Context, Embodiment and Online Surrogacy: the Impact of 4E Cognition Research on the Ontario SmartData Initiative
Submitting InstitutionUniversity of Stirling
Unit of AssessmentPhilosophy
Summary Impact TypeSocietal
Research Subject Area(s)
Psychology and Cognitive Sciences: Psychology, Cognitive Sciences
Philosophy and Religious Studies: Philosophy
Summary of the impact
Wheeler's internationally influential research on 4E
(embodied-embedded-extended-enactive) cognition has played a formative
role in the genesis and development of SmartData, an innovative
public-policy-realizing project spearheaded by the Office of the
Information and Privacy Commissioner of Ontario (IPC). IPC is an agent of
the Ontario legislature that acts to uphold and promote the protection of
personal privacy. The goal of SmartData is to enhance personal data
privacy on the Web, by developing Internet-based agents that act as online
surrogates for individuals, securely storing and intelligently disclosing
personal information as required. The identified research has influenced
the content and implementation of SmartData, by contributing to (i) the
IPC's understanding of the link between privacy and context-sensitivity,
(ii) the principles shaping the design of the online agents, and (iii) the
specification of the first `proof-of-concept' SmartData product, an
under-development mobile phone book ordering application.
Wheeler was appointed to a Senior Lectureship in Philosophy at Stirling
in 2004. He was promoted to Reader (2006) and then to Professor (2009).
Over this period, he has produced a widely discussed series of research
publications on the complex ways in which mind, body and world are
causally and constitutively inter-related. This work on so-called 4E
(embodied-embedded-extended-enactive) cognition includes a much cited
monograph, Reconstructing the Cognitive World (2005) and over 25
subsequent publications in peer-review journals and edited collections.
Reconstructing the Cognitive World, which is the principal source
of the opening phase of the impact described below, was completed
following Wheeler's appointment at Stirling and published by MIT Press in
2005. The book draws on sources as seemingly disparate as phenomenology,
evolutionary robotics and dynamical systems theory, in order to articulate
and defend a non-standard philosophical framework for cognitive science. Reconstructing
the Cognitive World explores a number of interconnected themes that
Wheeler has continued to investigate in subsequent research. This later
research has resulted in — and, in a manner to be described below, has
sometimes been carried out in interaction with — the second phase of the
impact described here. Two longstanding themes in particular should be
highlighted: (a) the challenges that confront the attempt to explain, in
purely mechanistic terms, the remarkable but routine capacity of human
intelligence to be fluidly and flexibly sensitive to what is relevant in
context (a question which is intertwined with that of replicating such a
capacity via artificial intelligence — AI) and (b) the ways in which, and
the extent to which, human thought and reasoning are determined by
features of our physical embodiment (a question which immediately raises
issues regarding the possibility of creating disembodied — e.g. online —
surrogates for our own decision-making).
With regard to (a), Wheeler has argued as follows (publications 1, 2, 3,
5, 6): there are in fact two problems of relevance, the first
concerns how a purely mechanistic system might achieve appropriate,
flexible and fluid action within a context, while the second concerns how
a purely mechanistic system might achieve appropriate, flexible and fluid
action in worlds in which adaptation to new contexts is open-ended and in
which the number of potential contexts is indeterminate; although a
potential solution to the first of these problems exists (in terms of
special-purpose, brain-body-environment couplings), our scientific
understanding of the second remains radically incomplete. With regard to
(b), Wheeler has argued as follows (publications 4, 6): although the role
of physical embodiment in intelligent human action has been
under-appreciated in most philosophy and cognitive science, what really
matters about that embodiment is adequately captured by a view according
to which the body is conceptualized as a material realizer of functionally
specified cognitive architectures; so the temptations of a more radical
(anti-virtuality) embodied position, according to which specific bodily
acts and structures make some nonsubstitutable contribution to thought and
reason, should be resisted. One outcome of this research is that Wheeler
now has a reputation as a philosopher whose work is highly relevant to the
understanding and future development of our online presence (see
corroborating source A).
References to the research
1. Wheeler, M., Reconstructing the Cognitive World: the Next Step,
MIT Press, Cambridge, Mass., 2005, 339pp. Peer reviewed publication.
