Extended Cognition

Hampshire College

Philosophy of Mind

Introduction

In their paper “The extended mind”, Clark & Chalmers argue that the traditional view of cognition and the mind as being “inside the skull” is incorrect. Instead, they argue, cognitive processes—and, indeed, the mind itself—are partially constituted by the surrounding environment. Clark & Chalmers’ position is more than simple externalism—the position that a consideration of external, environmental factors is important in understanding the workings of the mind—it is radical, active externalism, arguing that the external, environmental factors are part of the mind. Clark & Chalmers focus in particular on the ways that belief can be extended. In this paper, I argue that belief is not something that can be, as of now[1], extended, and that memory can much more plausibly be considered to be extended, and I propose a criterion by which we could consider something part of the mind. The mind is extended, just not in the particular way Clark & Chalmers argue for.

Clark & Chalmers’ Argument

In their paper, Clark & Chalmers spend most of their focus on beliefs as an example of extended cognition. Specifically, they introduce us to Otto, a man with Alzheimer’s who has developed a system for keeping track of the various information that he would like to recall: he writes it all down in a notebook, a notebook that is reliably available to him whenever he needs it. When Otto hears that the Museum of Modern Art is having an exhibition he would like to see, and that the MoMA is on 53^rd St., he writes this information down in his trusty notebook. For all intents and purposes, Clark & Chalmers argue, the content of Otto’s notebook constitutes his standing beliefs about the world. Because beliefs are part of the mind, Otto’s mind extends outside his skull, and, similarly, our own mind can be considered extended, since after all we quite often store information outside ourselves and there is no important difference between our writing stuff down in a notebook and Otto doing the same (except, of course, that we do not have Alzheimer’s).

Beliefs, and Why Otto’s Notebook Doesn’t Contain Them

Clark & Chalmers make a distinction between standing and occurrent beliefs, a distinction that is useful to go over before we continue. Occurrent beliefs are the beliefs that one is conscious of at any given moment: for example, the belief that is now brought to my mind that this essay is due today. Standing beliefs are those beliefs we have that we may not be consciously aware of at this moment, but which influence our actions and may, at any time, become occurrent. It is not even necessary that we be conscious in order to have standing beliefs: my standing belief that Mars is red exists even when I am in non-REM sleep.

It is my (occurrent and standing) belief that the contents of Otto’s notebook do not satisfy the definition of standing beliefs and also lack some important properties of such beliefs that are not contained in the definition (though perhaps they should be). Imagine some unfortunate person Alicia, who has sustained brain trauma that has made it very difficult for her to access semantic information about cars. Perhaps she can’t access it at all without some form of rehabilitation[2]. Furthermore, she has no implicit memories regarding cars: she behaves exactly as one who knows nothing about them. However, she does not lack the information: with appropriate help she can access it without having to relearn it. It is simply impossible for her to access it without significant effort. Does she have any beliefs regarding cars before she is able to access the information? It is apparent that she does not.

What is it about Alicia’s case that makes it apparent that she lacks beliefs about cars? First, that standing beliefs not only influence our behavior, but that they can do so without us being aware of them. So, for example, I do not have to consciously recall my beliefs about gravity to become concerned when some fragile object falls, even though without such belief that situation would have no reason to concern me (if I have no beliefs regarding gravity, for all I know the object may become safely suspended in midair). Or, for another example, at any particular moment I may act as if there is or isn’t a God without consciously referring to my beliefs on the matter. (Indeed, it seems that quite often the only times we become aware of many of our standing beliefs is when they are challenged in some way: things aren’t where we expected them to be, someone argues against our basic assumptions, etc.) Alicia’s information about cars cannot do this. And the content of Otto’s notebook cannot do this either: it cannot influence Otto’s behavior without him becoming consciously aware of it. Secondly, and on a related note, standing beliefs can, and often do, become occurrent beliefs automatically and/or without our conscious control. For example, when someone mentions the Empire State Building the belief that it is in New York may occur to me, quite without my willing it to. Once again, Alicia’s information about cars cannot do this and, once again, the contents of Otto’s notebook cannot do this: someone mentioning the MoMA would not cause any occurrent beliefs on the matter for Otto, unless he made a willful decision to check.

There are likely many more important differences between the sorts of things that standing beliefs can do and the sort of things that the contents of Otto’s notebook can do. At least in regards to the two I’ve described, which seem particularly essential, the first by itself is justification for rejecting the description of the contents of Otto’s notebook as “beliefs”, as those contents cannot influence behavior directly.

Memories

So while beliefs can’t be plausibly extended, does Clark & Chalmers’ thesis still hold: can the mind be extended? Is there some aspect of our self that reaches out into the world? I believe that memory fits the bill. When Otto writes down information in his notebook, what is contained in the notebook is not beliefs, but memories.

The distinction between memories and beliefs is perhaps subtle, but it is clear that there is a difference. Memories can do many of the things beliefs do that I described above: they can influence our behavior without us being aware of them, they can be recalled automatically, etc. In the case of memories, however, these properties are not in any way essential. Going back to the case of Alicia, we can ask whether she has memories regarding cars. It is apparent that she does: the memories were there, she just could not access them. What would she be re-accessing otherwise? Memories need be nothing more than the storage of information[3]. (Of course, one could ask whether, without access to information regarding cars, that information, and hence those memories, could be said to belong to Alicia. I discuss this question—that of the ownership of memories—later on.)

