Distributed Selves: Personal Identity and Extended Memory Systems

Abstract

This paper explores the implications of extended and distributed cognition theory for our notions of personal identity. On an extended and distributed approach to cognition, external information is under certain conditions constitutive of memory. On a narrative approach to personal identity, autobiographical memory is constitutive of our diachronic self. In this paper, I bring these two approaches together and argue that external information can be constitutive of one’s autobiographical memory and thus also of one’s diachronic self. To develop this claim, I draw on recent empirical work in human-computer interaction, looking at lifelogging technologies in both healthcare and everyday contexts. I argue that personal identity can neither be reduced to psychological structures instantiated by the brain nor by biological structures instantiated by the organism, but should be seen as an environmentally-distributed and relational construct. In other words, the complex web of cognitive relations we develop and maintain with other people and technological artifacts partly determines our self. This view has conceptual, methodological, and normative implications: we should broaden our concepts of the self as to include social and artifactual structures, focus on external memory systems in the (empirical) study of personal identity, and not interfere with people’s distributed minds and selves.

Full text

Synthese
DOI 10.1007/s11229-016-1102-4
Distributed selves: personal identity and extended
memory systems
Richard Heersmink1
Received: 12 February 2016 / Accepted: 15 April 2016
© Springer Science+Business Media Dordrecht 2016
Abstract This paper explores the implications of extended and distributed cognition
theory for our notions of personal identity. On an extended and distributed approach to
cognition, external information is under certain conditions constitutive of memory. On
a narrative approach to personal identity, autobiographical memory is constitutive of
our diachronic self. In this paper, I bring these two approaches together and argue that
external information can be constitutive of one’s autobiographical memory and thus
also of one’s diachronic self. To develop this claim, I draw on recent empirical work
in human-computer interaction, looking at lifelogging technologies in both healthcare
and everyday contexts. I argue that personal identity can neither be reduced to psy-
chological structures instantiated by the brain nor by biological structures instantiated
by the organism, but should be seen as an environmentally-distributed and relational
construct. In other words, the complex web of cognitive relations we develop and
maintain with other people and technological artifacts partly determines our self. This
view has conceptual, methodological, and normative implications: we should broaden
our concepts of the self as to include social and artifactual structures, focus on external
memory systems in the (empirical) study of personal identity, and not interfere with
people’s distributed minds and selves.
Keywords Personal identity ·Narrative self ·Extended mind ·Distributed
cognition ·Transactive memory ·Lifelogging
BRichard Heersmink
richard.heersmink@gmail.com
1Macquarie University, Sydney, Australia
123

Synthese
1 Introduction
As an adaptive strategy, the human cognitive system has evolved to offload and incor-
porate information in the environment to complement the limitations in its capability
to store and process information (Donald 1991). Thus rather than seeing our cog-
nitive system as instantiated only by the brain, we should see it as spread out across
embodied brains and information-structures in the environment (Hutchins 1995;Clark
1997;Rowlands 1999;Menary 2007;Sutton 2010). Memory systems, in particular,
are typically instantiated in distributed networks, as the environment affords easy
and often reliable storage of information (Michaelian and Sutton 2013). On such an
extended and distributed cognition view, our cognitive capacities depend on and are
(sometimes) constituted by a complex web of social and artifactual structures. Syn-
chronically, human cognizers thus form coupled systems with certain aspects of the
environment in which they are situated (Clark and Chalmers 1998).
Neo-Lockean views on personal identity emphasize memory as an important cri-
terion of persistence of personhood over time. Marya Schechtman (1994,1996), for
example, argues that we summarize and condense our past experiences into an auto-
biographical narrative, which is constitutive of our diachronic self. So who we are
as persons is defined and constituted by our autobiographical narrative. Given this
emphasis on memory, there is an obvious connection between extended and distributed
approaches to cognition and neo-Lockean approaches to personal identity1(Wilson
and Lenart 2014). In this paper, I further conceptualize this connection by developing
the following argument: if memory is often distributed and if the self is partly con-
stituted by our memory, then the self is also distributed. To develop this argument, I
draw on recent empirical work in human-computer interaction, looking at lifelogging
technologies in both healthcare and everyday contexts (Berry et al. 2007;Doherty
2012;Crete-Nishihata 2012). These are memory technologies that record personal
experiences. A key example is SenseCam, a small wide-angle camera worn around
one’s neck, taking a picture with a certain interval or when its sensor detects some
environmental change. These pictures are then edited into a visual lifelog with a cer-
tain narrative structure, transforming, aiding, and in some cases constituting one’s
autobiographical narrative.
The analysis unfolds as follows. First, I outline situated cognition theory with
a focus on extended and distributed memory systems. Second, I present the current
debate on whether the self is constituted by an embodied organism or partly constituted
by external informational structures. Third, I develop a view on the self as essentially
distributed across embodied agents and external information located either in other
agents (social) or in objects (artifactual). Fourth, I reflect on some of the normative
aspects of lifelogging and our distributed selves.
2 Situated and distributed memory
In the last thirty years or so, new views on human cognition emerged, arguing that our
cognitive systems are often scaffolded and sometimes constituted by external resources
1The terms “self”, “personhood”, and “personal identity” are used interchangeably throughout this essay.
123

