ive practices. Nor should it be rejected that cog- uted approaches to help ‘fill in the explanatory nitive practices are normatively constrained and gaps' while delivering a schematic but funda- that cognitive systems are deeply immersed in a mental account of the complex and comple- socio-culturally structured environment, which mentary roles of perception, action, attention, in turn provides these very norms through scaf- and environmental structuring." Once we take folding teaching. However, the important theor- the arguments and considerations in favour of etical contribution made by the prediction error EPP into account we have reasons to think that minimization framework is its providing of a EPP lends support to Clark's construal of the sub-personal, mechanistic description of the un- PP framework. This will become even more per- derlying neuronal and bodily sub-processes that suasive once we take empirical data and a turns out to be parsimonious, conceptually co- paradigm case of EPP into account.
herent, and empirically plausible. In addition, PP also offers a description of the close interac- 3 Reading acquisition: A case of
tion of the neuronal and bodily components constituting cognitive practices by offering a concise description of the ongoing, mutually So far, I have argued that the notion of encul- constraining interplay of perceptual and active turation and key claims made by CI can be en- inferences. More generally, this section should riched by taking the PP framework into ac- have established that all important claims and count. In particular, the hybridity, embodied- assumptions made by CI in favour of cognitive ness, and transformative character of encultur- ated cognition can be mechanistically described 7 This consideration was put forward by Richard Menary in personal in terms of prediction error minimization. How- Fabry, R. E. (2015). Enriching the Notion of Enculturation: Cognitive Integration, Predictive Processing, and the Case of Reading Acquisition - A Commentary on Richard Menary. In T. Metzinger & J. M. Windt (Eds). Open MIND:C). Frankfurt am Main: MIND Group. doi: ever, cognitive practices cannot be fully reduced normative insofar as they constrain the ways in to prediction error minimization, since they which combinations of letters are pronounced have a normative dimension that needs to be in- and written words are correctly related to vestigated on a personal level of description.
spoken words. The acquisition of this normative This section serves to illustrate the valid- knowledge needs "explicit instruction in the al- ity of the line of reasoning put forward in this phabetic principle" p. 57) commentary. This will be done by showing that It follows that learning these norms is socially reading acquisition, understood as another case structured and dependent upon the cooperation of enculturation next to mathematical cogni- of experts with novices. This fits neatly with tion, can be fruitfully described from the per- s p. 361) following assumption: spective of EPP.
Manipulative norms and interpretative 3.1 Scaffolded learning and the
norms apply to inscriptions of a public acquisition of cognitive norms
representational system and are never simply dependent on an individual. In- One crucial aspect of learning to perform a cog- deed, it is the individual who must come nitive practice is the acquisition of the relevant to be transformed by being part of the cognitive norms, where this class of norms "gov- community of representational system ern[s] manipulations of external representations, which aim at completing cognitive tasks" p. 238). In the case of reading, these Acquiring knowledge about grapheme-phoneme norms concern the recognition and identification correspondences, especially in an inconsistent of tokens of a representational writing system.
orthography such as English, puts demands not In alphabetic writing systems, important cog- only on the novice, but also on the teachers who nitive norms are derived from the so-called al- assist her in learning these correspondences. For phabetic principle, where this principle amounts the teachers, being experts in reading, need to to the "mapping [of] written units onto a small break down their automatic identification and set of elements – the phonemes of a language" recognition skills in order to be able to teach p. 33; see also the norms underlying the relationship between p. 87). Specifically, the correspondence of graphemes and phonemes. As p.
graphemes to phonemes puts culturally estab- 145) points out more generally, "[e]xpert per- lished, normative constraints on the ways in formance is often rapid and fluent, without ob- which individual letters (and combinations vious components. Learning from such perform- thereof) are related to phonological units. The ance is difficult. It becomes much easier if the normative scope of these correspondences is task is overtly decomposed into segments, each best illustrated by differences across languages of which can be represented and practiced indi- and orthographies. As pointed out by& vidually." In the present context, the most suc- p. 430), "[i]n some orthograph- cessful strategy of teaching grapheme-phoneme ies, one letter or letter cluster can have multiple correspondence has turned out to be so-called pronunciations (e.g. English, Danish), whereas phonics instruction (cf. pp.
in others it is always pronounced in the same 31f): "[…] teaching methods that make the al- way (e.g. Greek, Italian, Spanish).This phabetic principle explicit result in greater suc- demonstrates that the degree of consistency or cess among children trying to master the read- transparency of grapheme-phoneme correspond- ing skills than methods that do not make it ex- ences is subject to arbitrary stipulations by a plicit" p. 34). This goes along with teach- linguistic, literate community employing a spe- ing novices that spoken language consists of cific orthographic system. These stipulations are phonemes. That is, children's reading acquisi- 8 This phenomenon is also known as orthographic depth. For a recent 9 See also p. 219), p. 26), and review, see.
