Cognitive functions are not reducible to biological ones: the case of minimal visual perception

We argue that cognitive functions are not reducible to biological functionality. Since only neural animals can develop complex forms of agency, we assume that genuinely cognitive processes are deeply related with the activity of the nervous system. We first analyze the significance of the appearance of the nervous system in certain multicellular organisms (i.e., eumetazoa), arguing that it has changed the logic of their biological organization. Then, we focus on the appearance of specifically cognitive capacities within the nervous system. Considering a case of a minimal form of perceptual representation (as it happens in the visual system of cubozoan medusae), we analyze the specific functional role of this minimal form of (cognitive) activity in relatively earlier nervous systems, arguing that though this role is only understandable within a biological organization, yet it is not reducible to the underlying biological functionality. Finally, we conclude that the appearance of cognition is in turn linked to the emergence of an autonomous neurodynamic domain and a qualitative change in body complexity.

1. Some versions of the teleosemantic account consider that (for a biological structure) to play a role in a naturalistic account of mental content, the relevant selection process needs not be a natural selection operating over an evolutionary span of time—in some recent versions it is even not necessary to include a reproductive process. For example, Garson (2017) says that for a trait to possess a function, mere differential retention is sufficient regardless of whether it suffices for natural selection. And he argues that the idea of “natural selection” required for grounding a theory of functions does not need to talk of reproduction but just of “retention” among a set of entities constituting a “population”, which he defines as follows: “What sorts of interactions make a collection of individuals a population? Members of a population must engage in fitness-relevant interactions, whether competitive or cooperative. My behavior must have some effect on your chances of survival or reproduction, and vice versa. Clearly, we can place further restrictions on this idea, but this suffices for my purposes” (p. 536). But this is just half of the requirements for natural selection. “Fitness” in biology does not allow to avoid (the concept of) reproduction (see Lewontin 1970; Godfrey-Smith 2007): it is defined in terms of reproductive success. It describes individual reproductive success and is equal to the average contribution to the gene pool of the next generation that is made by individuals of the specified genotype or phenotype. In other words, fitness in the context of ‘natural selection’ requires and means survival of the form (phenotypic or genotypic) that will leave the most copies of itself in successive generations.

2. More technically, as formulated by Sebastian (2017), the idea is that “necessarily, some contentful states (….) require that (i) the state has been selected for indication in evolution or (ii) that it is the result of a device that has been selected for indication in evolution.” (p 319). See also Schulte (2015) for a teleosemantic defense of perceptual representations.

3. “Deflationist theories are part of a long, failed tradition of assimilating truth and accuracy to contribution to practical success, and falsity and inaccuracy to practical failure. Error need not be a failure or frustration of any independently identifiable biological function. Representational success need not fulfill any biological function.” (Burge 2010: 301).

4. And some other times, Burge is not clear and mixes things up, as he states that: “The idea is to explain error in terms of failure to fulfill biological function, and to explain veridicality in terms of fulfillment of biological function. Perceptual systems and some of their states surely have biological functions. Further, biological function is relevant to understanding both the content of perceptual states and their relation to actions that serve biological needs.” (Burge 2010, p. 299).

5. To be fair, under this label there is a variety of proposals: Bickhard (2000, 2004), Schlosser (1998), Collier (1999), McLaughlin (2001), Christensen and Bickhard (2002), Delancey (2006), Edin (2008), Mossio et al. (2009). But for the sake of precision, here we will commit to this late specific proposal.

6. The reason behind postulating such a type of organization is that it characterizes the differences between physical-chemical and biological systems. From the perspective of the OA, a truly naturalistic account of biological functions should rely on a conceptual framework that does not already presuppose a full-blown biological organization.

7. It has been argued in Arnellos and Moreno (2021) that a minimal form of PR (actually, a minimal form of content-based visual perception) appears in perhaps one of the simplest cases of behavior based on low-resolution spatial vision in nature–the visually-guided obstacle avoidance in the cubozoan medusa Tripedalia cystophora (TC). More specifically and based on the neurophysiologically-informed behavioral data available in the literature, the authors explained that the lower lens eye in TC employs (primitive) constancy mechanisms to construct a decoupled (from the retinal stimulations) neurodynamic structure, based on which it succeeds obstacle avoidance in virtue of this neurodynamic structure’s exploitable structural similarity with its target objects in the environment. Considering that the lens eyes of TC are the most basic case of true camera-type eyes in nature, obstacle avoidance in TC is guided by a minimal form of perceptual representation.