Synthese
such as other people and artifacts. Situated cognition theory is typically used as an
umbrella term to cover these views, including embodied, embedded, extended, distrib-
uted and collective cognition.2It is a set of loosely related approaches that underline
the importance of our embodied interactions with the socio-technological environment
for understanding our cognitive capacities, providing an alternative for neuro-centric
and individualist views on cognition. Situated views point out that our embodied
brains are powerful pattern recognition and completion systems, but have limited
information-storage and processing capacities. There are, for instance, limitations in
information-processing speed, accuracy, and capacity. To overcome these limitations,
we rely on and incorporate the complementary properties and functions of external
artifacts and other people (Sutton 2010). As Daniel Dennett writes: “The widespread
practice of offloading releases us from the limitations of our animal brains” (1996,p.
135). Below I briefly outline how we offload, rely on and incorporate information in
other people and artifacts, focussing on memory.
2.1 Transactive memory
Daniel Wegner and colleagues (Wegner et al. 1985;Wegner 1995; see also Theiner
2013) advanced an empirical view on social memory, arguing that dyads and other
intimate social groups often develop transactive memory systems. In such distrib-
uted memory systems, there is a cognitive interdependence between group members,
which means that agents rely on each other for their cognitive performance. Transac-
tive processes are essentially communicative or interactive processes between group
members, which most obviously occur at the encoding and retrieving stage of memory
processing. During transactive encoding, group members discuss incoming infor-
mation and negotiate who has to store the information and in which form. During
transactive retrieval, group members have to determine where the information is stored,
which can involve the integration of information stored in different brains. Consider
the following example of transactive retrieval from Harris et al. (2010):
F: And we went to two shows, can you remember what they were called?
M: We did. One was a musical, or were they both? I don’t ... no ... one ...
F: John Hanson was in it.
M: Desert Song.
F: Desert Song, that’s it, I couldn’t remember what it was called, but yes, I knew
John Hanson was in it.
M: Yes
In this example of transactive remembering, a long-married couple give each other
cues such that they jointly recall information that both individuals had forgotten.
Wegner et al. (1985) emphasize the interdependency of individual memory systems and
therefore point out that transactive memory systems cannot be reduced to individual
memory. It is an emergent group-level property that exists only when people interact
and communicate in the right sort of way (Theiner 2013). So, in order to empirically
2Sometimes the phrase “4e cognition” is used, which is an acronym for embodied, embedded, extended,
and enactive cognition.
123

Synthese
study and theoretically conceptualize such systems, we need to enlarge the unit of
analysis from a single agent to various agents interacting with each other. If we would
only focus on single agents in the Desert Sing example, we would not be able to
fully explain the outcome of the system. This real-world example thus suggests that
transactive memory systems are group-level properties and cannot be reduced to its
individual members.
2.2 Artifactual memory
In addition to relying on and collaborating with other people to perform memory tasks,
we also quote often use information stored in artifacts. An often mentioned example of
artifactual memory in the extended cognition literature is Otto and his notebook. Otto
is a man in the first stages of Alzheimer’s disease, using a paper notebook to aid and
supplement his biological memory systems (Clark and Chalmers 1998). He always
carries it with him because he needs the information in the notebook to successfully
get around in the world. Importantly, information in Otto’s notebook plays relevantly
similar functional roles as information stored in his biological memory systems in
guiding Otto’s actions. The information is further easily accessible, reliably there
when needed, trustworthy, has been endorsed at some point in the past and indeed
is there because of this endorsement. On an extended and distributed cognition view,
“a cognitive process is delimited by the functional relationships among the elements
that participate in it, rather than by the spatial colocation of the elements” (Hollan
et al. 2000, p. 176).3So the function and the way it is used and integrated into Otto’s
cognitive system are what matters, not that it is external to the organism.
Our memory is scaffolded by varies other external artifacts. For example, we put
labels on objects so we do not have to remember where they are; we use diaries and
calendars to remember our appointments; shopping lists to remember the items we
need to buy; and books, libraries, databases, and the Internet to store and look up all
kinds of information. These cognitive artifacts have been intentionally designed to
serve memory supporting functions, but we also often offload memory functions onto
the environment by improvising. Bartenders, for example, are typically under a lot of
pressure and have to deal with certain cognitive and environmental constraints. In order
to deal with the cognitive overload, they structure distinctively shaped drink glasses
such that they correspond to the sequence of the ordered drinks. Consequently, the
bartenders do not have to internally remember the order of the drinks, but offload it onto
their environment (Beach 1988). Cooks, likewise, offload memory tasks by actively
structuring their work-environment. The ingredients and equipment are arranged such
that they facilitate the cooking process by offloading the order of the steps of the
cooking process onto the environment. Kirsh (1995) refers to this as “the intelligent
use of space” in which the arrangement of artifacts is itself a cognitive artifact. Material
culture thus plays an important role in human memory (Donald 1991).
3I am using the phrases “extended cognition” and “distributed cognition” more or less interchangeably
in this paper, as both approaches claim that informational artifacts can be constitutive of a wider cognitive
system, but there are differences between the two approaches. See Hutchins (2014) for an analysis of some
of these differences in terms of examples, method, and explanatory scope.
123

Synthese
2.3 Situated autobiographical memory
As the Desert Song example shows, transactive memory often has a strong autobio-
graphical or episodic component, because transactive memory partners typically have
many shared past experiences. Most examples of artifactual memory in the situated
cognition literature, however, concern information that scaffolds and is integrated
with short-term or semantic memory systems but not so much with autobiographical
memory. It is, however, important to point out that artifacts also quite often scaffold
autobiographical memory. Some people, for example, have a journal in which they
write their daily experiences, which is a medium used to express oneself and organise
one’s thoughts, but also used as an external autobiographical memory system. Other
examples are photos and photo albums; travel blogs; significant objects like sou-
venirs, jewellery, or books; status updates on social media like Facebook and Twitter;
and uploaded pictures on social media like Flickr, Facebook or Instagram. Interacting
with these artifacts and the information they contain often evoke autobiographical and
sometimes emotional responses (Hoven 2014).
Relatively new and very powerful autobiographical memory technologies are now
emerging referred to as “lifelogging technologies” (O’Hara et al. 2008;Hoven et al.
2012). These are digital artifacts and applications that record aspects of one’s every-
day life. Such technologies include wearable cameras automatically taking pictures
with a certain interval (e.g., SenseCams), applications recording screen behaviour,
and wearable sensors recording biological variables (e.g., heart rate, temperature,
blood oxygen level, and EEG) or environmental variables (e.g., location or temper-
ature). Vannevar Bush’s concept of the “memex” (a portmanteau of memory and
index) was an early precursor of lifelogging. A memex is a “device in which an
individual stores all his books, records, and communications, and which is mech-
anised so that it may be consulted with exceeding speed and flexibility. It is an
enlarged intimate supplement to memory” (1945, p. 6). Bush suggested that scien-
tists could wear glasses that can take pictures of what they see and that entries in
the memex were annotated and easily searchable. Computer pioneer Gordon Bell
developed one of the most ambitious (and controversial) examples of contemporary
lifelogging called the “MyLifeBits” project. Inspired by Bush’s vision of the memex,
Bell digitized articles, books, music, phone calls, letters, photos, presentations, and
home movies, and stored them in a lifelog (Bell and Gemmell 2009). Software allows
him to search for dates and keywords in the lifelog, which, for Bell, functions as an
elaborate external autobiographical memory system. A practical problem with some
of these lifelogging technologies is that they capture too much information, leading
to information overload. So the challenge is to construct lifelogs that are manage-
able and cognitively meaningful to the user (Doherty 2012). More on this below in
Sects. 4and 5.
There is a small subculture of people who actively and consistently use these tech-
nologies to record their everyday life sometimes called the “quantified self movement”,
making elaborate lifelogs about their daily life. Most people, however, are not inter-
ested in creating elaborate lifelogs, but do sometimes use lifelogging applications,
for example when they make a (digital) photo album of their past holiday, track their
route with their smartphone when they go jogging, or do not delete emails such that
123

