Otto’s Notebook vs the Memex: Comparing the smartphone and the PC as two paradigms of cognitive enhancement


The history of computing has been characterized by an effort to replicate and augment human intelligence. Today’s personal computers are still substantially similar in their overall goals to the “memex,” a hypothetical computer prototype described by Vannevar Bush in 1945: both are designed to use computation to augment advanced forms of intellectual labor. Yet in the last decade, the rise of the smartphone has disrupted this long-standing paradigm of cognitive enhancement. It introduces a paradigm characterized by less complex but more pervasive kinds of cognitive enhancement. This is a kind of cognitive enhancement perhaps more akin to ever-present notebook that facilitated Otto the hypothetical Alzheimer’s patient in his daily activities (as described by Clark and Chalmers in their seminal paper on extended cognition in 1998). In this paper I compare and contrast some of the basic augmentations and affordances of the smartphone as compared to the PC (either desktop or laptop). Using Heersmink’s “multi-dimensional matrix for exploring cognitive-artifact relations” (Heersmink, 2012), I conducted a preliminary investigation suggesting some of the concrete differences between these two paradigms of computation and cognitive enhancement.

Download the PDF Here


The Nielsen US Digital Consumer Report of 2014 reported that the average amount of time US adults spent online on mobile surpassed the amount of time spent online on PCs for the first time this year (Nielsen, 2014). The easy-to-use always-on mobile computational device is a radical disjuncture from the laptop/desktop model of computation. The smartphone has a very different form-factor, use context, and computational and interface design, all leading this device to shape our thoughts and behavior in ways very different from a personal computer. In this paper I attempt to analyze and describe some of the differences between the two technologies in order to draw general conclusions about this paradigm shift. I do this primarily using Heersmink’s “multi-dimensional matrix for exploring cognitive-artifact relations” as a heuristic tool for analyzing these differences, specifically focusing on how they differ as cognitive tools. I hypothesize that the smartphone is a more practical kind of cognitive tool, providing assistance with everyday informational tasks and actions, whereas the PC is capable of significantly augmenting protracted and complex forms of cognition– in addition to the more basic cognitive functions supported by the smartphone. My research using Heersmink’s analytical tool draws out and describes those aspects of each artifact that creates their particular cognitive niche.

Historical and theoretical background

My analysis of these two paradigms of computing is based on an understanding of the smartphone and the personal computer as two different kinds of “cognitive artifacts.” Donald Norman defined cognitive artifacts as “those artificial devices that maintain, display, or operate upon information in order to serve a representational function and that affect human cognitive performance.” (Norman, 1991). Humans instinctively develop cognitive artifacts to enhance our mental performance, including everything from spoken language to the calendar, but the history of computing in particular is defined by the search to develop tools which can replicate or augment complex and high-level forms of human cognition. Perhaps the most classic example of this is the search to create a machine which can play chess as well as or better than the best human chess players.
This search to create a chess-playing computer is not purely incidental, but reflects the specific kinds of human thought and behavior that the earliest computational researchers were trying to replicate and augment. In 1962, Douglas Engelbart published an article titled “Augmenting Human Intellect: A conceptual framework.” This framework was specifically oriented towards developing artifacts and systems for enhancing human intelligence, and touched on the work of both Vannevar Bush and JCR Licklider, founding fathers in the history of computation. Engelbart would later go on to help develop the mouse, hypertext, and early precursors to the GUI, all incredibly important cognitive artifacts in the history of computing. But these great minds were primarily focused on developing computers that could enhance or augment a particular kind of cognition:

“By ‘augmenting human intellect’ we mean increasing the capability of a man to approach a complex problem situation, to gain comprehension to suit his particular needs, and to derive solutions to problems…And by “complex situations” we include the professional problems of diplomats, executives, social scientists, life scientists, physical scientists, attorneys, designers.”

