[1]
More precisely, according to Menary (2014, p. 293) it is scaffolded learning that renders LDP possible in the course of cognitive development of individuals: “Both structural and functional plasticity can result from both endogenous and exogenous sources, but here the focus is on structural and functional changes driven by scaffolded learning.”
[2]
On a neuronal level of description, hierarchical generative models are assumed to be neuronally realized by multiple connections across low- and high-level cortical areas. Each level within the cortical hierarchy is connected to the next subordinate and supraordinate level, thereby ensuring effective inter-level message passing (cf. Hohwy 2013, pp. 67f). According to Clark (2013, p. 187), predictive generative models are implemented in “a kind of duplex architecture”. This means that there are distinct neuronal units dedicated to the representation of predictions of environmental (or bodily) causes, so-called representation units, on the one hand, and those dedicated to the encoding of prediction error, so-called error units, on the other (cf. ibid.; Friston 2005, p. 829). To date, a detailed account of the concrete neuronal realization of these functionally distinct units of message-passing is still missing (cf. ibid.). However, it is hypothesized that representation units might correspond to superficial pyramidal cells, while error units might correspond to deep pyramidal cells (cf. Friston et al. 2012, p. 8; see also Clark 2013, pp. 187f).
[3]
The notion of two functions having “a different direction of fit” originates in J. L. Austin’s (1953, p. 234) speech act theory and in G. E. M. Anscombe’s (1963, p. 56) example illustrating how words and states of affairs can relate to each other. I would like to thank Thomas Metzinger for pointing out the philosophical history of this notion.
[4]
Thanks to Jennifer M. Windt for raising this point.
[5]
Thanks to an anonymous reviewer for helpful suggestions on this issue.
[6]
Thanks to Richard Menary for raising this important point in personal communication.
[7]
This consideration was put forward by Richard Menary in personal communication.
[8]
This phenomenon is also known as orthographic depth. For a recent review, see Richlan (2014).
[9]
See also Dehaene (2010, p. 219), Dehaene (2011, p. 26), and Frith (1985, p. 307).