The Intangible Bounds of Cognition: Adams and Aizawa Draw the Line
by Hayden Lamar
Franklin College of Arts and Sciences, Department of Philosophy
In pursuit of disproving an overeager conceptualization of active externalism, this paper considers the realism of Fred Adams and Ken Aizawa’s functionalism and applies it to Andy Clark and David Chalmers’ all-too-inclusive position on cognitive integration. By weighing the differing views between these two camps of functionalists and their respective peers, the latter’s propositions are whittled down into the more intuitive arguments of the former. Clark, Chalmers, and their contemporaries may act as the gas pushing the spatial limits of the human mind to their extremes, but this paper calls upon Adams and Aizawa to act as the brakes.
Functionalism, active externalism, tool use, cognitive bounds
Otto’s Notebook vs. A&A’s “Marks of the Cognitive”
As the interest in defining the limits of human interiority merges with skepticism about our physical limits, cognitive scientists find themselves testing the conventions of human cognition. A microcosm of this phenomena is Andy Clark and David Chalmers’ overzealous defense of active externalism, wherein they allege that the memento book of an Alzheimer’s patient is functionally equivalent to a neurotypical brain. By subjecting Clark and Chalmer’s argument to Adams and Aizawa’s tamer criteria for cognition as outlined in their two “marks of the cognitive,” this essay will prove that Otto’s notebook is not a literal extension of his mind.
Before analyzing their arguments, one must understand the extent of Clark and Chalmer’s functionalism by recognizing the principles they endorse. Primarily, Clark and Chalmers are active externalists, so they subscribe to the idea that the mind does not end where the skull or neural network does (C&C 1). This idea is why they so strongly adhere to the parity principle, which states that a process involving part of the world is actually a mental process so long as it’s functionally identical to a mental process. Both of the aforementioned concepts fall under the umbrella of functionalism, which defines minds through the functional roles of computation, cognition, and creation of mental states.
Backed by the beliefs above, Clark and Chalmers put forth a man with Alzheimer’s named Otto, who uses a notebook to keep track of important information to supplement his weak and fading memory (C&C 12). They compare Otto and his notebook to Inga, a neurotypical woman who doesn’t need any kind of material help to mentally navigate her life. They then measure Otto’s notebook and Inga’s brain as cognitive tools, asking whether they meet the four criteria for something to be part of a mind. The criteria are as follows: constancy, which means the cognitive tool is in constant use; direct availability, which means that it’s easily/practically mentally available; automatic endorsement, meaning the cognitive tool has high efficacy; and previous endorsement, which establishes the ethos of the cognitive tool (C&C 13-16). Because both Inga’s brain and Otto’s notebook satisfy the aforementioned requirements, Clark and Chalmers consider them functionally equivalent. According to them, “the essential causal dynamics of the two cases mirror each other precisely” (C&C 13). Because their respective use generally results in the same output, like recollecting addresses, remembering dates of events, etcetera, they conflate Otto’s notebook with Inga’s brain. Of course, this hypothetical is supposed to explain active externalism and consequently defend the parity principle to the extent that it proves Otto’s notebook is a functionally integrated part of his mind.
However, looking at Fred Adams and Ken Aizawa’s essay “The bounds of cognition,” one can see how their parameters for cognitive integration and physical boundaries for the mind contrast those of Clark and Chalmers. In specifying what any tool or process needs in order to be cognitively integrated, they refer to what they call the “marks of the cognitive”: the two necessary conditions for cognitive processes. The first condition is that “cognitive states must involve intrinsic, non-derived content,” meaning that the content that arises from any given process needs to arrive purely within one’s mind and be intrinsically representational (A&A 48). In this case, one can take “representational content” to mean content that is tied to or obviously representative of a belief about the external world. The second mark of the cognitive is that “cognitive processes are causally individuated,” which is to say that there is some “real causal process underlying them” (A&A 51). Although the point of the second mark is more complex than the first, it essentially boils down to there being an intentionality behind cognitive processes. One could think of it as the first-person understanding and willfulness behind our human cognition–features that are absent in the “thinking” processes of computers or robots that are merely programmed to perform a certain task. Together, these two marks of the cognitive clarify what an extracranial process would need in order to be functionally integrated into the mind.