Fabry, R. E. (2015). Enriching the Notion of Enculturation: Cognitive Integration, Predictive Processing, and the Case of Reading Acquisition - A Commentary on Richard Menary. In T. Metzinger & J. M. Windt (Eds). Open MIND:C). Frankfurt am Main: MIND Group. doi: tion is dependent upon, or at least co-develops coming proficient in applying the alphabetic with phonological awareness, where this is un- principle, getting to grips with phoneme-graph- derstood as "[…] the ability to perceive and ma- eme correspondences, and developing phonolo- nipulate the sounds of spoken words" gical and metalinguistic awareness are cases of & p. 78). The metalinguistic awareness that spoken language consists of phonemes must be explicitly acquired and al- 3.2 Reading acquisition and neuronal
lows the novice to learn that these units corres- pond to letters, or combinations thereof. It is still debated whether phonological awareness is Next to scaffolded learning, another crucial as- a prerequisite for learning to read or whether it pect of cognitive transformation is LDP (cf.
is co-emergent with basic letter decoding skills.
p. 356, p. 8). In- However, as suggested by & deed, in the case of reading acquisition, there is p. 104), "[…] it may not be possible for unequivocal evidence pointing to "[…] plastic phonemic awareness to be acquired at all in the changes in brain function that result from the absence of instruction on the links between acquisition of skills" p. 93). By phonemes and graphemes." Thus, it seems safe the same token, p.
to assume that phonological awareness clearly 2397) emphasize that "[…] culturally guided facilitates the ability to relate graphemes to education couples with experience-dependent phonemes. There are other components of meta- plasticity to shape both cortical processing and linguistic awareness that influence the successful reading development." As& application of norms governing alphabetic rep- p. 477) point out, the application of resentational writing systems. Beginning readers knowledge about grapheme-phoneme corres- are already proficient speakers of their native pondences in novice readers "[…] implicates the language and are able to fluently apply syn- formation of functional connections between tactic, semantic, and pragmatic norms in their visual object processing systems and systems in- everyday conversations. However, they are usu- volved in processing spoken language." The left ally unable to explicitly represent that utter- ventral occipitotemporal (vOT) area appears to ances are made up of sentences and that sen- play a crucial role in establishing these connec- tences are made up of combinations of words (cf. p. 308; p.
35). To novices, these basic properties must be there has been consensus on the contribution of made explicitly available in order to put those the vOT area to a neuronal reading circuit. In a novices in the position to apply knowledge series of experiments, Stanislas Dehaene, about them automatically and fluently at later Laurent Cohen and their colleagues have made stages of reading acquisition. Furthermore, the remarkable discovery that neuronal activa- novices need to be acquainted with the conven- tion in one particular region of the left vOT tion, which is fairly obvious to expert readers, area is reliably and significantly associated with that alphabetic writing systems are decoded visual word recognition in adult, non-patholo- from left to right and from the top to the bot- tom of a page. These basic personal-level com- ponents of the acquisition of reading skills provide the cognitive norms necessary for the . This region, especially the left ventral oc- development of reading understood as a cognit- cipito-temporal sulcus next to the fusiform ive practice. It is these norms that govern the gyrus, frequently responds to visually presented successful manipulation of representational words regardless of the size, case, and font in vehicles belonging to an alphabetic writing sys- which they are made available (cf. tem that need to be established by social inter- p. 143; p. 293).
action between learners and teachers. Thus, be- This consistent finding has led these researchers Fabry, R. E. (2015). Enriching the Notion of Enculturation: Cognitive Integration, Predictive Processing, and the Case of Reading Acquisition - A Commentary on Richard Menary. In T. Metzinger & J. M. Windt (Eds). Open MIND:C). Frankfurt am Main: MIND Group. doi: to call it the visual word form area (VWFA), acquired capacity to recognize visually presen- since it crucially contributes to "[…] a critical ted words (cf.& p. 468; process that groups the letters of a word to- see also p. 286). This "recyc- gether into an integrated perceptual unit (i.e. a ling" is in turn constrained by the overall ‘visual word form')" p.
evolved neuronal architecture and already ex- 293). However, it is debatable whether the left isting processing mechanisms (cf. vOT area is almost exclusively dedicated to pp. 146f). Thus, neuronal recycling is visual word recognition in expert readers, or just a special type of neuronal reuse (see whether this area serves several functions hav- for a discussion). There are cer- ing to do with the (visual) identification of tain conditions that need to be met if a spe- shapes more broadly construed (see & cific cortical area is to be ‘recycled' for a for a discussion). Neverthe- phylogenetically recent cognitive function (see less, the findings by Dehaene and his colleagues p. 288). In the case of visual that the left vOT area plays a crucial role in word recognition, the left vOT area is as- the overall visual word recognition process is sumed to exert certain "functional biases" important and widely acknowledged, although that make it most suitable for the recognition the interpretations of its functional contribution and identification of visually presented words: "(1) a preference for high-resolution foveal An important motivation for research on shapes; (2) sensitivity to line configurations; the overall function of the left vOT area stems and (3) a tight proximity, and, presumably, from considerations on the phylogenetic devel- strong reciprocal interconnection to spoken opment of visual word recognition. Consider- language representations in the lateral tem- ing that writing systems were invented only poral lobe" & 256).
approximately 5400 years ago, it is unlikely These "functional biases", however, do not that the ability to read is the result of an preclude that the left vOT area is still en- evolutionary process (cf. p.
gaged in other cognitive processes such as ob- 134, p. 5; p.
ject recognition in skilled adult readers (cf.
