Brette's thesis is that one main metaphor, that is, neural coding, has generated one main narrative, which carries a dominant standpoint on brain behavior that precludes other views, becoming almost undiscussed. By contrast, a scrutiny of the functioning of the coding metaphor shows a multipatterned usage of the metaphor and a broader metaphorical narrative. This plasticity of the coding metaphor clarifies why its pervasive and persuasive rhetoric spread across time almost invisibly.
Metaphors typically help us to understand complex ideas through more familiar elements by means of cross-domain mapping between “target” and “source” domains (Lakoff & Johnson Reference Lakoff and Johnson1980a). In “the brain is a computer,” for instance, the target is the brain and the source is the computer; thus, we can imagine the brain in terms of elements of the computer (i.e., computing and programming). Both target and source domains are essential: they shape the metaphor and affect its meaning. Substituting “brain” with “mind” gives the metaphor “the mind is a computer,” which changes its meaning; and replacing the source “computer” with “computational machine” gives “the brain is a computational machine,” which was considered not metaphorical (Eliasmith Reference Eliasmith2003). The multiple targets of the “electric switch” metaphor applied to epigenetics were considered a sign that the meaning of epigenetics is unclear and that it is a field in progress (Stelmach & Nerlich Reference Stelmach and Nerlich2015).
In Brette's examples, instead of a unified view there are both different targets and sources of the coding metaphor. The targets vary, referring mainly to “treating information” (e.g., Somjen Reference Somjen1972, in Brette, sect. 1, numbered list, item 2); “perceptions” (e.g., Pouget et al. Reference Pouget, Dayan and Zemel2003, in Brette, sect. 2.1, para. 3); “representations” (e.g., Bickhard Reference Bickhard2009, in Brette, sect. 3.1, para. 7); “neural activity” (Ashida & Carr Reference Ashida and Carr2011, in Brette, sect. 1, numbered list, item 1); and “brain activity” (Jazayeri & Movshon Reference Jazayeri and Movshon2006, in Brette, sect. 1, numbered list, item 2). The source domain “coding” is associated with “encoding” (e.g., Brette, sect. 2.4, para. 6) and “decoding” (e.g., Brette, sect. 4.1, para. 1). A broader metaphorical narrative develops, which more generally targets the ideas of “reading” and “communication,” connecting “coding” with “information” and with “encoding and decoding messages” to end with another metaphor about the brain “deciphering” hieroglyphs (Brette, sect. 5.2, last para.). This broad narrative, together with the varied targets and sources, indicates that the different voices in Brette's examples might stand for varied views of the coding metaphor instead of a unified one. This plasticity fulfills the well-known trait of metaphor in discourse, which needs to be “robust enough to carry certain implications from one context to another, but at the same time flexible enough to allow for different formulations in different contexts” (Hellsten Reference Hellsten2003).
For example, the two targets of “neural activity” and “brain activity” allow for two different meanings. In most of Brette's examples of neural coding (e.g., sect. 4.1, para. 1, 2), the coding metaphor entails a dualistic perspective on brain functioning that duplicates the agents, e.g., “the brain,” which makes decisions about stimuli deriving from “populations of neurons.” Here the coding narrative presents a scenario in which the target “brain activity” stands for higher hierarchical functions (like “deciding,” Brette, sect. 4.1, para. 1) than the target “neural activity.” In this scenario, the neurons encode messages while the brain decodes them. In a different scenario, instead, the two targets are conflated because neural activity is a metonymy for brain activity (there are no “‘little decoders’ inside the head. … Rather, they are embedded in the synaptic weights between neighboring neurons” (Eliasmith Reference Eliasmith2003, p. 506). Here, the coding metaphor applies equally to both targets (i.e., brain and neurons). To get a clearer picture of the coding narrative, the metaphor should be specified according to its specific targets and sources in any given context so that the diversity of the underlying theoretical models, as well as the flexibility of the coding metaphor, can be better appreciated instead of concealed.