Engelbart was imagining that these complex systems and computational artifacts could be used to help in a particular kind of intellectual labor. These machines were designed for augmenting advanced, formalized kinds of thought. Similarly, consider the following quote from Licklider’s visionary paper on “Man-computer Symbiosis”:

“Man-computer symbiosis is probably not the ultimate paradigm for complex technological systems. It seems entirely possible that, in due course, electronic or chemical “machines” will outdo the human brain in most of the functions we now consider exclusively within its province. Even now, Gelernter’s EBM-704 program for proving theorems in plane geometry proceeds at about the same pace as Brooklyn high school students, and makes similar errors [...] In short, it seems worthwhile to avoid argument with (other) enthusiasts for artificial intelligence by conceding dominance in the distant future of cerebration to machines alone.” (Licklider, 1960)

Licklider saw computers as tools for assisting with particular kinds of advanced, formal problem solving, and thus thought that man-computer symbiosis would only be a temporary phase, until machines inevitably were able to do this kind of work on their own. While it seems that in some ways our computers are moving towards this kind of independent processing (e.g., the complex networks of computers which independently trade stock based on mathematical formulas, or the automated processing of packet-switching that organizes all Internet traffic), it is increasingly clear that overall man-computer symbiosis is not a temporary state. This is because computation is no longer purely used for solving formalized complex problems. Computation has been incorporated into most everyday aspects of life, with the result that man-computer symbiosis is increasingly a fundamental part of the human experience. The smartphone as a cognitive artifact represents the movement of computation away from purely formal processes and towards integration into everyday forms of thinking and problem-solving.

This trajectory towards ubiquitous and everyday uses of computation seems obvious now, but is an unexpected twist that was not imagined by the early pioneers of computation. Even the personal computer, and the sometimes very simple forms of cognitive enhancement it offers, would probably have been a surprise to Licklider and company, let alone the application of advanced computational tools to mundane questions (e.g. “Where should I get lunch?”). This represents a dramatically different paradigm of computation altogether—one that has very little to do with formal problem solving and which certainly could never be done by machines alone.

It is interesting to consider just how invisible this turn towards more “personal” and everyday uses of computers was to the founding fathers of computation. In Vannevar Bush’s famous article “As We May Think,” published almost 15 years before “Man-Machine Symbiosis,” he actually comes quite close to an intuition of ubiquitous, mobile computing, and its integration with everyday activities. Consider, for example, the following quote:

“On a pair of ordinary glasses is a square of fine lines near the top of one lens, where it is out of the way of ordinary vision. When an object appears in that square, it is lined up for its picture. As the scientist of the future moves about the laboratory or the field, every time he looks at something worthy of the record, he trips the shutter and in it goes, without even an audible click.” (Bush, 1945).

This description of a pair of glasses with an embedded camera eerily foreshadows Google glass, that notorious artifact at the frontier of wearable technology. But note how Bush’s imagined user is still the scientist, the knowledge professional who needs to be able to perform incredibly advanced kinds of analysis. In a similar vein, although Bush acknowledged that there was a more general use and demand for specialized logic machines, he states: “With machines for advanced analysis no such situation existed; for there was and is no extensive market; the users of advanced methods of manipulating data are a very small part of the population.” Although this was true at the time, the introduction of the PC and accompanying cognitive artifacts has proved that “advanced methods of manipulating data” (which is precisely what is being done on the backend of nearly every app) can actually be quite useful to anyone with a place to be or a need to look up the president’s age. As of January 2014, 58% of American adults own a smartphone, and as such carry around tools for “advanced methods of manipulating data” in their pocket (Pew, 2014).

It is the subject of an entirely different project to ask what prevented these visionary men from seeing the potential applications of computation not just to intellectual laborers but to everyone in everyday situations. But whatever the reason, these early visionaries were very much focused on a paradigm of computation characterized by immobility/single-context use, directed towards advanced forms of formal thought and analysis, and utilized primarily by professionals and knowledge experts. This paradigm of computation has since become only one of many. The smartphone (and soon-to-come forms of wearable tech like the Apple watch or Google glass), in contrast, is fundamentally characterized by mobility and total integration with our everyday lives—and an emphasis on computation that helps us achieve practical, everyday goals more efficiently, rather than necessarily augmenting complex forms of higher thought or “intellectual labor.” This is a totally different paradigm of cognitive enhancement.