293). In a nutshell, the crucial question is how visual word recognition is possible given "[…] p. 257; & p. 478).
that the human brain cannot have evolved a Rather, it helps explain why this area is found dedicated mechanism for reading" & to be well-equipped for contributing to the p. 254). This is also referred to overall process of visual word recognition.
as the "reading paradox" p.
However, the question arises what the contri- 4). The solution to this paradox proposed by bution of the left vOT area to the overall Dehaene and his colleagues is to assume "[…] visual word recognition process is supposed to that plastic neuronal changes occur in the make. According to s & context of strong constraints imposed by the s Interactive Account (IA), the prior evolution of the cortex" as a result of contribution of the left vOT area can be best the human organism being exposed to tokens described and explained in terms of PP. In of a certain writing system & line with the general principles of the PP p. 254). Specifically, the idea is "[…] that framework presented above, they generally writing evolved as a recycling of the ventral hold the following assumption: "Within the visual cortex's competence for extracting con- hierarchy, the function of a region depends on figurations of object contours" (. This its synthesis of bottom-up sensory inputs con- view, which has been dubbed the neuronal re- veyed by forward connections and top-down cycling hypothesis (cf. p. 150), predictions mediated by backward connec- suggests that existing neuronal functions asso- tions" & p. 247). In other ciated with visual cognition are "recycled" for words, the suggested synthesis equals the pre- the phylogenetically recent, ontogenetically diction error that results from the discrepancy Fabry, R. E. (2015). Enriching the Notion of Enculturation: Cognitive Integration, Predictive Processing, and the Case of Reading Acquisition - A Commentary on Richard Menary. In T. Metzinger & J. M. Windt (Eds). Open MIND:C). Frankfurt am Main: MIND Group. doi: between top-down predictions and bottom-up as the proficiency in visual word recognition in- sensory information. Applied to the patterns of neuronal activation associated with visual word recognition, this assumption is specified In pre-literates, vOT activation is low be- cause orthographic inputs do not trigger appropriate representations in phonolo- For reading, the sensory inputs are written gical or semantic areas and therefore there words (or Braille in the tactile modality) are no top-down influences […]. In early and the predictions are based on prior as- stages of learning to read, vOT activation sociation of visual or tactile inputs with is high because-top-down predictions are phonology and semantics. In cognitive engaged imprecisely and it takes longer for terms, vOT is therefore an interface the system to suppress prediction errors between bottom-up sensory inputs and and identify the word […]. In skilled read- top-down predictions that call on non- ers, vOT activation declines because learn- visual stimulus attributes. & ing improves the predictions, which ex- plain prediction error efficiently […].
Accordingly, the vOT area is supposed to be as- That is, IA assumes that the level of activation sociated with a distinct level of the hierarchical within the left vOT area is dependent upon the generative model responsible for visual word re- general establishment and refinement of a gen- cognition mediating between higher-level, lan- erative model comprising both lower-level areas guage-related predictions and bottom-up visual associated with visual processing and higher- information. It follows that "[…] the neural im- level cortical areas associated with phonological plementation of classical cognitive functions and semantic knowledge. If this account turns (e.g. orthography, semantics, phonology) is in out to be correct, the blurredness of the distinc- distributed patterns of activity across hierarch- tion between perception and cognition as sug- ical levels that are not fully dissociable from one gested by becomes vitally import- another" p. 249). Specifically, IA pro- ant. For it is the mutual interplay of lower-level poses a neuronal mechanism that is able to processing stages (traditionally associated with demonstrate how linguistic knowledge about visual processing) and higher-level processing phonology and semantics, encoded in top-down stages (traditionally associated with phonolo- predictions, causally interacts with bottom-up gical and semantic processing) that renders the information. This is because it is held that a successful acquisition of visual word recognition prediction error is generated each time bottom- possible in the first place. Evidence in favour of up information diverges from the associated IA comes from studies demonstrating that there top-down prediction. In turn, the resulting pre- is a significant increase of activation in this area diction error is associated with significant activ- as a result of exposure to visually presented ation in the left vOT area. Empirical evidence words in beginning readers across different re- supporting this approach to the functional con- search paradigms and methodologies employing tribution of the left vOT area to visual word re- cognition in expert readers is widely available . Furthermore, two longitudinal ERP studies demonstrate that the left- lateralized occipito-temporal N1 effect, an effect In reading acquisition, the left vOT area associated with print sensitivity, does not de- appears to be an equally important contributor velop in a linear fashion in the course of reading to visual word recognition. According to acquisition. Rather, s p.