8. This stress on the qualitative novelty, however, is not to deny that any great emergent novelty in the history of the universe—the appearance of life or of consciousness, for example—is preceded by the appearance of a set of intermediate steps that in some sense “prepare” the “qualitative jump”. So, when we defend that we face a qualitative, emergent change in biological evolution, we are committed both to explain (1) why this change is radically new and plenty of far-reaching consequences, and (2) how this change could have been appeared, which forms of organization and trends have driven its appearance and why this appearance is not something “magical” or “out of the blue” (Arnellos and El-Hani ). In sum, we are committed to a naturalistic account (which is not a reductionist one) of the qualitative novelty.

9. In many cases, with the participation of other cells, like glial cells. In this paper, we are not interested in the generic concept of neural network (i.e., a population of interconnected neurons in a Petri dish), but in an organismically embodied set of neurons.

10. “The nerve impulse is a stereotyped electrical signal into which all modes of sensory stimuli are translated—photons, chemicals, heat, sound waves, and other mechanical types of energy. This common currency can be used to “map” the external world within the nervous system—that is, to create a “model” of the rapidly changing world surrounding the moving animal. Furthermore, nerve impulses are also used to map the internal environment—the organism’s body—within the same system. Importantly, the common currency makes it possible to bind together and integrate different types of stimuli that carry information from a single modality and, even more remarkably, from several different modalities. The nervous system can accomplish these feats because of the combination of its unique structural and functional properties: each neuron can form synapses with many other neurons, including very distant ones; electrical signaling is rapid; and there is an electric “common language.” (Ginsburg and Jablonka 2019:254).

11. A Central Pattern Generator—also called ‘pacemaker’—is defined as a set of neurons that operate together to generate a motor program. The term motor program identifies the time- and target-oriented output activity of the ensemble, which serves as a command to the muscles (Balaban et al. 2015 p. 42).

12. Since its ontogenetic origin, the NS starts modulating organ formation. For a more detailed description of the role played by the NS in development, see Cabej (2013).

13. The term “integration” can be broadly understood as a phenomenon that occurs when a set of different and initially independent processes begin to functionally cooperate and share their local functions, leading to the establishment of a wider collective and unique functional system/organization, in which some functional constraints of the constituent processes are interlocked and control each other. The emergence of a new functionally-integrated organization, therefore, requires a functional redefinition of the original processes (see also Arnellos and Moreno 2021, p. 6).

14. As Christensen (2007) has argued, “functional integration cannot be achieved, even partially, unless local controllers have information processing capacities that are able to determine precisely what the global context is and what the appropriate local response is”. And he adds that this requires “gathering information from wide-ranging sources and subjecting it to processing to extract highly specific control information. Regulatory systems provide spatially and/or temporally and/or qualitatively precise delivery of control signals across wide regions” (ibid., p. 267).

15. Advances in the neurosciences have revealed the staggering complexity of even “simple” nervous systems. This is reflected in their function, evolutionary history, structure, and the coding schemes they use to represent information. Any realistic notion of brain complexity must incorporate, primarily, the highly nonlinear, nonstationary, and adaptive nature of the neuronal elements themselves and, secondly, their nonhomogeneous and massive parallel patterns of interconnection whose “weights” can wax and wane across multiple time scales in behaviorally significant ways (Koch and Laurent 1999).

16. With this term we refer to the process by which, during evolution, the NS develops an operational center or “brain”. See also the discussion below in Sect. 3.1 ("Looking for the earlier ‘embrained’ NS").

17. Interestingly, the epithelium is also the place of the immune defense in Cnidaria and, according to Bosch (2016) “it represents the ancient system of host defense (in Cnidaria) and point to an origin of innate immunity in the Eumetazoan ancestor” (p. 397).

18. The role of the NS in the developmental processes of all eumetazoan has been stressed by Cabej (2013). Jekely et al. (2015) have also argued that one of the earlier roles of primitive NSs was the control of development. As they have pointed out, “the control of physiology (is) a second major role of nervous systems, (and) includes phenomena such as circadian and circalunar clocks, the control of metabolism, digestion and diuresis. Some borderline cases can be categorized either as behaviour or physiology, such as the feeding and peristaltic gut motion in a sea anemone. The control of development, also neglected in many discussions, is a fundamental role of all animal nervous systems. It includes the control of growth and metamorphosis, along with phenomena such as moulting and regeneration. These processes are controlled by hormonal signals emanating from the nervous system.” (ibid., p 3).

19. Most likely, in the earlier NSs this monitoring capacity was genetically fixed, but as they became more complex, they began to self-organize their own norms of integration.

20. Recent studies date back the origin of Cnidaria to 700 MYA (Schwentner and Bosch 2015; Dohrmann and Wörheide 2017.

21. In certain medusae, for example, pacemakers comprise approximately 2 × 30 giant neurons connected in ways that form subsystems (see e.g., Skogh et al. 2006).

22. As mentioned, even before this, the embryonic NS would also start to monitor the developmental processes of the animal.

23. For example, in cubozoan medusae central pattern generators (or pacemakers) are located close to the insertion of the rhopalial stalk (Skogh et al. 2006).