I would argue this new paradigm of computational cognitive enhancement has not been sufficiently analyzed or understood—nor indeed fully recognized. The smartphone is a computational technology that addresses an entirely different cognitive niche in our lives. Furthermore, this technology seems to bring us a step closer to a more pervasive kind of man-machine symbiosis and extended cognition, given the degree to which these technologies are physically, culturally, and cognitively embedded within every waking moment. This new paradigm in cognitive enhancement demands the same kind of advanced analysis and visionary thinking that undergirds the personal computing paradigm. An important first step towards this kind of visionary research around this kind of computation is to understand concretely how this new paradigm differs from the old.


In order to begin to explore in more concrete terms how the smartphone differs from the PC as a cognitive tool, I have drawn upon Richard Heersmink’s matrix for exploring cognitive-artifact relations. Heersmink’s matrix synthesizes suggested dimensions for analyzing cognitive artifacts proposed by several other researchers in the field. These dimensions and the resulting matrix are designed for “better understanding and exploring the different kinds of epistemic interactions and coupling between humans and cognitive artifacts.”(Heersmink, 2014) I put Heersmink’s dimensions into a matrix in Figure 1. I then applied each of these dimensions to a specific task on either a smartphone or PC, to create a complete matrix analyzing that task.

of analysis
(from heersmink, 2014)
type of information
flow: one way, two way, or reciprocal
“one-way information flow from artifact to agent. examples include clocks, compasses, slide rulers, road signs; two-way information flow, that is, from agent to artifact and then from artifact to agent…artifacts in two-way relations are often tailored for individual use and are frequently not part of publicly available artifacts or symbol systems. the third level is based on a reciprocal information flow. occasionally, cognitive artifacts are integral parts of ongoing information-processing tasks.”
reliability/durability/trustworthiness “the durability and repeatability of our relationship with cognitive artifacts often depends on the kind of epistemic action (and its epistemic purpose) one performs with it”
transparency “procedural and representational transparency…in clark’s words: “the notebook has become transparent equipment for otto, just as biological memory is for inga”…representational transparency concerns the effortlessness with which an agent can interpret and understand external information.”
individualization and
“individualization is changing, adjusting, or fine-tuning the artifact such that its use is more effective and efficient…entrenchment of cognitive artifacts implies a close equilibrium between agent and artifact in which both have been transformed in order to ensure the best possible fit between agent and artifact…otto’s is a case of entrenchment because ‘his behavioral and cognitive routines are sculpted by his notebook’”
speed of information
flow/ bandwidth
“information flow in human-artifact systems has a certain bandwidth, which is the amount of information that is exchanged per unit of time and depends on properties of both the agent and artifact…speed of information flow depends, on the one hand, on the cognitive and interpretation skills of the human agent and, on the other hand, on the informational and representational nature of the cognitive artifact…conversely, the speed with which one offloads information onto an artifact is also important. again, this depends on properties of both the agent and artifact. certain devices have input methods that are more efficient than others.”
distribution of
“the degree to which each element in a human-artifact system contributes to solving a problem depends on the distribution of computation.”
cognitive and
artifactual transformation
schemas are flexible as to incorporate 
tools into the sub-conscious representation of the body and its capabilities for
action. tool-use thus transforms the body schema. likewise, the use of cognitive artifacts and other external symbol systems transform the representational and cognitive capacities of the human brain…during ontogenetic development we interact with public 
systems such as mathematics and language. by so doing, we soak up and learn to think in those representational systems and the brain takes on the representational properties of those systems.”