& p. 248), the activation level of 756) comparison of their results obtained from the vOT area develops in a non-linear fashion, their child participants with an adult control Fabry, R. E. (2015). Enriching the Notion of Enculturation: Cognitive Integration, Predictive Processing, and the Case of Reading Acquisition - A Commentary on Richard Menary. In T. Metzinger & J. M. Windt (Eds). Open MIND:C). Frankfurt am Main: MIND Group. doi: group indicates that "[i]nstead of a linear in- special case of overall prediction error minimiza- crease with more proficient reading, the devel- tion according to the PP framework. On this opment is strongly nonlinear: the N1 specializa- construal, learning to read means becoming in- tion peaks after learning to read in beginning creasingly efficient in predicting linguistic, visu- readers and then decreases with further reading ally presented input as a result of long-term ex- practice in adults following an inverted U- posure to types of this input and the optimiza- shaped developmental time-course." In this vein, tion of hypotheses through perceptual inference.
p. 7942) interpret their res- The careful instruction in relating graphemes to ults by suggesting that "[t]he emergence of print phonemes, phonological and metalinguistic sensitivity in cortical areas during the acquisi- awareness, and the normatively constrained al- tion of grapheme-phoneme correspondences is in phabetic principle provides the environmental line with the inverse U-shaped developmental conditions for efficient and progressively more trajectory of print sensitivity of the ERP N1, accurate prediction error minimization. The sig- which peaks in beginning readers […]." nals delivered by this highly structured learning Another consequence ofs &s environment are estimated as being precise, PP account of reading acquisition is that such that the synaptic gain on error units re- the activation level within the vOT should be porting the discrepancy between (still inaccur- associated with the degree of accuracy of top- ate) predictions and prediction error is high. As down predictions in the face of bottom-up sig- learning to read proceeds, the predictions be- nals. This is supported by various studies come more accurate and the overall influence of demonstrating that higher-level activations of prediction error shows a relative decrease. This cortical areas associated with language pro- line of reasoning is supported by s & cessing are also present in beginning readers.
s p. 248) following suggestion: "At For example, p. 772) the neural level, learning involves experience-de- report that "[a]ctivity in the left ventral inferior pendent synaptic plasticity, which changes con- frontal gyrus increased with reading ability and nection strengths and the efficiency of percep- was related to both phonological awareness and tual inference." Understood this way, LDP and phonological naming ability. […] Brain activity the associated neuronal transformations can be in the anterior middle temporal gyrus also in- understood as being realized by prediction error creased with reading ability", where this area is minimization in the context of scaffolded learn- associated with semantic processing. Similarly, ing, which allows a beginning reader to become report activation in the ever more efficient and successful in this partic- middle temporal gyrus, which is frequently asso- ular cognitive practice. ciated with semantic processing in expert read- ers (e.g., p. 2; & 3.3 Reading acquisition and bodily
p. 4). Furthermore, they re- port significant activation patterns in left IFG, Starting from the hybrid mind thesis defended which is associated with both phonological and by CI, which states that certain cognitive pro- semantic processing. cesses are constituted by both neuronal and ex- In the light of much empirical evidence in tracranial bodily sub-processes, it seems natural favour ofs &s approach to to assume that reading acquisition also is asso- the neuronal changes corresponding to reading ciated with the transformation of bodily sub- acquisition, it seems safe to assume that it is processes. That is, in the course of encultura- empirically plausible and can account for many tion it is the enactment of bodily manipulation data derived from experiments in cognitive that is transformed in addition to the neuronal neuroscience. However, to what extent can this changes occurring as a result of LDP. In terms approach be conceptually enriched? Recall that of PP, this assumption leads to the suggestion learning a new skill such as reading is just a that it is not only perceptual inferences that are Fabry, R. E. (2015). Enriching the Notion of Enculturation: Cognitive Integration, Predictive Processing, and the Case of Reading Acquisition - A Commentary on Richard Menary. In T. Metzinger & J. M. Windt (Eds). Open MIND:C). Frankfurt am Main: MIND Group. doi: causally relevant for learning described in terms proficient readers show a "[…] refixation of prediction error minimization, but also active strategy, with initial saccade landing positions inferences that allow for ever more efficient sub- located closer to word beginnings." Similarly to personally employed strategies for "explaining away" incoming sensory input. Recall that eye port a longitudinal study comparing the per- movements are just a special case of active in- formance of 6- to 11-year-old children in a read- ference (see e.g., . Their ing task and a visual task. In line with the em- functional contribution to prediction error min- pirical evidence already mentioned, their results imization becomes vitally important for a com- indicate that "[w]ith age, children's reading cap- plete account of visual word recognition and its abilities improve and they learn to read by acquisition. This is because visual word recogni- making larger progressive saccades, fewer re- tion, in both novices and experts, is rendered gressive saccades and shorter fixations […]" possible by the coordination of perceptual and p. 6). Furthermore, it is active inference. From the perspective of CI, the demonstrated that the eye movement patterns idea here is that the ways in which an indi- employed in reading and in visual search di- vidual bodily manipulates a certain cognitive verge with increasing reading proficiency (cf.