To apply these criteria in a critical evaluation of Clark and Chalmers’ argument is to reveal that Otto’s notebook is certainly unworthy of being considered cognitively incorporated. Neither mark of the cognitive is present in the notebook’s function. First, it does not contain any non-derived content. Adams and Aizawa state that “strings of symbols on the printed page mean what they do in virtue of conventional associations between them and words of language” and that a reader’s ability to understand those associations is the sole reason symbols have representational value at all (A&A 48). This means that any understanding one gets from written or spoken language is non-derived content. Because everything in Otto’s notebook has obviously been written down or otherwise transcribed, the supplemental information it provides him is necessarily derived content. Sam Clarke, another extended mind theorist, attests to Adams and Aizawa’s issue with the representational value of Otto’s notebook and its alleged extension. Stating in his essay “Beyond the icon: Core cognition and the bounds of perception,” he characterizes Adams and Aizawa’s perspective as one concerned largely with the modularity of cognitive systems, noting that they may determine that “perception cannot be usefully demarcated by the format of its representations full stop” (Clarke). Essentially, Clarke recognizes that Adams and Aizawa take issue with the idea that text and lingual units impact perception rather than the other way around. In doing so, he acknowledges that his peers’ litmus for non-derived content is born from a refusal to assign inherent meaning to lingual units, encapsulating their first mark of the cognitive and succinctly showing why it cannot support Clark and Chalmers’ position.
Neither does Otto’s notebook satisfy the second mark of the cognitive, the need for a unified underlying causal process. Adams and Aizawa contend that cognitive processes must be undergirded by cognitive science, since “like other natural domains, namely, the cognitive must be discriminated on the basis of underlying causal processes” (A&A 51). However, there’s no clear way to apply the logic of cognitive science to what’s contained within Otto’s notebook. Taking cognition to be a “natural kind” as Adams and Aizawa propose, the information in Otto’s notebook would need to transcend language, culture, and other fluid properties in order to be considered cognitive. But it does none of those things, so clearly, the notebook is more akin to tool use than to an integrated cognitive process. This is because it isn’t a cognitive process at all, but rather an epistemic one. Put simply, Otto’s notebook is more reminiscent of one using a calculator than the intracranial process of mental math. It is an outsourcing of epistemic material, a reliance on a tool–-not an integrated part of his mind.
Still, other cognitive scientists like Matthew Overstreet diverge from Adams and Aizawa’s position by asserting that “social and material context lies at the very heart of the extended mind program” (Overstreet). This is inherently aligned with Clark and Chalmers, who must accept that external processes do not need to be superlingual nor supercultural to be cognitive. However, this once again exposes the epistemic nature of Otto’s notebook and renders it markedly non-cognitive according to Adams and Aizawa’s criteria. However, Matthew Overstreet corroborates Adams and Aizawa’s point about the cognitive qualities of the notebook in his own exploration of how various supplemental technologies can affect the functionality of writing. Though his essay “Writing as extended mind: Recentering cognition, rethinking tool use,” takes a position more favorable to Clark and Chalmers, Overstreet agrees that “the limited capacity of human working memory is undoubtedly one of the most important biological constraints a writer faces [and] given this situation, behavior which works to offload cognitive function is of great value to a writer” (Overstreet). This idea in itself isn’t incongruent with Adams and Aizawa’s argument; conversely, the idea that cognitive offloading is a benefit of writing suggests that writing itself acts as a tool, not an extension of the mind. Overstreet also states that the “writing act itself is an act of problem-solving, in the simple sense that it is sparked by a desire to move from present to intended state” (Overstreet). The language used here is critical: the idea that writing is a supplement to the finite space of one’s mind, that it is a means to an end, intuitively demonstrates its instrumentality.