I chose this matrix as a tool for analysis for two main reasons: 1) this matrix is a simplified modeling tool which helps to think more granularly about specific affordances, constraints, and general characteristics of each technology as it relates to cognition. It is in itself a cognitive tool. 2) Heersmink has very explicitly built into his matrix the idea that these dimensions of analysis can not be blindly applied to a piece of static technology, but that they arise out of particular combinations of “(1) the cognitive profile or cognitive capacities of the human agent; (2) the representational, functional, and technical properties of the cognitive artifact; (3) the task environment and context of use; and (4) the kind of epistemic action and its epistemic purpose.” (Heersmink, 2014) In other words, this analysis recognizes the integral importance of the specific user, the context of use, the goals of use, and last but not least the actual technological and representational affordances of the artifact.
With this context/user specificity in mind, it is important to note that my analysis using this matrix is only a preliminary experiment, and based on the single case study of myself[1], as a fairly native user, performing a few select tasks with a few select ends. A more comprehensive study would involve multiple users with different cognitive capacities, and a broader array of tasks done in different contexts and with different goals.

Although somewhat unconventional, the use of my own first-person experiences in this analysis actually falls in well with practices of introspection utilized historically in both phenomenological and some forms of psychological research. While a more extensive research project would include a wider variety of different users, particularly users with varying levels of familiarity with the device and the specific apps, my use of first-person experience in the context of a user experience study does not raise the same concerns about bias and self-reporting as other types of studies.

My analysis is done at the level of specific apps or tasks on both the smartphone and PC. Ideally this analysis would be even more granular; consider that within the task of “using

Facebook,” it is a very different thing to browse your News Feed or to create a large event, and these tasks would be performed very differently on a smartphone or a PC. Since I chose tasks that were familiar to me and because the dimensions of analysis are fairly abstract/conceptual, much of the analysis was done through simple reflection and mental comparison, although in some instances I found myself returning to the actual task to consider some questions more specifically (e.g. “How much information am I absorbing when I scan the News Feed via my phone versus on my computer?”). We can guess that more information is available via the larger screened device, but in some instances returning to the actual task revealed that, for instance, the information is organized in such a way on the mobile application that the difference in informational bandwidth is smaller than I might have guessed. A more rigorous and lengthy research procedure might ask users to always enact these procedures immediately before moving on to reflection, or even to return to them again after being told to pay attention to specific dimensions of analysis (e.g. check your Facebook News Feed on both your smartphone and your PC for 3 minutes, and pay particular attention to how much information you find yourself absorbing, and how the interfaces encourage specific flows of information and discourage others).

After doing this more granular task and application-based analysis, I then extrapolated back out to the level of comparing the two technologies overall based on Heersmink’s dimensions of analysis. The strongest conclusions resulted when similar results were reached across multiple tasks or applications, whereas widely differing results on the same dimension of analysis suggests that the differences in cognitive impact might not be best attributed to the devices, but to the particularities of a given task or application design.

Heersmink’s matrix is a complex and rigorous tool which could be used for a much more in-depth analysis of these two technologies across a number of different variables. What follows is merely a “quick and dirty” test run of this tool in order to understand some basic differences between the two technologies. I am certain that this is an over-generalized analysis that glosses over much of the complexity of these tools and their use in real-life contexts, but it did allow me come to a few concrete, high-level conclusions about how these two technologies operative differently as cognitive tools.


I have broken out my analysis along some of the matrices proposed by Heersmink, in some cases creating groupings for similar topics or themes.

  1. Reliability and Cognitive transformation/entrenchment

Due to it’s pocket and hand-sized form factor, the smartphone is much more reliable and accessible than the PC. The smartphone can be more closely integrated with your body, ready-to-hand in a Heideggerian sense. The computer, on the other hand, is much more restricted–you are less likely to take it with you to a party or keep it next to you in bed. Reliability is a fundamental difference between these two cognitive tools, because the smartphone’s ability to integrate more seamlessly with our bodily schema allows the smartphone to become much more closely integrated with our cognitive schema as well. For example, the reliability of always having Google maps on my person alters how I think about and approach moving around in space. In many ways it opens up an entire sphere of activity to have this device so reliably coupled with my own body.