resource is importantly improved in the course of cognitive transformation. Applied to reading An explanation of these results in terms of acquisition, this leads to the assumption that PP is straightforward. In beginning readers, the specific eye movement patterns become more ef- predictions initiating active inference occurring ficient as a result of reading instruction and it- in a highly-structured linguistic environment are erate exposure to a certain type of cognitive re- inaccurate, such that the generation and execu- source (say, sentences printed on a piece of pa- tion of eye movements in terms of active infer- ence is not as efficient as it is in the case of ex- Recently, it has become possible to invest- pert readers. By the same token, the inaccuracy igate eye movements in beginning readers by of the currently selected prediction makes it ne- employing eye-tracking methodologies. Conver- cessary to sample the visually available lin- ging evidence suggests that beginning readers guistic environment more thoroughly, explaining make more fixations (i.e., acquisition of visual the "refixation strategy" and the execution of information in the absence of oculomotor activ- comparatively more saccades. As reading skills ities), saccades (i.e., oculomotor activities), and improve, resulting from increasingly efficient regressions (i.e., backward saccades), and ex- prediction error minimization through percep- hibit longer fixation durations and smaller sac- tual inference as already suggested, the accur- cade amplitudes than proficient and expert acy of predictions becomes increasingly optimal, readers (cf. p. 3; therefore allowing for more efficient active infer- p. 46). More specifically, these tenden- ence. More efficient active inference, in turn, al- cies are assessed in a longitudinal eye-tracking lows for more efficient perceptual inference, study reported by. They since both types of inference mutually influence measured eye movements during an oral reading each other. This line of reasoning is supported task in second and fourth graders of a German by s p. 2957) claim in- primary school and additionally assessed overall formed by the results of their study "[…] that reading skills and oculomotor behaviour beyond only linguistic, not oculomotor skills were the reading (cf. p. 2949).
driving force behind the acquisition of normal Their results indicate that the fourth graders, in oral reading skills." Thus, the increase in effi- comparison to the second graders, show a de- ciency of eye movements in beginning readers crease of fixation duration, gaze duration, total does not result from an increase in oculomotor reading time, refixations, and saccadic amp- capabilities per se, but works in tandem with litudes (cf. p. 2956). higher-level linguistic knowledge encoded in pre- p. 2958) attest that the younger, less dictions, which are associated with representa- Fabry, R. E. (2015). Enriching the Notion of Enculturation: Cognitive Integration, Predictive Processing, and the Case of Reading Acquisition - A Commentary on Richard Menary. In T. Metzinger & J. M. Windt (Eds). Open MIND:C). Frankfurt am Main: MIND Group. doi: tions in higher-order cortical areas. As a result, for a conceptually coherent description of the the improvement of active inference in the interaction between brain, body, and environ- course of reading acquisition works in tandem mental cognitive resources. However, it needs to with the improvement of perceptual inference.
be supplemented by a sub-personal level de- This highlights that learning to read does not scription in terms of prediction error minimiza- only result in neuronal, but also in bodily trans- tion in order to be able to specify the neuronal formations. As such, the optimization of eye and functional underpinnings of the hybrid movements in the course of reading acquisition mind thesis, the bodily manipulation thesis, and highlights the importance of bodily manipula- the transformation thesis as defended by CI. At tion in the efficient enactment of reading under- the same time, the approach to reading acquisi- stood as a cognitive practice. This also means tion put forward in this commentary suggests to suggest that a complete account of encultura- that a vast array of empirical findings from cog- tion should not only pay attention to scaffolded nitive neuroscience and cognitive psychology learning and LDP, but also to the develop- can be unified for the first time by interpreting mental trajectory of bodily manipulation.
them from the new perspective of enculturated predictive processing. Thus, I submit that we 4 Concluding remarks
can only appreciate the cognitive assets rendered possible by our socio-culturally struc- This commentary on Richard Menary's paper tured environment once we account for the en- Mathematical Cognition: A Case of Encultura- abling conditions of sophisticated, neuronally tion started from the assumption that the gen- and bodily realized cognitive processes such as eral outline of enculturation and the associated mathematical cognition and reading. These con- claims made by CI provide important concep- ditions include socio-culturally established ways tual tools for the description of ontogenetically of learning and teaching, LDP, and the ability acquired, socio-culturally shaped cognitive pro- to adapt action patterns to the needs and re- cessing routines. However, I have argued that quirements of a certain cognitive task. My over- the idea of enculturation and its most import- all claim is that we need the EPP framework to ant aspects, namely cognitive transformation be able to approach the entire spectrum of and scaffolded learning, need to be enriched by these factors, whose complex interplay ulti- providing a detailed functional and neuronal de- mately leads to truly enculturated cognition.
scription on a sub-personal level of description.