MIT Press is arguably the world's leading publisher of philosophy of
cognitive science. The book received extremely positive reviews in highly
reputable publications such as the Times Literary Supplement (by
the philosopher of mind and psychiatry George Graham, 23/06/06, "[a]
challenging defence of anti-classical assumptions for cognitive science"),
Mind (by the philosopher of embodied cognition Shaun Gallagher,
July 2007, "packed full of good and innovative arguments"), and Artificial
Life (by the AI researcher Ezequiel Di Paolo, Spring 2007; "could
turn out to be the initial articulation for the most radical break in
cognitive science in decades"). The reception and sales of the book
resulted in its paperback publication by MIT in 2007. It has been the
topic of four international workshops: Tilburg University, Netherlands,
2006; the Free University of Amsterdam, Netherlands, 2008; Goldsmiths
University of London, 2009; the Goethe University of Frankfurt, Germany,
2. Wheeler, M. `Cognition in Context: Phenomenology, Situated Robotics
and the Frame Problem', International Journal of Philosophical Studies,
16(3), 323-49, 2008. Peer reviewed publication.
3. Wheeler, M. `Plastic Machines: Behavioural Diversity and the Turing
Test'. Kybernetes, 39(3), 466-80, 2010. Peer reviewed publication.
4. Wheeler, M. `Embodied Cognition and the Extended Mind', in Garvey, J.
ed., Continuum Companion to the Philosophy of Mind, Continuum,
London, 2011, pp.220-238. Invited contribution reviewed by editor.
5. Wheeler, M. `Naturalizing Dasein and other (Alleged) Heresies', in
Kiverstein, J. and Wheeler, M. eds., Heidegger and Cognitive Science,
Palgrave-Macmillan, Basingstoke, 2012, pp.176-212. This collection
includes chapters by Hubert Dreyfus, Matthew Ratcliffe, Andrea Rehberg and
Erik Rietveld that criticize arguments from Wheeler's book Reconstructing
the Cognitive World, including, prominently, those concerning the
problem of relevance. Wheeler's `Naturalizing Dasein' piece responds to
6. Wheeler. M. `What Matters: Real Bodies and Virtual Worlds', in Harvey,
I., Cavoukian, A., Tomko, G., Borrett, D., Kwan, H. and Hatzinakos, D.
eds., SmartData: Privacy Meets Evolutionary Robotics, Springer,
Berlin, 2013, pp.69-80. Peer reviewed collection. The reviewer rated the
paper "excellent" and "very compelling"; accepted with no changes.
Details of the impact
SmartData is a project of the Office of the Information and Privacy
Commissioner of Ontario (IPC), in collaboration with the Identity,
Privacy, and Security Institute of the University of Toronto (IPSI). IPSI
is the home of the encryption specialist Dr. George Tomko who originally
conceived the technical dimension of SmartData. IPC is an agent of the
provincial legislature of Ontario that acts "to uphold and promote open
government and the protection of personal privacy in Ontario"
(corroborating source B). One of IPC's flagship policies is `privacy by
design', as proposed by Ann Cavoukian, the Information and Privacy
Commissioner of Ontario, and endorsed by the International Data Protection
and Privacy Commissioners meeting in Jerusalem in October 2010. As
Cavoukian puts it, "[p]rivacy can no longer be assured solely by
compliance with regulatory frameworks; rather, privacy assurance must come
from designing privacy protection right into technology, business
practices and physical design" (corroborating source C).
Against this background, the goal of SmartData is to develop Web-based
autonomous agents that will act as an individual's online surrogate,
securely storing their personal information (e.g. financial details,
information regarding medical conditions and treatment), and intelligently
disclosing that information with appropriate sensitivity to the context of
any data request. The vision is one of protecting/reclaiming individual
informational privacy in an age of increasing worries about the ways in
which businesses, governments and criminal organizations may access and
control online personal information. The model for SmartData is
evolutionary computation and embodied cognition within a dynamical systems
framework. It is built on an optimism that, using this model, it will be
possible to artificially evolve online agents that are capable of
disclosing personal information in a context-sensitive fashion. The human
beings for which these virtual agents are online surrogates will need only
to set up the initial conditions for data release contexts and then inform
the agent should these conditions alter. Decisions regarding data release
will then be made by the agent through its autonomous context-sensitive
application of these conditions. Cavoukian has described SmartData as "an
innovation that is desperately needed because without it, our data will
move beyond our reach and our ability to ultimately maintain control"
(corroborating source D).
The initial phase of the impact described here was direct but unplanned.
As Michelle Chibba, the IPC's Director of Policy and Special Projects, put
it "[Wheeler's] book "Reconstructing the Cognitive World"... was the
inspiration for the Smart Data vision" (email of 27 October
2011; see corroborating sources E, F). The SmartData model combines
embodied cognition, evolutionary robotics and a dynamical systems approach
to intelligence. This was precisely the set of ideas that was explored and
developed, in a distinctive way, in Wheeler's Reconstructing the
Cognitive World. The book also focussed on context-sensitivity in
intelligence, and on the role of the set of ideas just mentioned in
showing us how the capacity for such sensitivity may be mechanized.
