Conference PaperPDF Available

Watching Me, Watching You. (Process Surveillance and Agency in the Workplace)

  • hrafARC and HRAF YALE - unpaid affiliation


The notion that computers are somehow separate from our lives is misleading and ignores the level of integration that has emerged. Most of the processes that dispense, load, and deliver the supplies that sustain cosmopolitan life are impacted by some form of computer in one way or another. The systems created when networks of computers intersect with networks of people are shaping our current cultural environment and the way that we exist in the world. This phenomena has created multiple types of interactions that are hybrids between humans and machines and at present, the balance of human behavior towards other humans is impacted by processes in business and elsewhere that have an over arching governance based on machines. This limits human agency and impacts understanding, service and privacy rights for humans. Further, these processes increasingly depend on greater and greater quantities of what had previously been considered personal information, often scraped from online processes people do not anticipate, yielding an often revealing portrait of themselves. Also, a poorly configured paradigm has created a culture where, when systems are required for big business, people more often alter their behavior to suit machines and work with them, rather than the other way around, and that this has eroded conceptions of agency. We explore the use of Thing Theory to implement a partial means of implementing mutual surveillance between management and workers to increase human agency while developing more adaptive and efficient business processes.
Pre-Pub Draft > CITE Publication to be available through (ISTAS’13) also
All material © 2013 Sally A. Applin and Michael D. Fischer. All rights reserved.
Pre-Pub Draft> CITE Publication to be available through (ISTAS’13) also
Watching Me, Watching You.
(Process Surveillance and Agency in the Workplace)
Sally A. Applin
Centre for Social Anthropology and Computing
School of Anthropology and Conservation
University of Kent, Canterbury, CT2 7NR
Michael D. Fischer
Centre for Social Anthropology and Computing
School of Anthropology and Conservation
University of Kent, Canterbury, CT2 7NR
Abstract The notion that computers are somehow separate from our
lives is misleading and ignores the level of integration that has emerged.
Most of the processes that dispense, load, and deliver the supplies that
sustain cosmopolitan life are impacted by some form of computer in one
way or another. The systems created when networks of computers
intersect with networks of people are shaping our current cultural
environment and the way that we exist in the world. This phenomena has
created multiple types of interactions that are hybrids between humans
and machines and at present, the balance of human behavior towards
other humans is impacted by processes in business and elsewhere that
have an over arching governance based on machines. This limits human
agency and impacts understanding, service and privacy rights for
humans. Further, these processes increasingly depend on greater and
greater quantities of what had previously been considered personal
information, often scraped from online processes people do not anticipate,
yielding an often revealing portrait of themselves. Also, a poorly
configured paradigm has created a culture where, when systems are
required for big business, people more often alter their behavior to suit
machines and work with them, rather than the other way around, and that
this has eroded conceptions of agency. We explore the use of Thing
Theory to implement a partial means of implementing mutual
surveillance between management and workers to increase human
agency while developing more adaptive and efficient business processes.
Keywords-surveillance; manufacturing; polysocial reality;
agency; multi-agent simulation
Surveillance isn't a new idea. People have been watching
others since they had others to watch. However, new ways of
collecting, aggregating and processing information is driving
surveillance to have the potential for a wide range of
constituencies. This includes the capacity to contribute to the
ongoing development of automation of processes which, when
adhered to, will greatly restrict human agency. If we consider
the workplace to be a structured environment that can, in part,
be thought of as a "canary in the coal mine" test suite for
pervasive surveillance, we gain insight on how surveillance
techniques might play out on a broader scale.
Surveillance techniques used in the workplace go beyond
merely capturing employees on cameras. Indeed, they
constrict in a more overt way--via processes that are tracked
and measured and require worker compliance. Furthermore,
workplace surveillance is unevenly distributed; where a
camera might record all parties for physical theft, process
adherence most affects workers on the lower rungs of the
corporate ladder. The lines between public knowledge and
personal privacy have been blurring for both workers and
management with the adoption of mobile devices and data
enabled smart phones. What process and surveillance most
affect is human agency, the capacity for humans to make
nondeterministic choices intended to advance toward a goal.
When agency is diminished, or disrupted via surveillance or
the process that requires them to follow a pre-determined
script, humans are restricted using their abilities to
independently think and problem solve creatively.
Part of the human experience is our heterogeneity, or
differences from one another. In a global workforce, the
heterogeneity of humans is compounded as multiple cultures
interact together to conduct their work via messaging through
communications channels such as correspondence, computers
and telephones. Simultaneously, the rise of new com-
munications behavior with the advent of the mobile web and
social media has imbued an additional layer of heterogeneity
to the global workforce on top of their usual communications,
creating options for behavior that can disrupt and interrupt the
production process as people check their devices for mail,
send more messages, look up information and/or talk while
they are working.
