We agree with Brette's assessment that the coding metaphor has become more problematic than helpful for theories of brain and cognitive functioning. In an effort to aid in constructing an alternative, we argue that joining the insights from the dynamical systems approach with the semiotic framework of C. S. Peirce can provide a fruitful perspective.

If sensory organs encode environment, this code must be decoded to perception. The currently dominant theory of perception – predictive coding – assumes a “Bayesian decoder,” a probability function, which will present (to whom?) an optimal guess, given previous encodings of the environment – old codes testing new codes. Such a process would delay perception noticeably. This is inconsistent with the perception of music, which for several reasons must be direct.

First, I argue that there is no agreement within non-classical cognitive science as to whether one should eliminate representations, hence, it is not clear that Brette's appeal to it is going to solve the problems with coding. Second, I argue that Brette's criticism of predictive coding as being intellectualistic is not justified, as predictive coding is compatible with embodied cognition.

Brette's criticism of the coding metaphor focuses on its presence in neurosciences. We argue that this problematic view, which we call “encodingism,” is pernicious in any model of cognition that adopts it. We discuss some of the more specific problems it begets and then elaborate on Brette's action-based alternative to the coding framework.

The challenges raised in this article are not with information theory per se, but the assumptions surrounding it. Neuroscience isn't sufficiently critical about the appropriate ‘receiver’ or ‘channel’, focuses on decoding ‘parts’, and often relies on a flawed ‘veridicality’ assumption. If these problematic assumptions were questioned, information theory could be better directed to help us understand how the brain works.

My commentary draws on extensive arguments against “coding in the brain” developed by my neuroscience mentor, the late Eugene Sachs, who summarized them as follows: “[T]he energy in the signal is the only code there is for information…. The code is the same for each cell, but each cell's location is different, and location is the only basis for significance” (p. 13).

Brette contends that the neural coding metaphor is an invalid basis for theories of what the brain does. Here, we argue that it is an insufficient guide for building an artificial intelligence that learns to accomplish short- and long-term goals in a complex, changing environment.

Brette criticizes the notion of neural coding because it seems to entail that neural signals need to “decoded” by or for some receiver in the head. If that were so, then neural coding would indeed be homuncular (Brette calls it “dualistic”), requiring an entity to decipher the code. But I think Brette's plea to think instead in terms of complex, interactive causal throughput is preaching to the converted. Turing (not Shannon) has already shown the way. In any case, the metaphor of neural coding has little to do with the symbol grounding problem.

It has been widely asserted that humans have a “Bayesian brain.” Surprisingly, however, this term has never been defined and appears to be used differently by different authors. I argue that Bayesian brain should be used to denote the realist view that brains are actual Bayesian machines and point out that there is currently no evidence for such a claim.

Brette argues that coding as a concept is inappropriate for explanations of neurocognitive phenomena. Here, we argue that Brette's conceptual analysis mischaracterizes the structure of causal claims in coding and other forms of analysis-by-decomposition. We argue that analyses of this form are permissible and conceptually coherent and offer essential tools for building and developing models of neurocognitive systems like the brain.

Besides failing for the reasons Brette gives, codes fail to help us understand brain function because codes imply algorithms that compute outputs without reference to the signals' meanings. Algorithms cannot be found in the brain, only manipulations that operate on meaningful signals and that cannot be described as computations, that is, sequences of predefined operations.

Brains are information processing systems whose operational principles ultimately cannot be understood without resource to information theory. We suggest that understanding how external signals are represented in the brain is a necessary step towards employing further engineering tools (such as control theory) to understand the information processing performed by brain circuits during behaviour.

Brains enjoy a bodily life. Therefore animals are subjects with a point of view. Yet, coding betrays an anthropomorphic bias: we can, therefore they must. Here I propose a reformulation of Brette's question that emphasizes organismic perception, cautioning for misinterpretations based on external ideal-observer accounts. Theoretical ethology allows computational neuroscience to understand brains from the perspective of their owners.

Shannon's theory lays the foundation for understanding the flow of information from world into brain: There must be a set of possible messages. Brain structure determines what they are. Many messages convey quantitative facts (distances, directions, durations, etc.). It is impossible to consider how neural tissue processes these numbers without first considering how it encodes them.

The long-enduring coding metaphor is deemed problematic because it imbues correlational evidence with causal power. In neuroscience, most research is correlational or conditionally correlational; this research, in aggregate, informs causal inference. Rather than prescribing semantics used in correlational studies, it would be useful for neuroscientists to focus on a constructive syntax to guide principled causal inference.

Representation and computation are the best tools we have for explaining intelligent behavior. In our program, we explore the space of representations present in the mind by constraining them to explain data at multiple levels of analysis, from behavioral patterns to neural activity. We argue that this integrated program assuages Brette's worries about the study of the neural code.

Brette criticizes the notion of neural coding as used in neuroscience as a way to clarify the causal structure of the brain. This criticism will be positioned in a wider range of findings and ideas from other branches of neuroscience and biology. While supporting Brette's critique, these findings also suggest the need for more radical changes in neuroscience than Brette envisions.

We extend the discussion in the target article about distinctions between extrinsic coding (external references to known things, as required by information theory) and the alternative we and the target article both favor, intrinsic coding (internal relationships within sensory and motor signals). Central to our thinking about intrinsic coding is population coding and the concept of high-dimensional neural response spaces.

The neural coding metaphor is so ubiquitous that we tend to forget its metaphorical nature. What do we mean when we assert that neurons encode and decode? What kind of causal and representational model of the brain does the metaphor entail? What lies beneath the neural coding metaphor, I argue, is a bureaucratic model of the brain.

The convincing argument that Brette makes for the neural coding metaphor as imposing one view of brain behavior can be further explained through discourse analysis. Instead of a unified view, we argue, the coding metaphor's plasticity, versatility, and robustness throughout time explain its success and conventionalization to the point that its rhetoric became overlooked.