In addition, it needs to be born in mind that enculturation is rendered possible by normative constraints developed by a large group of indi- The author wishes to thank the Barbara viduals sharing the same cognitive niche. To Wengeler Foundation for its generous financial this end, I have suggested that the notion of en- support. In addition, she is indebted to Thomas culturation and its associated constitutive as- Metzinger, Jennifer M. Windt, and an anonym- pects can be complemented in important ways ous reviewer for their helpful feedback on earlier by taking the PP framework into account. The versions of this commentary.
result is what I call enculturated predictive pro- cessing. Thus, the PP framework is capable of providing the conceptual resources necessary for a thorough description of the mechanistic un- derpinnings of cognitive practices and their ac- quisition. Lending further support to this line of reasoning, I have dealt with reading acquisition as a paradigmatic case of enculturated predict- ive processing. This should have been sufficient to establish that the CI framework is well-suited Fabry, R. E. (2015). Enriching the Notion of Enculturation: Cognitive Integration, Predictive Processing, and the Case of Reading Acquisition - A Commentary on Richard Menary. In T. Metzinger & J. M. Windt (Eds). Open MIND:C). Frankfurt am Main: MIND Group. doi: Cohen, L. & Dehaene, S. (2004). Specialization within the ventral stream: The case for the visual word form area.
Anderson, M. L. (2010). Neural reuse: A fundamental or- NeuroImage, 22 (1), 466-476.
ganizational principle of the brain. Behavioral and Brain Sciences, 33 (04), 245-266.
Dehaene, S. (2005). Evolution of human cortical circuits for reading and arithmetic: The "neuronal recycling" Anderson, M. L. & Finlay, B. L. (2014). Allocating struc- hypothesis. In S. Dehaene, J.-R. Duhamel, M. D.
ture to function: The strong links between neuroplasti- Hauser & G. Rizzolatti (Eds.) From monkey brain to city and natural selection. Frontiers in Human Neuros- human brain. A Fyssen foundation symposium (pp.
133-157). Cambridge, MA: MIT Press. Ansari, D. (2012). Culture and education: New frontiers (2010). Reading in the brain: The new science of in brain plasticity. Trends in Cognitive Sciences, 16 (2), how we read. New York, NY: Penguin Books. (2011). The massive impact of literacy on the Anscombe, G. E. M. (1963). Intention (2nd ed.). Cam- brain and its consequences for education. Human bridge, Mass: Harvard University Press. neuroplasticity and education. Pontifical Academy of Austin, J. L. (1953). How to talk: Some simple ways.
Sciences, 117, 19-32. Proceedings of the Aristotelian Society (53), 227-246. Dehaene, S. & Cohen, L. (2011). The unique role of the Bedo, N., Ribary, U., Ward, L. M. & Valdes-Sosa, P. A.
visual word form area in reading. Trends in Cognitive (2014). Fast dynamics of cortical functional and effect- Sciences, 15 (6), 254-262.
ive connectivity during word reading. PLoS ONE, 9 Dehaene, S., Cohen, L., Sigman, M. & Vinckier, F.
(2), e88940-e88940.
(2005). The neural code for written words: A proposal.
Ben-Shachar, M., Dougherty, R. F., Deutsch, G. K. & Wan- Trends in Cognitive Sciences, 9 (7), 335-341.
del , B. A. (2011). The development of cortical sensitivity to visual word forms. Journal of Cognitive Neuroscience, Feldman, H. & Friston, K. J. (2010). Attention, uncer- 23 (9), 2387-2399.
tainty, and free-energy. Frontiers in Human Neuros- Brem, S., Bach, S., Kucian, K., Guttorm, T. K., Martin, E.
& Lyytinen, H. (2010). Brain sensitivity to print emerges Friston, K. (2005). A theory of cortical responses. Philo- when children learn letter-speech sound correspondences.
sophical Transactions of the Royal Society B: Biolo- Proceedings of the National Academy of Sciences, 107 gical Sciences, 360 (1456), 815-836.
(17), 7939-7944.
Carreiras, M., Armstrong, B. C., Perea, M. & Frost, R.
(2010). The free-energy principle: A unified brain (2014). The what, when, where, and how of visual theory? Nature Reviews Neuroscience, 11 (2), 127-138.
word recognition. Trends in Cognitive Sciences, 18 (2), Friston, K., Adams, R. A., Perrinet, L. & Breakspear, M.