Synthese
they have a database of previous email communication. Importantly, some of these
technologies are used in healthcare to help patients with memory disorders. Sense-
Cams, in particular, are useful and effective to record aspects of a patient’s daily life
(Hodges et al. 2011). It has been shown that reviewing a visual lifelog has benefi-
cial effects on a patient’s autobiographical memory such as maintaining the integrity,
delay the disintegration, or in some cases replace autobiographical memory (Berry
et al. 2007;Doherty 2012;Crete-Nishihata 2012). Elsewhere I argued that lifelogging
technologies change our embodied self, as the information they provide about bodily
and environmental states transforms the kind of relationship we have to our embod-
iment (Smart et al. 2016). In this paper, I elaborate on this claim and argue that it is
not just our embodied self but also our narrative self that is transformed by lifelogging
technologies.
3 Extended cognitive systems and personal identity
Having briefly outlined situated cognition theory and the distributedness of both
semantic and episodic memory, I now continue by summarizing the current debate
on the extended self. Some theorists argue that the notions of extended and distrib-
uted cognition imply an extended and distributed self (Clark and Chalmers 1998;
Clark 2005;Anderson 2008), whereas others deny this (Baker 2009;Olsen 2011;
Buller 2013). In his early work, Clark (1997) speculates about the possibility of an
extended self and agency, but ultimately remains agnostic whether self and agency are
extended. Clark writes: “I am content to let the notions of self and agency fall where
they will” (1997, p. 218). However, in his co-authored article with David Chalmers,
The Extended Mind, Clark argues that the extended mind does suggest an extended
self.
“Does the extended mind imply an extended self? It seems so […] The infor-
mation in Otto’s notebook, for example, is a central part of his identity as a
cognitive agent […] To consistently resist this conclusion, we would have to
shrink the self into a mere bundle of occurrent states, severely threatening its
deep psychological continuity. Far better to take the broader view, and see agents
themselves as spread into the world” (Clark and Chalmers 1998, p. 18).
So, on Clark and Chalmers’ view, self and agenthood are extended and spread into
the world. Baker (2009) denies this claim. She seems sympathetic to the extended mind
thesis, but denies it entails an extended person. On an extended mind view, persons
are constantly shifting collections of biological and artifactual parts, implying there
is no continuity of personhood over time. Baker finds this hard to accept, because
on her view there are enduring persons. It is important to note that artifacts might
actually provide some stability for the continuity of personhood over time. Information
stored in artifacts is typically more durable and stable than information stored in
biological memory (Donald 1991). External information such as pictures, journal
entries, significant objects, or indeed lifelogs is often very durable, potentially existing
in our lifeworld for decades. Our biological bodies and cognitive capacities change,
sometimes quite dramatically, during our lifetime. But some of the artifacts we keep
123

Synthese
in our homes or stored in our computers are quite durable, providing long-lasting
information we can lock onto at various stages in our life (Clowes 2012).
A more specific example concerns Alzheimer’s patients who structure their homes
such that the location of objects such as clothing, food, cups, utensils, the telephone,
the remote control, and so forth, is familiar to them. “Taking them out of their homes
is literally separating them from large parts of their minds” (Dennett 1996, p. 138).
Their self-structured environment, rather than being a constantly shifting collection
of objects, is—in fact—quite stable, providing continuity for their memory and per-
haps also for their self. A final example where external artifacts can provide stability
for memory and self is Leonard Shelby, the protagonist from Christopher Nolan’s
film Memento. Due to an “accident”, Leonard has a severe form of amnesia: the
memories of his past before the accident are uneffected, but he is now only able
to remember the last seven minutes of his conscious life. So every seven minutes,
he reawakens as if he sees the world for the first time. To deal with his memory
loss, he develops strategies to communicate to his future self (Schechtman 2008).
He makes polaroid photos which he annotates and labels with information; par-
ticularly crucial information is tattooed on his body. Leonard’s tattoos have both
semantic and episodic components (Sutton 2009) and function to partly replace his
biological memory systems, in that way providing some stability for memory and
self.
Olsen (2011) argues that if Otto’s notebook is part of Otto as a person, it means
that persons are not purely biological organisms but organisms-plus-external-parts.
This implies that Otto is not his body, a conclusion difficult to accept for Olson who
defends a view on personal identity, arguing that we basically are human animals.
On Olsen’s view, we are our biological bodies and our mental life and memories
are not terribly important for our personal identity. Ultimately, Olson’s denial of the
extended self thesis is based on the denial of psychological properties (whether inter-
nally or partly externally instantiated) as a criterion for personal identity. I neither
have the space nor the intention to try to solve the debate between psychological
and biological approaches to personal identity. In this paper, I am assuming that psy-
chological or cognitive properties are essential for both synchronic and diachronic
aspects of personhood. The majority of philosophers working on personal identity
agree with this, but some philosophers, such as animalists, would not agree with this
claim.
Like Baker, Buller (2013) accepts the extended mind thesis, but denies it entails
an extended person. A person, on his view, is an embodied agent with, in some cases,
an extended mind. “Otto qua person coincides with the skin-and-skull boundary even
though his mind is extended beyond this boundary” (2013, p. 600). Buller takes con-
scious experience as an essential element of personhood and because experience cannot
be extended but only realized by an embodied brain, personhood cannot be extended
either. I agree with Buller that consciousness or conscious experience is not extended,
but I am not sure whether that means personhood is not extended either. Buller seems
to imply that only conscious states are relevant for the self. But it is odd to say that, for
instance, non-occurrent beliefs or non-occurrent information in memory are irrelevant
for the self. Reducing the self to conscious states would, as Clark and Chalmers put it,
123