As the channels of synchronous and asynchronous
communications increase, more extensive and complex
interactions form and the information available to individuals
becomes more variable. This results in the emergence of a
structure we call PolySocial Reality (PoSR). Too much
asymmetric information exchanged without enough overlap
can lead to missed meaning. Unless this imbalance is
compensated for critical information that can impact human
cooperation and behavior in the workplace can be lost,
misunderstood or forgotten.
Independently, due to globalization, companies have sought to
control quality of production by defining more and more
detailed processes and process scripts in an attempt to manage
their highly heterogenous workforce. From a management
perspective, devising a 'common' language such as a process
that is surveyed seems a smart way to keep workers focused.
Pre-Pub Draft > CITE Publication to be available through (ISTAS’13) also
All material © 2013 Sally A. Applin and Michael D. Fischer. All rights reserved.
Pre-Pub Draft> CITE Publication to be available through (ISTAS’13) also
However, it also limits their agency and surveillance
techniques only further isolate and reinforce worker hetero-
geneity. Bi-directional or multi-directional surveillance
between worker and worker, and worker and management,
might address this problem, enabling workers to retain some
agency and quality of life, within the structure of the
We introduce Thing Theory, which describes a communal
surveillance agent, a Thing-agent. A Thing-agent develops
relationships between parties and shapes knowledge and
responses for different actors within the system that it serves.
A Thing-agent utilizes knowledge of a local environment,
gained via its relationship to all people, roles and processes in
the environment and to system-subagents and meta-agents in
other systems.
The structure in an organization, particularly a global
corporation, is complex. In order to contain the seemingly
overwhelming act of managing a consistently shifting global
structure, intermediate processes are introduced that help to
run various tasks and functions. These procedures and
processes create both a power structure (that is not always
overtly recognized as such), and a process structure that is
used to track and maintain the actions of a more massive
organization. Historically, large organizations were generally
governed by outcomes, with much of the process more or less
under the control of groups or individuals who were evaluated
mostly on those outcomes, and not the processes. Gradually,
corporations began to dictate more and more of the content of
the processes, partially in response to greater legal liability
and reporting requirements, but also due to early 'data mining.'
This led to corporate theories of how to maximize the impact
of the quarterly report, including the idea that there might be
some benefit from more control over processes. Some
organizations even developed 'scripts' or near 'scripts' for
many tasks. These scripts were (and are) put in place for
various employees to follow in order to complete the tasks
required by the organization to function. In a decision tree,
there is the idea that contingency exists. In the case of scripts,
the decision tree is moderated by a weak or false notion that
the process could be intended to be scripted with almost all
potential outcomes pre-calculated. This is different than
expecting any type of free will on the part of a worker to
execute the idea of the script rather than the script itself.
Doing so would cause non-standard responses that would be
more difficult for the corporation to quantify and utilize as
data. This also may explain why such a surprise and
puzzlement is sometimes evoked in the corporate sector when
an employee/visitor/customer asks for something that is
outside the range of the script.
These scripts are a type of automation for coordinating people
as a part of the process and thus, require surveillance to evolve
and be successfully maintained as such. Pervasive automation,
such as the control mechanisms that enable the function of
drones, remote cameras, and the streamlined organizational
process depends upon surveillance that must be put into place
to both control and monitor performance and outcomes.
In the current workplace, corporations are using scripts to
collect data with the goal of streamlining their internal
processes. Recently, there has been much discussion of the
tentative use of sensor-based tracking of workers to add to this
process. In some ways, this is not new. Trucks have had
monitoring sensors for some time to track driver attention on
the road [1-3], and floor supervisors in the factories of the past
certainly watched and paid attention to workers movement and
habits. Every person carrying a mobile device with GPS is
capable of being tracked, but it is not habitually used to affect
or control their production. That said, the way that this
particular type of sensor based tracking and surveillance is
done in the workplace has generated discussion around
privacy, human dignity, and the boundaries between people
and their livelihoods.
In general, surveillance is a vast subject and we’ve narrowed
the discussion of this aspect of process surveillance to the
workplace. Current process script design with or without
sensor-surveillance might be described by three principles:
1. Processes are created in the interest of cost saving,
streamlining, efficiency, etc. and limit the free will of
those enacting these by narrowing choice and experience
that could be applied to the job.
2. The actions of individuals within these systems are
recorded as data that is collected, aggregated, analyzed
and reviewed.