Castles, A. & Coltheart, M. (2004). Is there a causal link (2012). Perceptions as hypotheses: Saccades as Experi- from phonological awareness to success in learning to ments. Frontiers in Psychology, 3.
read? Cognition, 91 (1), 77-111. Frith, U. (1985). Beneath the surface of developmental Clark, A. (2012). Dreaming the whole cat: Generative dyslexia. In K. E. Patterson, J. C. Marshall & M. Colt- models, predictive processing, and the enactivist con- heart (Eds.) Surface dyslexia. Neuropsychological and ception of perceptual experience. Mind, 121 (483), 753- cognitive studies of phonological reading (pp. 301-330).
Hillsdale, NJ: Erlbaum. (2013). Whatever next? Predictive brains, situated Gaillard, W., Balsamo, L., Ibrahim, Z., Sachs, B. & Xu, agents, and the future of cognitive science. Behavioral B. (2003). fMRI identifies regional specialization of and Brain Sciences, 36 (03), 181-204.
neural networks for reading in young children. Neuro- logy, 60 (1), 94-100.
(2015). Embodied prediction. In T. Metzinger & Hohwy, J. (2011). Phenomenal variability and introspect- J. M. Windt (Eds.) Open MIND. Frankfurt a. M., ive reliability. Mind & Language, 26 (3), 261-286.
GER: MIND Group. Fabry, R. E. (2015). Enriching the Notion of Enculturation: Cognitive Integration, Predictive Processing, and the Case of Reading Acquisition - A Commentary on Richard Menary. In T. Metzinger & J. M. Windt (Eds). Open MIND:C). Frankfurt am Main: MIND Group. doi: (2012). Attention and conscious perception in the (2014). Neural plasticity, neuronal recycling and hypothesis testing brain. Frontiers in Psychology, 3.
niche construction. Mind & Language, 29 (3), 286-303.
(2013). The predictive mind. Oxford, UK: Oxford (2015). Mathematical cognition: A case of encul- University Press. turation. In T. Metzinger & J. M. Windt (Eds.) Open (2014). The self-evidencing brain. Noûs.
MIND. Frankfurt a. M., GER: MIND Group. Menary, R. & Kirchhoff, M. (2014). Cognitive transform- (2015). The neural organ explains the mind. In T.
ations and extended expertise. Educational Philosophy Metzinger & J. M. Windt (Eds.) Open MIND. Frank- and Theory, 1-14.
furt a. M., GER: MIND Group. Olulade, O. A., Flowers, D. L., Napoliello, E. M. & Eden, Huestegge, L., Radach, R., Corbic, D. & Huestegge, S. M.
G. F. (2013). Developmental differences for word pro- (2009). Oculomotor and linguistic determinants of cessing in the ventral stream. Brain and Language, 125 reading development: A longitudinal study. Vision Re- (2), 134-145.
search, 49 (24), 2948-2959.
Price, C. J. & Devlin, J. T. (2003). The myth of the Joseph, H. S. S. L., Liversedge, S. P. & Paterson, K. (2013).
visual word form area. NeuroImage, 19 (3), 473-481.
Children's and adults' on-line processing of syntactical y ambiguous sentences during reading. PLoS ONE, 8 (1), (2004). The pro and cons of labelling a left oc- cipitotemporal region "the visual word form area".
Kherif, F., Josse, G. & Price, C. J. (2011). Automatic NeuroImage, 22 (1), 477-479. top-down processing explains common left occipito- (2011). The interactive account of ventral occipito- temporal responses to visual words and objects. Cereb- temporal contributions to reading. Trends in Cognitive ral Cortex, 21 (1), 103-114.
Sciences, 15 (6), 246-253.
Kronbichler, M., Hutzler, F., Wimmer, H., Mair, A., Staffen, Price, C. J. & Mechelli, A. (2005). Reading and reading W. & Ladurner, G. (2004). The visual word form area disturbance. Current Opinion in Neurobiology, 15 (2), and the frequency with which words are encountered: Evidence from a parametric fMRI study. NeuroImage, 21 Rayner, K., Foorman, B. R., Perfetti, C. A., Pesetsky, D.
(3), 946-953.
& Seidenberg, M. S. (2001). How psychological science Maurer, U., Brem, S., Kranz, F., Bucher, K., Benz, R., informs the teaching of reading. Psychological Science Halder, P., Steinhausen, H.-C. & Brandeis, D. (2006).
in the Public Interest, 2 (2), 31-74. Coarse neural tuning for print peaks when children learn to read. NeuroImage, 33 (2), 749-758.
Richlan, F. (2014). Functional neuroanatomy of developmental dyslexia: The role of orthographic depth. Frontiers in Hu- McCandliss, B. D., Cohen, L. & Dehaene, S. (2003). The man Neuroscience, 8.
visual word form area: Expertise for reading in the Roepstorff, A. (2013). Interactively human: Sharing time, fusiform gyrus. Trends in Cognitive Sciences, 7 (7), constructing materiality. Behavioral and Brain Sci- ences, 36 (03), 224-225.