Synthese
“shrink the self to a bundle of occurrent states” (1998, p. 18). This seems too narrow
aviewtome
4.
The above criticism, in one way or another, emphasize the unity of the biological
organism as essential for personhood. But does personhood necessarily coincide with
one’s biological organism? In the next section, I suggest that the answer this question
is “no”. But before doing that, let me briefly point out that certain prostheses can com-
municate with the central nervous system and are physically connected to the body.
These prostheses are under full control of their users and after a long and sometimes
difficult period of training experienced as a transparent part of their embodied self
(Clark 2007). Likewise, a blind man’s cane is not so much part of his external envi-
ronment, but absorbed by his body schema, transparent in use, and therefore part of
the apparatus with which he encounters the world. It seems that personhood might
not always coincide with the biological body. If prostheses and other tools can be
experienced as transparent parts of the embodied self, then why not a notebook or
other cognitive artifact?
4 Distributed selves
Two questions are often distinguished in relation to personal identity. First, what
defines being a person? Second, is personhood persistent over time? The first ques-
tion is concerned with synchronic aspects and the second with diachronic aspects of
personal identity. Extended and distributed cognition have important implications for
both of these questions.
4.1 External resources and the synchronic self
Rather than stipulating the metaphysical essence of personhood in terms of necessary
and sufficient conditions, I use the concept of practical identity (Mackenzie 2008;
Mackenzie and Walker 2015). Notions of practical identity try to answer questions
about who we are as persons and what characterizes our first-personal perspective. An
important element of our practical identity are our cognitive capacities, i.e., what we are
cognitively capable of.5If our cognitive capacities are important for our personhood
and if cognitive artifacts are “the things that make us smart” (Norman 1993), then such
artifacts are important for a synchronic conception of personhood. This is supported by
the fact that humans often define themselves partly in terms of the cognitive skills they
have. We see ourselves as analytical thinkers, as creative, as good problem-solvers, as
good at language, and so forth. This is sometimes reflected in the jobs we have. We
become engineers because we are good at analytical thought, we become designers
4A reviewer pointed out that Locke’s original memory criterion focused on consciousness too. In fact,
memory was Locke’s way of explaining the notion of temporally extended consciousness; memory makes
the idea of a temporally extended consciousness fall within the experience-centered empiricist outlook.
Many neo-Lockeans, including Schechtman, include non-occurrent states in their notion of psychological
continuity.
5There are, of course, other capacities that are important for our practical identity such as our capacity
for consciousness and emotions.
123

Synthese
because we are creative, we become writers because we are good at language, etc.
Having certain cognitive capacities is thus important for the kind of person we are.6
Extended and distributed cognition theory have extensively studied how we use
artifacts to aid our cognition and problem-solving abilities (Norman 1993;Hutchins
1995;Kirsh 1995;Hollan et al. 2000). We have seen that embodied agents like us
have limited information-storage and processing capacities. To complement these lim-
itations, we use maps, diagrams, models, diaries, timetables, textbooks, calculators,
computer systems, and many other artifacts to help us perform our cognitive tasks
such as planning, navigating, calculating, learning, and remembering. We also rely on
and interact with other people to complete our cognitive tasks (Wegner 1995;Sutton
et al. 2010;Harris et al. 2010). Our cognitive capacities are thus relational in that they
often depend on social and artifactual structures in the environment in which we are
situated. Ontogenetically, our cognitive capacities develop by interacting with par-
ents, caregivers and teachers as well as artifacts like abacuses, multiplication tables,
textbooks, learning applications in computer systems, and so on. Other people like
teachers, parents, or peers often teach us how to successfully use such artifacts to
achieve our cognitive goals. Our embodied brains learn to incorporate people and arti-
facts into our problem-solving routines, in that way transforming and augmenting the
way we think and remember. If we remove the external resource from the integrated
system, our cognitive competences will drop. Without incorporating other people and
artifacts into our cognitive practices, we simply would not have the same cognitive
capacities. For this reason, external resources are essential for our cognitive capacities
in the here-and-now and thus also for our synchronic personhood.
4.2 External resources and the diachronic self
Psychological continuity theories of personal identity often focus on memory as the
criterion for persistence of personhood over time. Such views, dating back to John
Locke, argue that someone in the present is the same person as someone in the past
when that former person remembers having the experiences of the later person. For
personhood to persist over time, there must thus be direct memory connections to past
experiences. So the content of a memory at t2provides a cognitive connection to an
experience had at t1, in that way creating a link between two time-slices of a person.
On such a neo-Lockean view, our memory system is seen as a storehouse in which
individual memories are stored for later retrieval in their original form and content.
However, Schechtman (1994) argues that human biological memory often does not
provide such direct memory connections between two discrete states of consciousness.
Rather, we summarize and condense important life experiences into a narrative. This is
not to say that we do not have direct and specific memories of past experiences, we do.
Schechtman’s point, I take it, is that most of our autobiographical memories are also
integrated into a larger narrative structure. We often remember that we did something,
but rarely remember specific details of an event. For example, we may remember
6If, for example, someone has a stroke or other kind of brain injury and loses part of one’s memory or
language capacity, that would surely be a change to the person.
123