3. New processes and procedures are created based on the
interpretation of this data with little regard to the context
of the human experience, except as required by law--and
even then companies often attempt to skirt these laws [4].
The first principle of sensor-based surveillance inspired
process design can be illustrated by the warehouse computer
systems used in industry. Paperless order picking systems
have been in practice for some time and usually involve a
small computer, such as the Motorola WT400, strapped to the
forearm of a warehouse worker. This allows for hands free
movement, while the instructions are shown on the display [5].
The computer selector electronically receives an assignment,
which directs a worker to the location, where they pick the
product, and are able to stage it on a pallet [6]. In his
dissertation, "Order Picking Supported by Mobile
Computing," Hannes Baumann, suggests that "the
development of these type of paperless systems were focused
on the tasks of inspection, maintenance, manufacturing, repair,
and training as potential areas where wearable computing
might prove beneficial" [7]. In this way, paperless systems
were developed to replace having to carry paper and pens and
report in by turning in sheets at the end of the day, as the data
from the paperless ordering picker system generated reports
via computer. In addition, Baumann adds that "the fields of
Pre-Pub Draft > CITE Publication to be available through (ISTAS’13) also
All material © 2013 Sally A. Applin and Michael D. Fischer. All rights reserved.
Pre-Pub Draft> CITE Publication to be available through (ISTAS’13) also
wearable and ubiquitous computing have evolved from the
creation of laboratory prototypes to examining systems
deployed in workers’ and consumers’ everyday lives" [7].
Although Baumann's dissertation contains a thorough analysis
of Order Picking, he neglects to indicate that there is a new
area that these types of paperless systems might be focused on
besides cutting down on paper usage. The sensors within them
can be used for surveillance of worker movement and further,
for monitoring compliance to script processes. This type of
surveillance moves the rationale from tracking of data, to
tracking of behavior.
A former Tesco staff member alleged that a Tesco warehouse
in Ireland took liberties with their paperless order picking
system computers, using the system armbands to keep track of
their workers habits and productivity [8]. In summary, Tesco
was attributed to using both gamification and surveillance type
capabilities in addition to the function of collecting data about
their work processes. Worker performance was monitored via
a percentage score, which was claimed to be lowered if
workers took a longer lavatory break and raised if they worked
more quickly. Tasks were required to be completed within
fixed time frames as well. The staff member added that people
would be called before management if not "working hard
enough." Of note is that many of the former staff member's
colleagues who had to use the system armbands in the
workplace were eastern European immigrants with limited
English skills. The Tesco management had allotted a fixed
time of 25 minutes per day for stoppages, but other time was
monitored [8].
The difference in electronic tracking vs observational tracking
is that there can be mistakes attributed solely to data analysis
that lack proper context. Within this range, there is a narrow
window of tracked behavior that must be adhered to. There is
no room for fixing mistakes longer than the pre-determined
time that mistakes must take, and even less room for
individual creativity or ingenuity. The job expressed must only
fall within the requirements of the job's specifications and now,
if the claims are found to be true, the specifications of a time
duration as well. In this way, the capacity for problem solving
for the workers is limited. They have authority to fix problems,
but only for the time allotted by Tesco. The windows of their
ability to exercise free will decrease.
Sometimes, a corporation will engage in a behavioral tracking
type of surveillance on their customers to inform their
processes. This example is included because it is a useful
illustration of how important context is when data and process
are seemingly 'violated' in the eyes of corporate power
structures. Elon Musk of Tesla was offended when a New
York Times reporter gave Tesla's car a less than stellar review
[9]. In this case, the reporter had free will for the drive and
was free to write up the review. However, when the review
was unfavorable, Musk turned the tables, saying that Tesla had
recorded the drive and that they would release the reviewer's
drive details, because all of the data of the drive proved that
the reporter did not drive the vehicle correctly (violated the
unknown process script) and "drove in circles for over half a
mile in a tiny 100-space parking lot" [10]. As the story broke,
many noted technologists called for the reporter to be fired.
The reporter wrote a thoughtful reply in more depth about his
drive; it was colder than usual, he couldn't find the high
powered charging outlets and didn't know he could use the
slower ones, he didn't know he had to charge the battery to full
instead of just topping it up a bit to keep going, etc. [11].
Proper context made a great change in the story evoked by
analysis of the data and criteria for review. For surveillance
data the context must be accurate and present, in order for the
'true' story to emerge. The situation we are seeing now is that
stories are being created from data, with insufficient regard to
the full context; data analysis and its subsequent story is used
as the 'best' fit to the context expected and desired by the
organization, not the context which is in place. If the reporter
had had a script to follow, or was aware of one, in the way that
Musk intended, perhaps he could have averted some of the
problems he found on the road. In the absence of having
specific instructions that addressed the reporter's unique
driving experience and context, even the Tesla employees
gave him differing opinions and instructions on how to
troubleshoot the problems he'd encountered whilst driving.