This app is, of course, only one of many powerful tools that smartphone users carry around in their pocket. The smartphone allows us to walk around with the certainty that, at any moment, we can access nearly any piece of information, communicate with nearly anyone, be entertained and consume media any number of ways, and to document and share our thoughts, surroundings, or a given moment of experience. The physical reliability and accessibility of the smartphone, combined with the breadth of expertise afforded by this toolbox, is a strong impetus for building cognitive reliance upon the phone in our daily lives. We take this tool for granted because it is so tightly embedded in our bodily and cognitive schema, and when it is lost or forgotten we experience a sense of anxiety over what is essentially a part of ourselves[1].

This reliance on the smartphone is partially due to the cycle of increasing returns we receive the more closely we are coupled with the tool. To return to the example of Google maps, if I do not always have my phone—if it is not reliable—then I must have duplicate system—either a map, or to make the effort to mentally store and remember transit information—to cover those times when I do not have it. If I am confident that I always have my phone on me, then I am relieved of the cognitive burden of these “backup” systems. In this way, the reliability of the phone allows for a more complete change in my behavior and cognition. Similarly, if I am trained to check my phone every few minutes, then it is increasingly unlikely that I will miss a time-sensitive notification about, say, cancelled dinner plans as I am getting ready to leave the house or a meeting I am about to miss. This inherent reward system of checking the phone contributes to changed behavioral patterns (e.g. increasingly frequent phone-checking). (Oulasvirta, 2012) In their seminal paper on extended cognition (Clark, 1998), Chalmers and Clark use the thought experiment of a man suffering from Alzheimers named “Otto” who has externalized his memory into a notebook that he always keeps on his person. Heersmink notes that in the case of Otto and his notebook, “[Otto’s] behavioral and cognitive routines are sculpted by his notebook.”(Heersmink, 2014) Heersmink calls this kind of behavioral and cognitive change “entrenchment” or “cognitive transformation.” In this instance, it is very clear that the smartphone creates extensive behavioral and cognitive transformation, in large part due to the reliability and physical accessibility of this cognitive tool[2].

In contrast to this very pervasive form of entrenchment between the smartphone and its user, the PC is obviously more physically distant and, as a result, is generally less embedded within everyday cognition and behavior than the smartphone. The form-factor prevents you from having it physically on your person at all times, and thus the PC does not allow you to develop the same reliance and completeness of cognitive offloading as the smartphone. In terms of cognitive transformation, the PC does have a much bigger transformative effect within the context of specific tasks. For example, over time my actual cognitive processes when writing a paper have been transformed by the affordances of a word processor, such as copy/paste and quick search. Similarly, I find that when doing research online my thoughts tend to follow a kind of branching pattern similar to a chain of hyperlinks. These are complex and very significant changes in our cognition, but these cognitive transformations are limited to specific kinds of intellectual work. These cognitive transformations are more task-specific, whereas with the smartphone the entire device becomes part of my cognitive-bodily schema, transforming my behavior and cognition in a more holistic way.

The PC effects complex changes to particular cognitive processes, but these changes are isolated to those tasks alone. The smartphone changes my behavior and cognition in a way that is perhaps more simple (e.g. creating a compulsive phone checking habit) but is much more pervasive across a variety of context and tasks.

  1. Bandwidth

Another major and obvious difference between these two cognitive tools relates to what Heersmink calls “bandwidth.” This refers to the speed of information flow between the artifact and the user. This flow of information is not just from tool to user, but also from user to tool.

In the case of the smartphone, it is well known that this tool is extremely limited in terms of user input via the tiny keyboard. Lengthy writing tasks remain the exclusive domain of the keyboard and mouse, not the touchscreen[3]. In addition to the constraints of a user’s ability to input information, the smartphone is limited in its ability to return information to the user because the relatively small screen can only display so much information at once. Applications like Twitter exemplify the attempt to develop information streams that are more consumable via the mobile phone’s smaller informational bandwidth, but richer informational sources (e.g. spreadsheets, lengthy articles, even long-form videos) can be difficult to effectively consume on the mobile phone. The smartphone also tends to be less reliable in terms of its network connection speed, which is often subpar compared to the stable wifi connection found on a PC.