Regardless of Overstreet’s overarching idea that viewing writing as an extension of the mind may be beneficial, he gives significant merit to the view that writing is purely a technology and not necessarily an extension of the mind. He goes as far to reference the brand of functionalism endorsed by Clark and Chalmers. Referencing the instrumentality of a supplemental writing program, he states that individuals may “transform the physical world from a problem space to a problem-solving resource, as Andy Clark might say” (Overstreet). But once again, I return to Adams and Aizawa’s insinuation that these “problem-solving resources” are best described as tools, not functionally integrated processes. The labelling of physical media or instruments as modes of external cognition jumps the gun–for the calculator is no more the prefrontal cortex than a browser’s tab-organizer is a network of synapses.
This isn’t to say that Clark and Chalmers (or Overstreet, necessarily) would see a calculator as any less cognitively integrated than Otto’s notebook. However, that is the very issue: their examples of extracranial cognition are too easily rebutted by references to natural processes or mere tool use. As follows, it isn’t that Adams or Aizawa state that active externalism is impossible, but rather that Otto’s notebook is a terrible attempt at proving it. They say that “the cognitive processes we find in the real world all happen to be brain bound. It appears to be just a contingent empirical fact that cognitive processes are not transcorporeal processes” (A&A 46). They are not directly disagreeing with Clark and Chalmers’ functionalist perspective. In fact, they are only reasserting that tool use is unlikely to be coextensive with cognition (A&A 47). The point is that Otto’s notebook cannot be deemed cognitively equivalent to Inga’s brain simply because they share the same epistemic function.
Ultimately, what Adams and Aizawa do in their argument is reject the parity principle, not disprove the idea of active externalism. Their two marks of the cognitive are intuitive ways to explain integrated cognition without precluding the possibility of extracranial cognitive processes. In doing so, they effectively reduce Otto’s notebook from Clark and Chalmers’ idea of a functionally integrated process to just tool dependence. They are careful to make the distinction between epistemic function and cognitive function. While their essay doesn’t provide an example of extracranial cognition that would meet their criteria, as functionalists, Adams and Aizawa still concede that it’s possible. However, some may take issue with their rejection of the parity principle, saying that such a position is incompatible with functionalism. If this were accurate, then their appraisal of Otto’s notebook would be baseless. However, this objection is misleading, as it frames Adams and Aizawa’s marks of the cognitive as incongruous with functionalism rather than an addendum to functionalism.
By this, I mean that a functionalist stance isn’t in any way undercut by the two marks of the cognitive nor the rejection of the parity principle. To reiterate, the given definition of functionalism is that which defines minds through the roles of computation, cognition, and creation of mental states. The two marks of the cognitive put asterisks on cognition and creation of mental states, stipulating what those two functions themselves must entail. This is merely a narrowing of the scope of functionalism, a specification of its terms, an increase in its precision; it isn’t a fundamental infringement on its very ideology. Adams and Aizawa aren’t immediately disqualified from the functionalist camp because they mention non-derived content to particularize the creation of mental states, nor because they point to cognitive science to give parameters for cognition. In fact, these stricter criteria carve out an area of cognition that is more intuitive by excluding tool use and other obviously non-cognitive processes from consideration. For example, these qualifications would prevent a system like that in the Chinese Room Thought Experiment from being considered capable of cognition, whereas Clark and Chalmers’ criteria for a mind would lump that obviously non-thinking system into the cognitive. On the whole, Adams and Aizawa are not contradicting themselves in their rejection of the parity principle. Rather, they’re creating a more logical and specific understanding of functionalism that prevents non-cognitive objects like Otto’s notebook from entering the domain of the mind.