Synthese
that we went to see the film Interstellar last year, but have trouble remembering
specific images or details of its plot. Whilst we may remember a few specific details,
we certainly do not remember the entire film. Memories are rarely exact, detailed
replicas of an event or experience but are often condensed into general summarised
structures, containing only the gist of things and are in some cases even inaccurate.
Thus, focussing on direct memory connections to past experiences is insufficient for
establishing psychological continuity of personal identity.
Schechtman (1996) has therefore developed a notion of personal identity that seeks
to explain the diachronic constitution of our self in terms of a narrative. She argues
that “We constitute ourselves as persons by coming to understand our lives as narra-
tives with the form of the story of a person’s life” (2012, p. 335). Summarising and
condensing autobiographical memories makes sense as a cognitive strategy because
we receive an overwhelming amount of information about our own lives. “In autobio-
graphical memory the information about our lives that is constantly coming in should
be reconstructed as a more concise and comprehensible narrative which emphasises
the most significant factors of past experiences and depicts general and longstanding
patterns or activities with representative examples” (1994, p. 11).
To better understand the relation between autobiographical memory and one’s narra-
tive, Schechtman presents an analogy between an archive and a biography. An archive
contains all the historical facts of a person, whereas a biography is an integrated and
coherent story of someone’s life. To get a sense of the person behind the historical
facts we need a biographer to make sense of the events, looking for patterns, anom-
alies, significant and insignificant events, and causal connections between events.
Only when the facts are interpreted and integrated into a narrative, does a real per-
son emerge. There is, however, an important difference: we constantly construct and
revise our biography throughout our lives. Because our life is continuously unfolding,
our narrative keeps changing. Over time, we have new significant experiences (e.g.,
graduating, marriage, divorce, new job, moving, traveling, death of a loved one, new
friendships, and so on), which are then integrated into the overall narrative. Our nar-
rative is thus a dynamic structure and the relationship we have to our narrative might
change. Events that used to be important to us may now appear to be less important,
and vice versa. Another difference is that most people do not have a biographer that
interprets the facts and constructs a narrative, but make their own autobiographical
narratives.
How do cognitive artifacts relate to our autobiographical narrative? I now argue that
external artifacts can, in some cases, play an essential and constitutive role in one’s
autobiographical narrative and thus also in one’s personal identity. This is particularly
the case for people with memory disorders such as traumatic brain injury, mild cogni-
tive impairment, Alzheimer’s disease or other forms of dementia. Such patients suffer
from a decline in their memory capacities, including their autobiographical memory.
Psychological research has shown that this has a negative impact on their sense of
self (Addis and Tippett 2004). On Schechtman’s narrative self-constitution view, “the
limits of a person are determined by the limits of a narrative, and the integrity of a
single person consists in the unity of a narrative” (2012, p. 336). If memory disor-
ders disintegrate one’s autobiographical narrative, then they also disintegrate one’s
personal identity.
123

Synthese
In a somewhat similar way as Otto extends his semantic memory when using his
notebook, artifacts may extend one’s autobiographical memory. Specifically, cognitive
artifacts can help to maintain the integrity, delay the disintegration, or in some cases
replace one’s autobiographical narrative. Lifelogging technologies, in particular, are
useful to achieve these goals (Berry et al. 2007;Doherty 2012;Crete-Nishihata 2012).
A striking example in this context is SenseCam, creating a visual narrative of one’s
daily life by taking a picture with a certain interval or when its sensors detect a change
in environmental conditions. Berry et al. (2007) performed one of the first empirical
studies of a patient with amnesia, using SenseCam to develop a photographic diary as to
improve autobiographical memory. They showed that reviewing SenseCam images of
personal experiences improves consolidation and retrieval of autobiographical memo-
ries for specific events, both on a short and long term. In similar research, Loveday and
Conway (2011) demonstrate that SenseCam images are more effective than written
records of personal experiences in cueing what they refer to as otherwise “inaccessible
memories”. So the images are necessary to evoke memories, otherwise they are likely
to remain inaccessible. SenseCam images have this causal efficacy because they are
cognitively comparable with biological memory in that “they are visual, represent
short time slices of experience, they are time compressed and fragmentary, they are
formed outside intentional awareness, they preserve the perspective of the individual
at the time of the event, and they are temporarily ordered” (Loveday and Conway 2011,
pp. 701–702). There are thus a lot of similarities between biological autobiographical
memory and SenseCam images.
More recently, Crete-Nishihata (2012) used SenseCams and other lifelogging tech-
nologies to create a digital narrative of people with memory disorders. Their project
aims to understand the implications of personal memory technologies for cognitively
impaired persons. They report on three of their projects. In the first, they created a “mul-
timedia biography” for twelve people with mild cognitive impairment or Alzheimer’s
disease. These multimedia biographies consist of photos, home movies, documents,
letters, music, and narration complied into a digital video that represents a person’s
life history. The patients and their family members provided the material and also
directed the script and storyboard of the narrative. These biographies are between 15
and 60 min and are divided into a series of acts, each with a number of scenes. The
acts are often major life events or stages such as adolescence, marriage, career, and
hobbies, which are told chronologically. The researchers write that “family members
and participants perceived the multimedia biographies as a means for preserving the
personhood of their loved one…” (2012, p. 101, italics added).
Personalised photos and music selections often elicited the strongest emotional
responses. The researchers learned that telling a life story is essentially a reconstructive
process. It is not so much “about documenting clear objective series of facts. Instead,
each user of personal memory technologies may perceive the past in different ways
and want their interpretation reflected in the media” (2012, p. 103). This shows that
one’s autobiographical narrative is socially constructed. An important conclusion is
that “intimate technologies for constructing personal identity and experiences must
allow editing of the “past” to match one’s current viewpoint” (2012, p. 103). This
is consistent with Schechtman’s narrative view, as she pointed out that the relation
we have to our narrative changes: events that used to be important can now be less
123