  1. Limited Reciprocal information flow

            One related aspect of information flow on a typical mobile device is that this flow is organized into discrete streams of information, each corresponding to a particular app. I discovered this when initially trying to analyze the smartphone and the PC based on a particular task. I came to realize that on a phone, it was easy to choose a particular application and analyze activity within that application, but on a PC you are much more likely to switch applications quickly and work in parallel across multiple different information streams and tasks. For example, when thinking about mobile email, it is easier to imagine the general use-case, which tends to be reading email, deleting email, and simple replies. On the PC, “doing email” often involves cross-referencing other pieces of information in totally separate applications (calendar, documents, etc), incorporating that information into responses, and generally is much more integrated with different tasks, applications, and informational flows. I would argue that the bandwidth limitations of the smartphone described above, combined with the difficulty of switching between different streams of information, is what makes it nearly impossible to do the kind of integrative, cross-referential activity that characterizes focused intellectual labor.

Heersmink also defines three basic types of information flow: one-way, two-way, and reciprocal. In keeping with the limited informational bandwidth of the smartphone interface and the restrictiveness of app-based computing, information flow on the smartphone is often two-way, consisting basically of human input and tool response. In a reciprocal information flow, cognitive artifacts “are integral parts of ongoing information-processing tasks.” I was hard pressed to think of a situation in which the smartphone is truly a part of an “ongoing information-processing task.” Mobile phone use is characterized more by a kind of “checking” than by complex and iterative forms of thought. However, this kind of information flow is precisely what characterizes use of the PC. Browsing, researching, general work activities all utilize the PC as an integral part of the information processing taking place. As described by Heersmink:

“We offload small bits of information onto the artifact, and the nature and content of the offloaded information contributes to and partly determines the next step in the overall process. For example, when writing an academic paper one often starts with a rough outline, which may prompt ideas about how to fill in the details. Filling in the details may then prompt an adjustment of the outline, which may in turn prompt further details…Each step in the overall process builds and depends on previous steps. The human agent and cognitive artifact continuously exchange information and so there is a reciprocal and cumulative information flow that constantly transforms the agent-artifact system. There is, in Clark’s words, ‘continuous reciprocal causation’ between agent and artifact.”

An important element of this kind of reciprocal information flow is that it seems to represent rapid cumulative knowledge development between man and machine. This is truly thinking with the machine, rather than simply informing our independent cognitive process with information provided by a machine. I would argue that the lack of complex reciprocal information flow between the user and a smartphone indicates a certain limit on the cognitive enhancement that a smartphone can offer. In contrast, the PC is a powerful cognitive tool that is much more integral to complex and extended forms of thought. Of course, this is largely a formal way of explaining what most of us know intuitively: that the smartphone is more convenient for simple tasks and for basic entertainment, whereas serious “work” is better accomplished on a larger computer.

  1. Distribution of cognition

Heersmink also includes “distribution of cognition” as one dimension of analysis. Our above analysis would suggest that in many cases the PC is capable of taking on a larger share of the computation in a given problem. It can significantly facilitate even more complex forms of thought and analysis. In contrast, the application-based computing that characterizes the smartphone limits computation to relatively simple and singular tasks, which I then use as a means to inform a much more complex overall task. For example, when I look up the weather (relatively simple computation by the computer), I must still determine the proper attire for the day given the weather and other complex factors. Or I check my Twitter feed, but still must parse through the content and figure out what I want to read, what I want to share, etc. Of course, the same would be true if I wanted to complete these tasks on my PC—but the PC also allows me to do complex reciprocal processes such as writing a paper, whereas the smartphone is generally limited to these simple, two-way information flows. Overall, it seems that the PC is capable of somewhat more evenly distributed cognition between cognitive artifact and the user than the smartphone.