Moreover, the coding metaphor does not stand alone in the debate. At the core of the coding narrative there are two primary metaphorical models: “the brain is a computational machine” and the “information flow” of cognition (Dretske Reference Dretske1981; Searle Reference Searle1980). For more than 60 years, these models outlined different views about cognition and brain functioning according to three main approaches: symbolicism, which focuses on a formal view of computation via symbols; connectionism, which emphasizes the parallelism of neural structures with networks; and dynamicism, which is an “environmentalist” view pointing to interactive and time-laden cognitive behaviors (Eliasmith Reference Eliasmith2003). Outlining Brette's criticism on neural coding against these pervasive cognitive metaphorical models frames his position within an older debate, that is, against a symbolicist and a computational view of the brain and in line with the dynamic approach.
We now turn to Brette's second point, that is, regarding the coding metaphor that establishes one dominant narrative of information that prevents imagining alternatives. Lakoff and Johnson (Reference Lakoff and Johnson1980b) clarified the Gestalt mechanism of metaphor: by highlighting a given element (such as the computational view of the brain) other elements necessarily move to the foreground. In other terms, through metaphor the complexity of a phenomenon is restricted to show only one perspective at a time (Hellsten Reference Hellsten2003). This better explains Brette's claim on the coding metaphor for which “the notion of information it carries seems to be the only possible one.” Over the years, the narrative developed by the metaphor became conventionalized and established “self-evident ways of seeing things, even to the degree where no alternatives are imagined” (Hellsten Reference Hellsten2003). It is the plasticity of the coding metaphor, that is, its robustness through change, that explains why for more than 60 years it spread and served across time and debates, allowing for its conventionalization, similarly to “brain-computer” and “information” metaphors. The pervasive and persuasive effects of the metaphorical narrative hinder the fundamental self-correcting trait of science that aims to provide counterexamples of dominant theories instead of just supporting them. This is part of the well-known circuit of metaphor in scientific discourse that entails potential risks, from misunderstanding to hype in communication, calling for more responsibility in language use, which should reinforce our vigilance about metaphor and metaphorical narratives in science (Ball Reference Ball2011; McLeod & Nerlich Reference McLeod and Nerlich2018).
Brette's thesis is that one main metaphor, that is, neural coding, has generated one main narrative, which carries a dominant standpoint on brain behavior that precludes other views, becoming almost undiscussed. By contrast, a scrutiny of the functioning of the coding metaphor shows a multipatterned usage of the metaphor and a broader metaphorical narrative. This plasticity of the coding metaphor clarifies why its pervasive and persuasive rhetoric spread across time almost invisibly.
Metaphors typically help us to understand complex ideas through more familiar elements by means of cross-domain mapping between “target” and “source” domains (Lakoff & Johnson Reference Lakoff and Johnson1980a). In “the brain is a computer,” for instance, the target is the brain and the source is the computer; thus, we can imagine the brain in terms of elements of the computer (i.e., computing and programming). Both target and source domains are essential: they shape the metaphor and affect its meaning. Substituting “brain” with “mind” gives the metaphor “the mind is a computer,” which changes its meaning; and replacing the source “computer” with “computational machine” gives “the brain is a computational machine,” which was considered not metaphorical (Eliasmith Reference Eliasmith2003). The multiple targets of the “electric switch” metaphor applied to epigenetics were considered a sign that the meaning of epigenetics is unclear and that it is a field in progress (Stelmach & Nerlich Reference Stelmach and Nerlich2015).