Synthese
important to us. So being able to change or edit one’s lifelog such that it is consistent
with one’s narrative is an important property of lifelogs and multimedia biographies.
In their second and less extensive study, they made a multimedia biography that also
included recently taken SenseCam images for a patient with mild Alzheimer’s disease.
In this study, the aim was to examine the social-psychological effects of multimedia
biographies by studying in more detail the authoring and viewing processes, both
of which had a positive impact on the patient’s apathy and self-image. So, not just
autobiographical memory, but also one’s self-image improves when using lifelogging
technology.
In their third study, they compared two ways of viewing SenseCam images. An
unedited version containing all images and an edited version containing only 24 images
and narrated by the partner of the cognitively impaired person. The results were mixed
as three out of five patients preferred the unedited version as there is more information
and detail to prompt their memories. Noteworthy is that the partners were the editors
and narrators, the patients themselves were not involved. One of the patients pointed out
that the edited version made her feel a little annoyed because her partner picked things
that meant something to him, but not things that meant something to her. So, again,
authorship of and agency over the content of the lifelog is thus (perhaps obviously)
an important factor for its cognitive and narrative efficacy.
Are multimedia biographies and lifelogs mere cues to activate or trigger internally
encoded memories or can they perhaps be seen as genuine external memories? The
research above seems to suggest they are cues to internal memory. However, if someone
is unable to activate or retrieve an autobiographical memory and can only retrieve it
by using an external cue, then it seems plausible to say that the external cue is part
of the overall memory system. The processing of memories is often divided in three
stages: encoding, storing, and retrieving. Memory disorders may occur at any of these
stages. An autobiographical memory may still be stored in the brain, but an agent
may not be able to retrieve it without help (Loveday and Conway 2011). If someone
needs an external cue to successfully retrieve a memory, then we might say the cue
is part of the overall distributed memory system. Crete-Nishihata (2012) looked at
patients with mild cognitive impairment or in the first stages of Alzheimer’s disease.
These patients still have a reasonably intact memory system. However, in some cases,
patients will have (much) lesser intact memory systems and have trouble encoding,
storing and retrieving memories. For those patients, a lifelog can literally be their
autobiographical memory in the same way as Otto’s notebook is literally his semantic
memory.
5 Enhancement or diminishment?
In the previous subsection, I mainly focussed on people with memory impairments
and how they use lifelogging technologies such as SenseCams to aid or supplement
their autobiographical memory. It is difficult to see these assistive technologies as
cognitively, morally, or culturally undesirable as they help people to deal with their
cognitive impairments, thereby increasing their autonomy, which results in a more
positive self-image and a better quality of life. However, lifelogging technologies are
123

Synthese
also used by cognitively unimpaired agents (O’Hara et al. 2008). Some lifeloggers,
such as Gordon Bell, take it quite seriously and have extensive lifelogs of personal
experiences and other personally-relevant information, going back decades. Bell and
most other members of the quantified self movement have an intact autobiographical
memory system and therefore a different relation to their lifelog. For some of the
above discussed patients, their lifelog or multimedia biography is (at least potentially)
constitutive of their autobiographical narrative. Whether lifelogs are constitutive of
other agents’ autobiographical memory, depends on the way they are used and how
deeply they are integrated into one’s cognitive system. The degree of integration of
a lifelog into autobiographical memory depends on factors such as the availability of
the lifelog, how often it is used, its transparency in use, the ease with which it can be
edited, and its capacity to capture relevant experiences that are consistent with or fit
into one’s narrative. Also, it is likely that lifelogs contain images of experiences that
its user has (long) forgotten. If the SenseCam had not captured the experience, its user
might have never thought about it again, implying that the lifelog could be a necessary
cue for internally encoded memories.
Regardless of whether lifelogs are constitutive of autobiographical memory, it is
likely that they transform onboard autobiographical memory systems. An effect of
lifelogging could be that lifelogs reduce the detail and amount of internally encoded
autobiographical memories. If users know their SenseCam reliably records images
of personal experiences and regularly review their lifelog, they might put less effort
(consciously or unconsciously) into encoding the experience in biological memory.
Empirical research in cognitive psychology shows that when we know information is
available in some external media, we tend to put less effort into memorizing that infor-
mation internally (Sparrow et al. 2011). Similarly, people taking pictures of objects
in an art museum, have less memories of the photographed objects as compared to
museum objects they had not photographed (Henkel 2014). When writing about lifel-
ogging technology, Sherry Turkle asks: “If technology remembers for us, will we
remember less?” (2011, p. 300). Research thus suggests the answer to this question
might be “yes”. Given the human cognitive system is a pragmatic system evolved to
make the best of its informational environment, this should not come as a surprise. In
our recent cognitive evolution, we have consistently relied on external information to
perform our cognitive tasks (Donald 1991). Modern lifelogging technology is just the
latest in a long history of memory technologies.
Furthermore, narrative identity is as much about forgetting as it is about remember-
ing. For whatever reason, we sometimes want to forget an experience, but lifelogging
may not always allow one to do that, given its focus on “total recall” (Mayer-
Schonberger 2011;Jacquemard et al. 2014). Commenting on this issue, Luciano Floridi
points out: “The more memories we accumulate and externalise, the more narrative
constraints we provide for the construction and development of personal identities.
Increasing our memories means decreasing the degree of freedom we might enjoy in
defining ourselves” (2011, p. 562). It seems true that a lifelog not only helps one to
remember experiences—which is often beneficial, depending on one’s situation—but
also limits the options one has to construct one’s narrative. Awkward discrepancies
may potentially emerge between one’s lifelog and internally constructed narrative.
However, having control over the contents of one’s lifelog and being able to edit it,
123

Synthese
seems to, at least potentially, solve this issue.7It is also important to note that similar
uncomfortable moments may occur when reading one’s past journal entries or seeing
photos of past events that remind one of experiences one rather forgets. Moreover,
sometimes awkward experiences may actually be meaningful for one’s autobiograph-
ical narrative. In the film, Eternal Sunshine of the Spotless Mind, the protagonist
removes the memories of his failed romantic relationship, thereby removing it from
his overall narrative. The point of the film, however, is that such experiences are part
of a meaningful and authentic life.
One might ask what would happen if photos of events that one did not actually
experience are somehow included in someone’s lifelog or multimedia biography.8
The person in question, further, does not realise those photos are not taken by his
or her own SenseCam and believes them to be true and part of the narrative. This
means that one’s digital narrative can be inconsistent with (at best) or potentially
even distort reality (at worst).9However, internally stored narratives are not without
inconsistencies and inaccuracies either. Schechtman outlined two narrative constraints:
the reality and the articulation constraint. The reality constraint says that the narrative
must be consistent with the laws of nature. So the protagonist cannot live as long as
Methuselah, travel faster than the speed of light, or be at two places at the same time.
The articulation constraint says that the person must be able to articulate parts of the
narrative when appropriate. But note that Schechtman does not claim that the narrative
must be consistent with actual events. So false or inaccurate memories may (and often
are) be part of one’s narrative, either internally or externally stored. People sometimes
leave out events or exaggerate others. One’s narrative rarely is a complete and accurate
reflection of one’s life.
Finally, a number of situated cognition theorists argue that the human self is essen-
tially a soft self. On Clark’s extended cognition view, for example, “our best tools and
technologies literally become us: the human self emerges as a soft self, a constantly
negotiable collection of resources easily able to straddle and criss-cross the boundaries
between biology and artifact” (Clark 2007, p. 278). If cognition and self are indeed
distributed, the constitutive parts of those distributed systems obtain a particular moral
status. Destroying Otto’s notebook, for example, “has an especially worrying moral
aspect: it surely is harm to the person, in about as literal a sense as can be imag-
ined” (Clark 1997, p. 215). Neuroethics and the cognitive enhancement debate have
7A reviewer pointed out that emphasizing control over the contents of one’s lifelog can be interpreted as
claiming that the locus of personhood is still brain-based. My view is that agency and the locus of personhood
can be distributed across humans and cognitive artifactssuch as SenseCam-generated lifelogs. I think agency
is distributed because the SenseCam automatically takes pictures whose content we typically cannot change,
but the human agent often has to edit the lifelog. So both components in the larger systems have some form
of agency and the equilibrium may shift between the artifactual and biological components, depending on
each specific case. Furthermore, in the typical extended mind cases (Clark and Chalmers 1998), agency is
largely located in the biological organism. However, in larger distributed cognitive systems, for example
Hutchins’ (1995) example of ship navigation, humans may not necessarily be the center and controller of
distributed systems. For more discussion on this topic see Hutchins (2014).
8This is related to the problem of q-memory as outlined in Parfit (1984).
9These are important epistemological issues to further explore. They also tie in with the extended knowl-
edge debate (e.g., Michaelian 2014), which has to do in part with the possibility of assigning credit to agents
who rely on external cognitive resources for their epistemic accomplishments, including those are false.
123