Otto’s, and Our, Memories

Memories seem to be much more plausibly extended than beliefs[4]. Information stored outside the body seems to serve the same functional role as memory, for the same reasons Clark & Chalmers argued that such information serves the same function as belief. If we follow Otto around, we will see that where we would commit some information to memory, he writes it down, and where we would consult our memory, he consults his notebook. It is more difficult and effortful, perhaps, for him to access the information in his notebook than it normally is for us to access our memories while in a normal state; however, it is no more difficult for him than it would be for us to access our memories if we were tired, distracted, or had suffered the same sort of brain trauma as Alicia. In fact, even in a normal state the access of our own memories can be difficult: I often find, for instance, that the name of some person is inaccessible to me (despite the fact that I know it) and that I have to perform some sort of effortful memory search to re-access it. There is perhaps a qualitative difference in our memories: whereas my memories “feel” like information that I knew all along, the contents of the notebook would not. But this only reflects the fact that we are accessing the memories through different modalities.[5]

Who’s Memories?

In this discussion, the question may arise: if memories can be stored outside of the body, what makes those memories one person’s instead of another’s? Or, furthermore, what makes some information I’ve written down my memory, instead of just a record of information? I think this is the wrong question to ask. It assumes that my memories are encased in a unitary shell, that one person’s memories cannot be another’s, and that something is either my memory or it isn’t. I propose, instead, as an object of further study, that there is a degree to which a memory can be considered to belong to a particular individual. This degree is determined by a combination of properties, neither of which are necessary in and of themselves, although it is necessary that there be at least one or some combination of them: first, how easily is the information recalled by that particular person, and secondly, how important is the information to the identity of that person? By the first criteria, anything stored “in the head”, which when recalled has the qualitative sense of being “mine”, is indisputably my memory, regardless of how relevant it may be to my identity; and by the second criteria, a record of my life, sealed away so that I even cannot review it, would also constitute my memory. These are the extreme examples—most memories share both properties to some extent. In Otto’s case, the information in his notebook serves as his memory because of its reliability and ease of access, and also because without his notebook he would likely feel that he had lost an important piece of himself. [6]

But What of the Mind?

Perhaps, at this point, the reader will object that while memories may be extended, they are not aspects of the mind; therefore, the mind is not extended. Perhaps the mind is only those operations we perform on our memories; after all, there are many sources of information that the mind draws from in order to function that are not themselves part of the mind—sensory data, for instance, or, if you feel that the senses are part of the mind, the content of that data—and perhaps memory is such a source. To this, I would reply that in my formulation my “mind” is identical with my “self”, and it is our memories that make us who we are: if you remove some memory that is important to my conception of self, I would not longer be the same person, and, therefore, I am not of the same mind. Therefore, memories are part of the mind.

I think the difficulty lies in thinking of “the mind”: of the self as being one unitary thing, that if extended is extended in its entirety. Instead, I propose as a question for further study that the mind is made up of several different parts. Our conscious awareness is one such part, but our memories are also important. This non-unitary structure explains how outside sources of information can be memories and part of the mind, even though our recall of them is qualitatively different than that of memories stored “in the head”: the conscious aspect of our mind recalls them in different ways. Of course, now we need to ask how we determine what, exactly, can be considered part of the mind. There are many important factors that allow us to think—the beating of our heart, etc. How do we determine which of these should be considered part of the mind? I propose that the factors that are part of our minds are those that contribute functionally to the workings of the mind in such a way that they cannot be replaced without changing the identity of that mind—regardless of whether these factors are inside the head or not.

---

[1] I don’t consider it at all impossible that beliefs could be extended through the use of neural prosthetics in the future, as long as they satisfy the requirements listed later on in this paper.

[2] That, for the purposes of this thought experiment, does not include relearning the information.

[3] Although one can argue about whether and what certain types of stored information can be considered “memories”. I do not discuss this here.

[4] Furthermore, it seems much more natural to extend them this way: colloquially, we talk about a book of recollections, or a box full of memories, or even of places that remind us of the past as somehow containing the memories themselves (“this house is full of memories”), in a way that we do not for beliefs. While this fact about natural usage of terms does not show that it is in fact wrong to speak of beliefs as being extended (although as seen earlier it seems that it is), it does lend support to that assertion, and it make it even more likely that memories can be plausibly extended, as the idea does not seem too far from our natural intuitions about the use of our terms.

[5] One objection could be raised related to recent research on memory: it seems that when we recall memories, we re-imagine and rewrite them. Otto does not do this with the contents of his notebook (or, at least, it isn’t necessarily part of the process), therefore the contents are not memories. I would reply by saying that this fact about memory as it happens to work in no way describes an essential property: it could have turned out that the recollection of memories did not work this way, yet we would still call it memory.

[6] It is also possible for entities other than individuals to have memories: collective memory is allowed for, if the information is easily accessed by and/or important to the identity of the people as a collective.