In Brette's examples, instead of a unified view there are both different targets and sources of the coding metaphor. The targets vary, referring mainly to “treating information” (e.g., Somjen Reference Somjen1972, in Brette, sect. 1, numbered list, item 2); “perceptions” (e.g., Pouget et al. Reference Pouget, Dayan and Zemel2003, in Brette, sect. 2.1, para. 3); “representations” (e.g., Bickhard Reference Bickhard2009, in Brette, sect. 3.1, para. 7); “neural activity” (Ashida & Carr Reference Ashida and Carr2011, in Brette, sect. 1, numbered list, item 1); and “brain activity” (Jazayeri & Movshon Reference Jazayeri and Movshon2006, in Brette, sect. 1, numbered list, item 2). The source domain “coding” is associated with “encoding” (e.g., Brette, sect. 2.4, para. 6) and “decoding” (e.g., Brette, sect. 4.1, para. 1). A broader metaphorical narrative develops, which more generally targets the ideas of “reading” and “communication,” connecting “coding” with “information” and with “encoding and decoding messages” to end with another metaphor about the brain “deciphering” hieroglyphs (Brette, sect. 5.2, last para.). This broad narrative, together with the varied targets and sources, indicates that the different voices in Brette's examples might stand for varied views of the coding metaphor instead of a unified one. This plasticity fulfills the well-known trait of metaphor in discourse, which needs to be “robust enough to carry certain implications from one context to another, but at the same time flexible enough to allow for different formulations in different contexts” (Hellsten Reference Hellsten2003).
For example, the two targets of “neural activity” and “brain activity” allow for two different meanings. In most of Brette's examples of neural coding (e.g., sect. 4.1, para. 1, 2), the coding metaphor entails a dualistic perspective on brain functioning that duplicates the agents, e.g., “the brain,” which makes decisions about stimuli deriving from “populations of neurons.” Here the coding narrative presents a scenario in which the target “brain activity” stands for higher hierarchical functions (like “deciding,” Brette, sect. 4.1, para. 1) than the target “neural activity.” In this scenario, the neurons encode messages while the brain decodes them. In a different scenario, instead, the two targets are conflated because neural activity is a metonymy for brain activity (there are no “‘little decoders’ inside the head. … Rather, they are embedded in the synaptic weights between neighboring neurons” (Eliasmith Reference Eliasmith2003, p. 506). Here, the coding metaphor applies equally to both targets (i.e., brain and neurons). To get a clearer picture of the coding narrative, the metaphor should be specified according to its specific targets and sources in any given context so that the diversity of the underlying theoretical models, as well as the flexibility of the coding metaphor, can be better appreciated instead of concealed.
Moreover, the coding metaphor does not stand alone in the debate. At the core of the coding narrative there are two primary metaphorical models: “the brain is a computational machine” and the “information flow” of cognition (Dretske Reference Dretske1981; Searle Reference Searle1980). For more than 60 years, these models outlined different views about cognition and brain functioning according to three main approaches: symbolicism, which focuses on a formal view of computation via symbols; connectionism, which emphasizes the parallelism of neural structures with networks; and dynamicism, which is an “environmentalist” view pointing to interactive and time-laden cognitive behaviors (Eliasmith Reference Eliasmith2003). Outlining Brette's criticism on neural coding against these pervasive cognitive metaphorical models frames his position within an older debate, that is, against a symbolicist and a computational view of the brain and in line with the dynamic approach.
We now turn to Brette's second point, that is, regarding the coding metaphor that establishes one dominant narrative of information that prevents imagining alternatives. Lakoff and Johnson (Reference Lakoff and Johnson1980b) clarified the Gestalt mechanism of metaphor: by highlighting a given element (such as the computational view of the brain) other elements necessarily move to the foreground. In other terms, through metaphor the complexity of a phenomenon is restricted to show only one perspective at a time (Hellsten Reference Hellsten2003). This better explains Brette's claim on the coding metaphor for which “the notion of information it carries seems to be the only possible one.” Over the years, the narrative developed by the metaphor became conventionalized and established “self-evident ways of seeing things, even to the degree where no alternatives are imagined” (Hellsten Reference Hellsten2003). It is the plasticity of the coding metaphor, that is, its robustness through change, that explains why for more than 60 years it spread and served across time and debates, allowing for its conventionalization, similarly to “brain-computer” and “information” metaphors. The pervasive and persuasive effects of the metaphorical narrative hinder the fundamental self-correcting trait of science that aims to provide counterexamples of dominant theories instead of just supporting them. This is part of the well-known circuit of metaphor in scientific discourse that entails potential risks, from misunderstanding to hype in communication, calling for more responsibility in language use, which should reinforce our vigilance about metaphor and metaphorical narratives in science (Ball Reference Ball2011; McLeod & Nerlich Reference McLeod and Nerlich2018).