Synthese
traditionally focussed on technologies that interact with or alter the brain such as tran-
scranial magnetic stimulation, psychopharmaceuticals, brain-computer interfaces, and
so on. However, the argument developed in this paper implies that cognitive artifacts
should get a much more prominent role in neuroethics (Levy 2007a,b). Thus, if the
mind and self are, in some cases, partly constituted by external information, we ought
to treat that information in a similar way as we treat mental states instantiated purely
by the brain. What does this mean concretely? In addition to traditional legal and
moral issues of artifact-ownership, it means we should not interfere with people’s dis-
tributed minds and selves (Heersmink 2015). So we should not interfere with people’s
lifelogs, diaries, journals, photo albums and other artifacts related to autobiographical
memory. On this view, we also should not take Alzheimer’s patients out of (or change)
their personalized, self-structured environments, as that would reduce their minds and
selves and worsens their situation.
6 Conclusion
In this paper, I argued that not just memory and cognition, but also personal identity
depends on and is sometimes constituted by the social and technological environment
in which we are situated. The complex web of relations we develop and maintain with
other people and artifacts partly constitutes our self, implying that we are essentially
a soft self. Personal identity can thus neither be reduced to psychological structures
instantiated by the brain nor by biological structures instantiated by our biological
organism. If our minds and selves are indeed changing combinations of biological
and artifactual resources, we should be careful in selecting and creating the most
appropriate resources available, as these artifacts partly determine who we are. For
this reason, neuroethics should focus more on artifacts and material culture and develop
approaches to determine the best ways to distribute our selves. The view developed
in this article has conceptual, methodological, and normative implications: we should
broaden our concepts of the self as to include social and artifactual structures, focus
on external memory systems in the (empirical) study of personal identity, and not
interfere with people’s distributed minds and selves.
Acknowledgments I would like to thank Neil Levy for helpful advice, Paul Smart and Robert Clowes
for discussion on embodiment and the self, and two anonymous reviewers for their helpful comments and
suggestions.
References
Addis, D. R., & Tippett, L. J. (2004). Memory of myself: Autobiographical memory and identity in
Alzheimer’s disease. Memory,12(1), 56–74.
Anderson, J. (2008). Neuro-prosthetics, the extended mind, and respect for persons with disability. In M.
Düwell, C. Rehmann-Sutter, & D. Mieth (Eds.), The contingent nature of life: Bioethics and limits of
human existence (pp. 259–274). Heidelberg: Springer.
Baker, L. R. (2009). Persons and the extended mind thesis. Zygon,44(3), 642–658.
Beach, K. (1988). The role of external mnemonic symbols in acquiring an occupation. In M. M. Gruneberg,
P. E. Morris, & R. N. Sykes (Eds.), Practical aspects of memory (pp. 342–346). Chichester: Wiley.
123