  1. Consumptive Activity across Smartphone and PC

In reality, there is a substantial amount of activity done on both the phone and the PC that is very similar. In my analysis of using Facebook on both the PC and the smartphone, it seems that these two activities are largely similar across all the different dimensions of analysis. Indeed, this seems to be true of most passive information or media-consuming tasks. In these instances, there is very little informational input from the user, and very little task switching or parallel information processing that needs to be done. In this case, the usual limitations of the smartphone in terms of information flow are largely irrelevant. The implication of this is that the mobile device, while perhaps an ineffective tool for complex forms of computation, is equally effective for simpler forms of media consumption.

Interestingly, this corresponds strongly with patterns of user behavior which suggest that the mobile phone user often turns to the mobile phone for media consumption even when at home and a PC is also available (Nylander, 2009). The authors cite user explanations that this is partially because the phone is always accessible. Thus, in many cases the immediacy of access to the smartphone trumps the possible increased informational bandwidth of the PC. I would also suggest that many of our activities are characterized by either simple two-way information flow interactions (e.g. looking up the weather), or by some form of relatively passive information consumption, such as skimming through the Facebook newsfeed or quickly reviewing incoming email.

In a kind of parallel to Bush’s suggestion that “the users of advanced methods of manipulating data are a very small part of the population,” (Bush, 1945) it seems we only require tools for enhanced cognition and complex, iterative forms of thought during select time periods, largely when doing some kind of work. During all other parts of the day, much simpler enhancements and forms of entertainment suffice.


To summarize my above analysis, it seems that there are some clear and essential differences between these two cognitive artifacts:

  • The smartphone is much more reliable, and thus much more pervasively entrenched, both behaviorally and cognitively. However, it has limited informational bandwidth. It is much better suited for simple two-way information flows and passive information consumption than complex forms of thought and idea-development.
  • The PC is much less accessible and pervasive, and thus shapes our cognition only with regards to specific tasks and activities. The bigger informational bandwidth in terms of both input and output makes it a much more effective tool for extended and complex forms of thought. Use of the personal computer creates a cumulative effect whereby thought is substantially enhanced and extended beyond what is possible within the user’s own mind.


This analysis helps us to see precisely how the smartphone breaks with the previous paradigm of computation and cognitive enhancement. If the memex represents the basic prototype of a computational cognitive artifact, then the PC actually seems to be the most up-to-date iteration of this same prototype. Like the memex, it helps to augment human thought via a complex, reciprocal exchange of information, specifically by augmenting the kind of cross-referential and integrative thinking that characterizes advanced intellectual labor. This interaction creates a kind of cumulative effect wherein our thought can be advanced significantly beyond the capability of the human mind alone.

In contrast, the smartphone acts more as a simple informational aid. It augments that part of our thought and activity that is not engaged in complex, extended cognition. And for many people and much of the time, the wide variety of simple enhancements offered by the smartphone, coupled with its physical accessibility, make for both a necessary and sufficient cognitive tool. It is not meant for advanced forms of intellectual labor, instead simplifying and slightly augmenting basic referential and communicative activities.

In this way, the smartphone represents a break with the kind of computation imagined by Bush, Licklider and Engelbart. These founding fathers were designing computational tools for the academic, the scientist, for themselves in many ways. The smartphone is a kind of “everyman’s” computational device, much better fitted for the kinds of simple informational tasks that facilitate everyday life and occupy our free time, outside of the office or school. Steve Jobs is perhaps an early visionary of this kind of computation, but I would suggest that we need much more research and thought dedicated to developing this form of computation that is so integrated with our daily lives. Thad Starner’s article on Google Glass is one promising example of advanced consideration about how this technology can best fit its cognitive niche, but it only represents a small slice of the field of mobile computation more generally (Starner, 2013).


This concludes my quick-and-dirty attempt to analyze these two cognitive technologies. There are many more ways that these two cognitive artifacts could be analyzed, ideally with greater rigor and specificity than I have been able to here. Indeed, it would be very interesting to do a similar analysis using the methods laid out by Xia and Maes in their reformulation of Engelbart’s conceptual framework for augmenting human intellect. However, we must proceed very carefully. Even their framework falls short in that it is explicitly meant to enhance our intelligence, and if this paper has shown anything I think it is that the smartphone is not a device intended to augment our intellect, strictly speaking. This framework, and likely most frameworks for analyzing computational tools, is still very much entrenched in the PC/memex paradigm of augmenting only certain forms of advanced cognition.

