A major contribution of Bruineberg and colleagues' analysis is to reduce potential time and confusion involved in trying to work out what might be done in practice with Markov blankets. In particular, they provide succinct comparative analyses of the many different descriptions of Markov blankets. Importantly, their comparative analysis includes illustrative cases, which is a proven strategy for improving learning (Roelle & Berthold, Reference Roelle and Berthold2015). However, Bruineberg and colleagues' analysis reveals that both Pearl and Friston blankets are of limited usefulness for practice. In particular, Bruineberg and colleagues clarify that Pearl blankets is only an auxiliary construct that can be used to describe conditional independence on random variables. They also clarify that what can be done with Friston blankets in 2021 is a matter of ongoing debate that includes dispute over conceptual issues and mathematical details.

To use an automotive vehicle analogy for brevity, Bruineberg and colleagues' analysis can be distilled into the following summary: Pearl blankets can contribute to describing what is going on “under the hood” while Friston blankets can contribute to describing what is going on “in and around the hood.” From an instrumentalist systems design perspective, both Pearl and Friston blankets are human ascriptions made in order to model real-world systems. Colloquial description using the “hood” analogy is not appropriate for natural science and social science, but is appropriate for action science within which the need for practical framings is emphasized (Friedman & Putman, Reference Friedman, Putman, Coghlan and Brydon-Miller2014).

The limited practical usefulness of distinguishing between Pearl and Friston blankets can be illustrated with the following example related to functional disorders: that is, medical conditions without complete medical explanation that impair normal functioning of bodily processes (Stone, Reference Stone2009). Better healthcare systems are needed to provide support for people suffering with functional disorders (Stone, Reference Stone2016). Gait issues are involved in functional disorders (Espay et al., Reference Espay, Aybek, Carson, Edwards, Goldstein, Hallett and Morgante2018). Gait is related to personality in ways that are not fully understood (Sun et al., Reference Sun, Wu, Shen, Yang, Li, Liu and Chen2018). This ascription problem is exacerbated by the difficulty of defining where one personality type ends and another begins (Haslam, Reference Haslam2019). Furthermore, gait is related to memory in ways that are not fully understood (Michalak, Rohde, & Troje, Reference Michalak, Rohde and Troje2015). This ascription problem is exacerbated by the difficulty of defining what aspects of memory are in the mind and what aspects are in the body (Tozzi, Reference Tozzi2014). Thus, distinguishing between what is “under the hood” and what is “in and around the hood” is very difficult. For example, are personality type and body memory “under the hood” while gait is “in and around the hood”? If so, how is the fascia's connection of bones and muscles in gait related to the fascia system holding body memories (Tozzi, Reference Tozzi2014)? Should the same fascia be described with both Pearl and Friston blankets? This example illustrates that distinguishing between Pearl and Friston blankets does not necessarily make modelling complex systems any easier. Nor does it necessarily end potential argument that there should only be one type of Markov blanket or potential argument that Markov blankets could be superseded, for example, by causal blankets (Rosas, Mediano, Biehl, Chandaria, & Polani, Reference Rosas, Mediano, Biehl, Chandaria, Polani, Verbelen, Lanillos, Buckley and De Boom2020).

Although reading The Emperor's New Markov Blankets reveals limitations of Pearl and Friston blankets, this does not mean that nothing practical can be done. In particular, it has been argued that it may be useful to frame systems in terms of constructs such as Markov blankets, but without applying all technical details and associated mathematics (Fox, Reference Fox2021). There is precedence for this in Kurt Lewin's force field analysis diagram being used widely, but many technical details and associated mathematics of his field theory not being used. This is a relevant example as force field analysis is applied to determine interactions between forces for change and forces against change during the evolution of organizations (Burnes & Cooke, Reference Burnes and Cooke2013).

Apropos, Bruineberg, and colleagues provide another useful contribution with their Figure 5, which has potential as a boundary object: that is, as a means of providing meaningful information to different parties who have different backgrounds (Bowker & Star, Reference Bowker and Star2020). In particular, their Figure 5 provides a succinct explanation of interactions between main constructs in the free energy principle. Importantly, it does not involve single line demarcations, which can facilitate the reification fallacy (Mishra & Mishra, Reference Mishra and Mishra2010). Moreover, it has much potential to provide the basis for an interactive multimodal symbol system, which can facilitate effective communication among people who have different backgrounds (Fox, Moreno, & Vahala, Reference Fox, Moreno and Vahala2019) during the design and implementation of biosocial–technical systems (Fox, Griffy-Brown, & Dabic, Reference Fox, Griffy-Brown and Dabic2020). For example, facilitate effective communication among individuals suffering with functional disorders and healthcare practitioners who could provide them with support (Allen, Reference Allen2009; Stone, Reference Stone2016).