Synthese
Bell, G. M., & Gemmell, J. (2009). Total recall: How the e-memory revolution will change everything.New
York: Dutton.
Berry, E., et al. (2007). he use of a wearable camera, SenseCam, as a pictorial diary to improve autobi-
ographical memory in a patient with limbic encephalitis: A preliminary report. Neuropsychological
Rehabilitation,17(4—-5), 582–601.
Buller, T. (2013). Neurotechnology, invasiveness and the extended mind. Neuroethics,6(3), 593–605.
Bush, V. (1945). As we may think. The Atlantic,176(1), 101–108.
Clark, A. (1997). Being there: Putting brain, body, and world together again. Cambridge, MA: MIT Press.
Clark, A. (2005). Intrinsic content, active memory and the extended mind. Analysis,65(285), 1–11.
Clark, A. (2007). Re-inventing ourselves: The plasticity of embodiment, sensing, and mind. Journal of
Philosophy and Medicine,32(3), 263–282.
Clark, A., & Chalmers, D. (1998). The extended mind. Analysis,58(1), 7–19.
Clowes, R. W. (2012). Hybrid memory, cognitive technology and self. In Y.J. Erden & J.M. Bishop
(Eds.),Proceedings of AISB/IACAP world Congress (pp. 4–13).
Crete-Nishihata, M., et al. (2012). Reconstructing the past: Personal memory technologies are not just
personal and not just for memory. Human-Computer Interaction,27(1–2), 92–123.
Dennett, D. (1996). Kinds of minds: Towards an understanding of consciousness. New York: Basic Books.
Doherty, A. R., et al. (2012). Experiences of aiding autobiographical memory using the SenseCam. Human-
Computer Interaction,27(1–2), 151–174.
Donald, M. (1991). Origins of the modern mind: Three stages in the evolution of our cognitive system.
Cambridge, MA: Harvard University Press.
Floridi, L. (2011). The informational nature of personal identity. Minds and Machines,21, 549–566.
Harris, C. .B, Keil, P. G., Sutton, J. & Barnier, A. J. (2010). Collaborative remembering: When can remem-
bering with others be beneficial? In W. Christensen, E. Schier & J. Sutton (Eds.), Proceedings of the
9th conference of the Australasian Society for Cognitive Science (pp. 131–134).
Henkel, L. A. (2014). Point-and-shoot memories: The influence of taking photos on memory for a museum
tour. Psychological Science,25(2), 396–402.
Heersmink, R. (2015). Extended mind and cognitive enhancement: Moral aspects of cognitive artifacts.
Phenomenology and the Cognitive Sciences. doi:10.1007/s11097-015- 9448-5.
Hodges, S., Berry, E., & Wood, K. (2011). SenseCam: A wearable camera that stimulates and rehabilitates
autobiographical memory. Memory,19(7), 685–696.
Hollan, J., Hutchins, E., & Kirsh, D. (2000). Distributed cognition: Toward a new foundation for human-
computer interaction research. Transactions on Computer-Human Interaction,7(2), 174–196.
Hutchins, E. (1995). Cognition in the wild. Cambridge, MA: MIT Press.
Hutchins, E. (2014). The cultural ecosystem of human cognition. Philosophical Psychology,27(1), 34–49.
Jacquemard, T., Novitzky, P., O’Brolchain, F., Smeaton, A. F., & Gordijn, B. (2014). Challenges and
opportunities of lifelog technologies: A literature review and critical analysis. Science and Engineering
Ethics,20(2), 379–409.
Kirsh, D. (1995). The intelligent use of space. Artificial Intelligence,73(1–2), 1–52.
Levy, N. (2007a). Rethinking neuroethics in the light of the extended mind thesis. American Journal of
Bioethics,7(9), 3–11.
Levy, N. (2007b). Neuroethics: Challenges for the 21s t century. Cambridge: Cambridge University Press.
Loveday, C., & Conway, M. (2011). Using SenseCam with an amnesic patient: Accessing inaccessible
everyday memories. Memory,19(7), 697–704.
Mackenzie, C. (2008). Introduction: Practical identity and narrative agency. In C. Mackenzie & K. Atkins
(Eds.), Practical identity and narrative agency (pp. 1–28). New York: Routledge.
Mackenzie, C., & Walker, M. (2015). Neurotechnologies, personal identity, and the ethics of authenticity.
In J. Clausen & N. Levy (Eds.), Handbook of neuroethics (pp. 373–392). Dordrecht: Springer.
Mayer-Schonberger, V. (2011). Delete: The virtue of forgetting in the digital age. Princeton, NJ: Princeton
University Press.
Menary, R. (2007). Cognitive integration: Mind and cognition unbound. London: Palgrave McMillan.
Michaelian, K. (2014). JFGI: From distributed cognition to distributed reliabilism. Philosophical Issues,
24(1), 314–346.
Michaelian, K., & Sutton, J. (2013). Distributed cognition and memory research: History and current
directions. Review of Philosophy and Psychology,4(1), 1–24.
Norman, D. (1993). Things that make us smart: Defending human attributes in the age of the machine.New
York: Basic Books.
123

Synthese
O’Hara, K., Tufflied, M. M., & Shadbolt, N. (2008). Lifelogging: Privacy and empowerment with memories
for life. Identity in the Information Society,1(1), 155–172.
Olsen, E. T. (2011). The extended self. Minds and Machines,21(4), 481–495.
Parfit, D. (1984). Reasons and persons. Oxford: Oxford University Press.
Rowlands, M. (1999). The body in mind: Understanding cognitive processes. Cambridge: Cambridge Uni-
versity Press.
Schechtman, M. (1994). The truth about memory. Philosophical Psychology,7(1), 3–18.
Schechtman, M. (1996). The constitution of selves. Ithaca, NY: Cornell University Press.
Schechtman, M. (2008). Diversity in unity: Practical unity and personal boundaries. Synthese,162(3),
405–423.
Schechtman, M. (2012). The story of my (second) life: Virtual worlds and narrative identity. Philosophy
and Technology,25(3), 329–343.
Smart, P., Heersmink, R., & Clowes, R. (2016). The cognitive ecology of the internet. In S. Cowley & F.
Vallée-Tourangeau (Eds.), Cognition beyond the brain: Computation, interactivity and human artifice
(2nd ed.). Dordrecht: Springer.
Sparrow,B., Liu, J., & Wegner, D. M. (2011). Google effects on memory: Cognitive consequences of having
information at our fingertips. Science,333(6043), 776–778.
Sutton, J. (2009). The feel of the world: Exograms, habits, and the confusion of types of memory. In A.
Kania (Ed.), Philosophers on Memento (pp. 65–86). London: Routledge.
Sutton, J. (2010). Exograms and interdisciplinarity: History, the extended mind, and the civilizing process.
In R. Menary (Ed.), The extended mind (pp. 189–225). Cambridge, MA: MIT Press.
Sutton, J., Harris, C. B., Keil, P., & Barnier, A. J. (2010). The psychology of memory, extended cognition,
and socially distributed remembering. Phenomenology and the Cognitive Sciences,9(4), 521–560.
Theiner, G. (2013). Transactive memory systems: A mechanistic analysis of emergent group memory.
Review of Philosophy and Psychology,4(1), 65–89.
Turkle, S. (2011). Alone together: Why we expect more from technology and less from each other.New
York: Basic Books.
van den Hoven, E. (2014). A future-proof past: Designing for remembering experiences. Memory Studies,
7(3), 373–387.
van den Hoven, E., Cas, C., & Whittaker, S. (2012). Introduction to this special issue on designing for
personal memories: Past, present, and future. Human-Computer Interaction,27(1–2), 1–12.
Wegner, D. M. (1995). A computer network model of human transactive memory. Social Cognition,13(3),
319–339.
Wegner, D. M., Giuliano, T., & Hertel, P. T. (1985). Cognitive interdependence in close relationships. In W.
J. Ickes (Ed.), Compatible and incompatible relationships (pp. 253–276). New York: Springer-Verlag.
Wilson, R., & Lenart, B. (2014). Extended mind and identity. In J. Clausen & N. Levy (Eds.), Handbook of
neuroethics (pp. 423–439). Dordrecht: Springer.
123