It would also be interesting to attempt a similar analysis taking into account the current horizon of computational technologies. The tablet and new forms of wearable tech already are and will continue to disrupt the current paradigm of computation I describe here, changing both how we design software and how we implement these different tools into different cognitive niches in our lives. In order to understand this increasingly varied toolset of cognitive artifacts, we must develop a more advanced theory of computation and cognitive artifacts directed towards different kinds of cognition and behavior, rather than focusing on advanced forms of intellectual labor.


[1] Although somewhat unconventional, the use of my own first-person experiences in this analysis actually falls in well with practices of introspection utilized historically in both phenomenological and some forms

[2] See, for example, “Nomophobia,” a proposed phobia characterized by “discomfort, anxiety, nervousness or anguish caused by being out of contact with a mobile phone or computer” (Bragazzi, 2014). Although a “phobia” is defined as an irrational fear, I would argue that in some ways the anxiety felt when separated from the mobile phone is perfectly rational, akin to the fear we might experience if tried to move in the world with a broken leg or in a new city without a map.

Bragazzi, N. L., & Del Puente, G. (2014). A proposal for including nomophobia in the new DSM-V. Psychology Research and Behavior Management, 7, 155–160. doi:10.2147/PRBM.S41386

[3] In this light, it is easy to see the value of wearable tech such as the Apple watch or Google glass. Users/wearers of these technologies can attain almost complete reliability through literal physical coupling and reap the benefits associated with frequent checking while minimizing the annoyances of compulsively checking an “external” device.

[4] Speech-to-text has made some slight improvements here, but it is still problematic in many contexts and for specific uses.


Works Cited

Absar, R., Halbert, H., O’Brien, H., & Trumble, A. (2014, February). Searching for Information On-the-Go: Understanding User Engagement with Mobile Devices. In Proceedings of the Annual Conference of CAIS/Actes du congrès annuel de l’ACSI.

Bush, V. (1945). As we may think. The Atlantic Monthly, 176(1), 101-108.

Bragazzi, N. L., & Del Puente, G. (2014). A proposal for including nomophobia in the new DSM-V. Psychology Research and Behavior Management, 7, 155–160. doi:10.2147/PRBM.S41386

Clark, A., & Chalmers, D. (1998). The extended mind. Analysis, 7-19.

Engelbart, D. C. (1963). Conceptual Framework for the Augmentation of Man’s Intellect.

Heersmink, R. (2012). Mind and artifact: A multidimensional matrix for exploring cognition-artifact relations.

Licklider, J. C. R. (1960). Man-computer symbiosis. Human Factors in Electronics, IRE Transactions on, (1), 4-11.

Nielsen. (2014) The US Digital Consumer Report. Retrieved from:

Norman, D. A. (1990). Cognitive artifacts. Department of Cognitive Science, University of California, San Diego.

Nylander, S., Lundquist, T., Brännström, A., & Karlson, B. (2009). “It’s Just Easier with the Phone”–A Diary Study of Internet Access from Cell Phones. InPervasive Computing (pp. 354-371). Springer Berlin Heidelberg.

Oulasvirta, A., Rattenbury, T., Ma, L., & Raita, E. (2012). Habits make smartphone use more pervasive. Personal and Ubiquitous Computing, 16(1), 105-114.

Pew Research Internet Project. (2014). Mobile Technology Fact Sheet. Retrieved from:

Starner, T. (2013). Project glass: An extension of the self. Pervasive Computing, IEEE, 12(2), 14-16.

Xia, C., & Maes, P. (2013, March). The design of artifacts for augmenting intellect. In Proceedings of the 4th Augmented Human International Conference(pp. 154-161). ACM.