24. Various electrophysiological and neurophysiological studies have consistently shown that the visual control of the pacemaker in Cubozoa is not just an additive result of individual inputs; nor is it simply the result of a hierarchy between the stimulated eyes. Rather, some form of integration must be taking place within the eyes and the pacemaker (see Garm and Bielecki 2008; Garm and Mori 2009, and Parkefelt and Ekström 2009 for details).

25. Given that these functions are related (as it is argued in Arnellos and Moreno 2021) with perceptual activities—in a minimal but quite genuine form—we consider them ‘cognitive’.

26. This evolution includes, of course that of sensors and effectors (and in fact, of the whole body).

27. For a recent detailed description of the neural organization of Aglantha digitale, see Norekian and Moroz (2020).

28. As Ginsburg and Jablonka (2019) have pointed out, “some responses are complex and suggest that discrimination, integration, and action selection are enabled by their nervous system. For instance, Aurelia’s response to touch depends on both mechanoreceptors and chemoreceptors, and its reaction depends on the type of touch. While a touch by a conspecific elicits a pause in swimming followed by a resumption of the previous swimming pattern, a touch by a silicon ball initiates reorientation and swimming upward. Catching prey also elicits different behavior depending on the conditions (…) The mechanisms underlying this behavior require modulations of default exploratory activity.” (p. 281).

29. This is basically a direct signal-motor, reflex-arc type of response, but A. digitale is also able to swim gently and to display other forms of behavior by integrating different simple signals sent from its diverse sensors to its pacemakers (Mackie 2004).

30. Notice that these NDSs (1) are autonomously generated by those NS capable to achieve a certain degree of integration; (2) they are there insofar as they contribute to the maintenance of this specific neurodynamic organization (namely, that which support successful distal obstacle avoidance swimming).

31. For an analysis of the neurophysiological characteristics and properties of the process of obstacle avoidance in the cubozoan Trypedalia cystophora, see Arnellos and Moreno (2021), pp. 4–6.

32. This is known as the underdetermination problem in visual perception, which is solved through the application of constancy mechanisms by the sensory state (for details see Burge 2010, Schulte 2020, and Arnellos and Moreno 2021).

33. Regardless of any eventual subsequent biological damage (i.e., a skin injury).

34. In these very primitive NSs we can talk about a certain “basic form of self-maintenance”—in the sense only of a dependence of the neural structure on the activity of the neurons—but not in the deeper sense of the capacity of generating constraints on which the far-from-equilibrium neurodynamic organization will depend.

35. The idea that the neurodynamic organization may constitute an autonomous domain has been first presented in Barandiaran and Moreno (2006a, b) and then further developed in Barandiaran (2008). However, the way this idea is presented here differs in several points and aspects with that of the aforementioned works.

36. Of course, not all neural structures have a cognitive function.

37. That they occur at the neurodynamic domain (NdD) does not mean that they are disembodied, on the contrary. What we want to stress is that (a) they require a neurally controlled body; and (b) that within a neurally controlled body, they occur mainly at the neural level. The digestive processes in an animal require also a neural control, but they do not mainly occur at the neural level.

38. This process has been accompanied also by an increase in the integration of the body (Arnellos and Moreno 2016; Arnellos and Keijzer 2019).

39. Actually, we believe that when Ginsburg and Jablonka (2019) propose that the expansion of associative learning (which enables an organism to ascribe motivational value to a novel, compound, non-reflex-inducing stimulus or action) drove the Cambrian explosion and its massive diversification of organisms, they are also suggesting this very issue, namely, that, at some point, the NS managed to organize its own functional domain, and that neural evolution has been the specific and predominant driving force of animal evolution ever since.

The authors acknowledge funding from the Basque Government (Project: IT1228-19 for AA and AM), Ministerio de Ciencia, Innovación y Universidades, Spain (research project PID2019-104576 GB-I00, for AM), and the John Templeton Foundation (research project ‘Agency, Directionality & Function’ - Science of Purpose for AA and AM).

Authors and Affiliations

1. Department of Product and Systems Design Engineering, Complex Systems and Service Design Lab, University of the Aegean, Syros, Greece

   Argyris Arnellos

2. Department of Logic and Philosophy of Science, IAS-Research Centre for Life, Mind and Society, Donostia - San Sebastián, Spain

   Argyris Arnellos

3. Donostia International Physics Center (DIPC), Donostia - San Sebastián, Spain

   Alvaro Moreno

Corresponding authors

Correspondence to Argyris Arnellos or Alvaro Moreno.

Competing Interests

The authors declare that the research was conducted in the absence of any commercial, financial or non-financial relationships that could be construed as a potential conflict of interest, and they have no such relevant interests to disclose.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Reprints and permissions