Although undoubtedly more compelling, one has to wonder how this idea of functionalism would apply to less intuitive examples of active externalism. For example, with the relative rise of transcranial technologies like brain implants, their marks of the cognitive may qualify a man-made tool as part of the mind. Neuralink, one of the most popular brain implant technologies, places a small medical-grade device underneath the skull, where it “decode[s] the information represented by [neural] cells” (Neuralink). Using that information, the device allows its users to navigate computer interfaces (often augmentative and alternative communication, or AAC, systems) at higher efficiency (Neuralink). So, the device is quite literally in the brain, and it also works directly with neural firings as they happen in order to complete its function. With this understanding alone, it seems as if the Neuralink is functionally integrated into the mind. It also certainly passes Adams and Aizawa’s first mark of the cognitive, as it’s unequivocally devoid of any derived content. The information Neuralink uses comes directly from the neural electrons, which are of course contained entirely within the brain and occur organically. In this way, the device seems to exclusively use non-derived content, so it conforms to the first mark of the cognitive.
However, applying the second mark of the cognitive is trickier. Although the Neuralink’s underlying causal process is arguably cognitive science since it operates secondary to neural firings, it is hard to say that the device and its function exist independently of fluid properties and is a true “natural kind.” Trying to qualify a brain implant as something that exists independently of culture and language is tricky. However, the fact that Neuralink works concomitant with organic neurological processes makes such objections difficult to accept. There is a “real causal process” underlying it, and one that is entirely intracranial at that (A&A 51). Despite Neuralink’s emergence from a cultural movement toward technological innovation for the body, its actual process is supercultural and superlingual. It passes the second mark of the cognitive because it’s coextensive with neural firings and therefore with cognitive processes. This is evidence to the fact that Adams and Aizawa’s criteria truly do conform to functionalism and allow for instances of real active externalism. Their marks of the cognitive can permit the introduction of things outside the brain (or at least manufactured outside the brain) to cognition, proving that Clark and Chalmers’ example of extracranial cognitive integration is merely insufficient. Clearly, Adams and Aizawa are committed to defending an idea of active externalism that is realistic, not overzealous with a low standard of admission.
Ultimately, Clark and Chalmers fail to provide an account of Otto’s notebook that would safeguard it against Adams and Aizawa’s objections. Because the latter give a stronger and less encompassing idea of what qualifies as cognition, they are able to successfully show that Otto’s notebook is merely a supplemental tool to his mind, not a functionally integrated piece of his mind. It isn’t the same as Inga’s hippocampus, nor is it evidence to the emergence of an extracranial cognitive process, although Adams and Aizawa are careful not to dismiss such a possibility. By the mark of the cognitive’s standards, Otto’s notebook is no different from a monkey with a stick; it’s merely a means to an end, and certainly not an integral part of Otto’s mind.
References
Adams, Frederick, and Kenneth Aizawa. The Bounds of Cognition. Blackwell Pub., 2008.
Clark, Andy, and David Chalmers. “The Extended Mind.” Analysis, vol. 58, no. 1, 1998, pp. 7–19. JSTOR, http://www.jstor.org/stable/3328150. Accessed 22 Nov. 2025.
Clarke, Sam. “Beyond the Icon: Core Cognition and the Bounds of Perception.” Mind & Language, vol. 37, no. 1, Feb. 2022, pp. 94–113. EBSCOhost, https://doi.org/10.1111/mil a.12315. Accessed 10 Mar. 2026.
Overstreet, Matthew. “Writing as Extended Mind: Recentering Cognition, Rethinking Tool Use.” Computers and Composition: An International Journal for Teachers of Writing, vol. 63, Mar. 2022. EBSCOhost, https://doi.org/10.1016/j.compcom.2022.102700. Accessed 10 Mar. 2026.
Neuralink. “Technology.” Neuralink, neuralink.com/technology/. Accessed 7 Dec. 2025.
Citation Style: MLA