The second phase of the impact involved a planned intervention. Wheeler
was invited by IPC to take part in the IPSI SmartData International
Symposium (`Privacy meets Evolutionary Robotics: Protecting our Freedoms
with Virtual Tools', Toronto, May 2012). This symposium refined SmartData
by bringing together academics, members of IPC (including the
Commissioner) and experts from the computer and Web industries (e.g. a
senior researcher from IBM's Blue Gene supercomputing project).
Representatives from Google and Facebook, plus privacy stakeholders in the
public sector, were present. Wheeler gave two presentations (Philosophical
Reflections on the Design of Autonomous Agents: the Problem of Relevance'
and `What Matters: Real Bodies and Virtual Worlds') in which he applied
arguments and insights from the research described above to SmartData. He
also took part in a 4 hour closed meeting which brought together selected
invited speakers, the local SmartData team, and representatives from IPSI
and IPC, to shape the future of the project. Wheeler was asked to submit a
paper to the symposium proceedings, published by Springer in their
`Security and Cryptology' series (publication 6). This allowed his
experience at the symposium to influence his research, demonstrating a
mutually profitable interaction between impact and research. (The
corroborating sources for this paragraph are detailed in G.)
IPC subsequently invited Wheeler to be a named collaborator on a major
grant application (1 million Canadian dollars) made via IPSI to the
University of Toronto's 2013 Connaught Global Challenges Fund Award. This
award aims to support a project that "focus[ses] intently on a leading
global challenge of the 21st century" (corroborating source H). SmartData
was one of three shortlisted applications (corroborating source H).
Although the bid was ultimately unsuccessful, IPC has been encouraged to
submit to the 2014 competition. In addition, in 2012, Wheeler drew on the
research described above to contribute to the design of a proof-of-concept
SmartData study, a mobile phone application for ordering books online, in
which a spoken request such as "Buy me a paperback copy of X, lowest
price, delivery time unimportant" will be carried out (including seller
credibility checks and payment) by an application that adapts over time to
user preferences. This study will test the SmartData concept, set a
baseline, and gain further insights into the theoretical and technical
issues. It is in progress.
Sources to corroborate the impact
A. For example, Wheeler was an invited speaker at `The Philosophy of the
Web' (Sorbonne and Pompidou Centre, Paris, 2012), a seminar organised in
part by Harry Halpin of the World Wide Web Consortium (W3C), the leading
international agency for Web standards (http://www.w3.org/).
B. Quotation from
C. Quotation from
D. Quotation from
E. In an email of 27 October 2011 to Wheeler from Michelle Chibba, she
writes: "[i]t was your book on "Reconstructing the Cognitive World" that
was the inspiration for the Smart Data vision. The thought leader behind
this new research area, Dr. George Tomko, read your book and the questions
that you raised resonated with him. It is for this reason we have reached
out to you because your grounding in philosophy alongside your interests
in artificial intelligence and artificial life is a much-needed area of
input as we formulate this vision of Smart Data agents."
F. See http://www.ipsi.utoronto.ca/sdis/bibliography.pdf#
for a list of "[w]orks related to the development of SmartData". Two of
Wheeler's publications (1 and 2 above) are mentioned.
G. See http://www.privacybydesign.ca/index.php/time-to-get-smart-about-big-data-enter-smartdata/
for a piece by Ann Cavoukian in which she writes: "More than 20 of the
world's most innovative thinkers from such fields as evolutionary
robotics, engineering, cognitive science, brain imaging, computer
sciences, philosophy and privacy, gathered recently at the University of
Toronto from May 14 to 16, for an International Symposium... All of the
speakers are world renowned experts in their fields including... Dr.
Michael Wheeler of the University of Stirling, Scotland... We were also
honoured to have Dr. Kirk Jordan from the Computational Science Center at
IBM Watson Research Center and Bob Blainey, an IBM Fellow at the Hardware
Acceleration Laboratory IBM Software Group presenting as well. These
experts turned their minds to launching this revolutionary concept." (The
same piece appears on the website of the Ontario Ministry of Research and
http://www.ipsi.utoronto.ca/sdis/index.html for the SmartData
symposium website. See http://www.ipsi.utoronto.ca/sdis/program.html
for a video of Ann Cavoukian's contribution to the Toronto symposium in
which she highlights the importance of context. See http://www.ipsi.utoronto.ca/sdis/program.html
for videos of Wheeler's talks at the symposium.
H. Quotation from http://www.research.utoronto.ca/u-of-ts-connaught-fund-announces-innovation-and-summer-institute-awards/.
The shortlist for the award is also published here.