Menary, R. (2006). Attacking the bounds of cognition.
Rowlands, M. (1999). The body in mind: Understanding Philosophical Psychology, 19 (3), 329-344.
cognitive processes, Cambridge studies in philosophy. Cambridge, UK: Cambridge University Press. (2007a). Cognitive integration: Mind and cognition Schlaggar, B. L. & McCandliss, B. D. (2007). Develop- unbounded. New York, NY: Palgrave Macmillan. ment of neural systems for reading. Annual Review of (2007b). Writing as thinking. Language Sciences, Neuroscience, 30 (1), 475-503.
29 (5), 621-632.
(2010). Cognitive integration and the extended Schurz, M., Kronbichler, M., Crone, J., Richlan, F., mind. In R. Menary (Ed.) The extended mind (pp. 227- Klackl, J. & Wimmer, H. (2014). Top-down and bot- 243). Cambridge, MA: MIT Press. tom-up influences on the left ventral occipito-temporal (2013). The enculturated hand. In Z. Radman (Ed.) cortex during visual word recognition: An analysis of The hand, an organ of the mind. What the manual tells effective connectivity. Human Brain Mapping, 35 (4), the mental (pp. 561-593). Cambridge, MA: MIT Press. Fabry, R. E. (2015). Enriching the Notion of Enculturation: Cognitive Integration, Predictive Processing, and the Case of Reading Acquisition - A Commentary on Richard Menary. In T. Metzinger & J. M. Windt (Eds). Open MIND:C). Frankfurt am Main: MIND Group. doi: Seassau, M., Bucci, M.-P. & Paterson, K. (2013). Reading and visual search: A developmental study in normalchildren. PLoS ONE, 8 (7), e70261-e70261.
Seth, A. K. (2015). The cybernetic Bayesian brain: From interoceptive inference to sensorimotor contingencies.
In T. Metzinger & J. M. Windt (Eds.) Open MIND.
Frankfurt a. M., GER: MIND Group. Snowling, M. J. (2000). Dyslexia. Malden, MA: Blackwell Sterelny, K. (2012). The evolved apprentice: How evolu- tion made humans unique,The Jean Nicod lectures:Vol. 2012. Cambridge, MA: The MIT Press. Turkeltaub, P. E., Gareau, L., Flowers, D. L., Zeffiro, T.
A. & Eden, G. F. (2003). Development of neural mech-anisms for reading. Nature Neuroscience, 6 (7), 767-773.
Twomey, T., Kawabata Duncan, K. J., Price, C. J. & Devlin, J. T. (2011). Top-down modulation of ventraloccipito-temporal responses during visual word recogni-tion. NeuroImage, 55 (3), 1242-1251.
Vinckier, F., Dehaene, S., Jobert, A., Dubus, J. P., Sig- man, M. & Cohen, L. (2007). Hierarchical coding ofletter strings in the ventral stream: Dissecting the in-ner organization of the visual word-form system.
Neuron, 55 (1), 143-156.
Vogel, A. C., Church, J. A., Power, J. D., Miezin, F. M., Petersen, S. E. & Schlaggar, B. L. (2013). Functionalnetwork architecture of reading-related regions acrossdevelopment. Brain and Language, 125 (2), 231-243.
Vogel, A. C., Petersen, S. E. & Schlaggar, B. L. (2014).
The VWFA: It's not just for words anymore. Frontiersin Human Neuroscience, 8.
Ziegler, J. C. & Goswami, U. (2006). Becoming literate in different languages: similar problems, different solu-tions. Developmental Science, 9 (5), 429-436.
Fabry, R. E. (2015). Enriching the Notion of Enculturation: Cognitive Integration, Predictive Processing, and the Case of Reading Acquisition - A Commentary on Richard Menary. In T. Metzinger & J. M. Windt (Eds). Open MIND:C). Frankfurt am Main: MIND Group. doi:

Source: http://loft9492.serverloft.de:9181/openmind/papers/enriching-the-notion-of-enculturation-cognitive-integration-predictive-processing-and-the-case-of-reading-acquisition-a-commentary-on-richard-menary/at_download/paperPDF

Die schwangerschaft

Die Geburten meiner zwei Töchter – einmal ohne und einmal mit Traude. Kein Vergleich! BEL und Sophie Licht! Die Schwangerschaften Ich habe zwei gesunde Kinder (Sophie-Therese 2003, Amelie-Louise 2006) zur Welt gebracht und dennoch lässt sich dieses Kapitel ganz kurz halten: ich bin/war sehr gerne schwanger! Ich konnte es jedes Mal unheimlich genießen


Management of Biochemically Recurrent Prostate Cancer After Local Therapy: Evolving Standards of Care and New DirectionsChanning J. Paller, MD, and Emmanuel S. Antonarakis, MD Drs. Pal er and Antonarakis are Assistant Abstract: Among men treated with prostatectomy or radiation Professors of Oncology at the Sidney