Agency is the capacity to make and execute nondeterministic
choices intended to advance to a goal as events unfold. For
example, humans exercise agency when deciding whether to
move towards a light or to walk across a street to avoid a
possible obstacle. Agency implies that people's future choices
are not intrinsically fixed or stochastically predictable except
on the basis of secondary principles of reasoning, such as
rationality, cooperation or enmity.
Choices in this context are not generally preset or fully
specified, but are relative to the goals of a person. That is,
people are likely to dynamically construct their choices as they
advance towards their goals, adapting to changes in
circumstances and their progress as they do so.
The complexity of any situation with more than one person
(exercising agency) increases rapidly, often requiring novel
choices on the part of all people whether they are competing
or cooperating since no person is solely responsible for most
changes. The capacity for anticipating other people's point of
view and what goals they may have and actions they may take
is critical. People not only have personal points of view, but
they must often be able to anticipate other people's points of
view to make progress towards their goals and to instantiate
their intentions in a manner that works.
The foundation of any social relationship requires a mutual
presumption of agency on the part of the other; social relations
require that each party assumes the other has some level of
agency [12]. In an automation scenario, such as executing
some type of corporate script, both worker and corporation
forfeit some or most of their agency to the script.
Pre-Pub Draft > CITE Publication to be available through (ISTAS’13) also
All material © 2013 Sally A. Applin and Michael D. Fischer. All rights reserved.
Pre-Pub Draft> CITE Publication to be available through (ISTAS’13) also
Agency is moderated by social settings, but always with the
knowledge that there are limits to this moderation if an
individual is abused or over-exploited in the setting. Social
interactions are moderated by cooperation and power, and
both have limits.
Agency is also limited in corporate settings. How agency is
conceptualized and instantiated in contemporary large
organizations is in part, structured by the type of surveillance
put upon the worker at their particular level of participation. In
the performance of duties, agency is assumed to be mainly at
the top, and decreases as one moves to the lower orders of the
organization, usually the place where workers are required to
execute scripts. As scripts are invoked, individuals have less
and less overt agency (a kind of cultural version of the
stigmergy that organizes ants). Perhaps less recognition of
others' agency emerges in corporations because most people in
the corporate chain seem to have so little. In this conception,
workers who are following scripts have little agency, but are
able to express some solely through ineptitude or failure to
follow instructions. Alternatively, people cloak their agency
by appearing to follow script instructions as a cover for other
activities that achieve the desired ends in an unauthorized
manner. This organizational design is further undermined by
the heterogeneity of staff from different parts of the world
with different understandings and languages (remember our
immigrant Tesco workers). However, the design of the system
is set up to largely deny any attribution of individual agency
by most of the underling workers.
In the case of scripts automating a process in the workplace,
the agency of people is formally limited. If the scripted
process is flawed, more ingenuity and 'disguised' agency is
required on the part of workers to make the deficient system
operate because their experiences are channeled into narrower
and narrower processes. This is due to the design of the
original scripts, the limitations of the computers monitoring
them and the software for analyzing the data about their
The last principle of process design is that, unfortunately, the
outcome of the first two principles is often used to design the
next set of processes. This is the critical mistake, as the scope
of human agency and the range in which we can cooperate
becomes more constricted, the processes required for success
become more fragile, and require more and more stealth-
agency on the part of workers to maintain. In addition, as
earlier generations of workers who developed a capacity for
such adaptation retire, increasingly the workers in their place
may come to accept the organizational design and may either
not care to 'rescue' organizational processes, or not be inclined
to do so
Society has become increasingly heterogenous, comprised of
people from many points-of-view using a range of channels
for communication in multiple languages on multiple devices
running multiple apps, in multiple time zones [13]. PolySocial
Reality (PoSR) emerges as people adapt to the pressures of
having to manage multiplexed communication between
themselves and others, and themselves and machines.
Machines are not exempt. The more surveillance that is
required from a machine, the more communications that
machine must engage in with other machines. PoSR also
arises in those circumstances.
PoSR is a conceptual model of the global interaction context
that emerges when people and/or machines communicate.
Each individual can exchange information with several
different networks of individuals, and thus each individual has
different access to information overall and thus a unique
viewpoint. The outcome of the aggregate of these partially-
intersecting networks and viewpoints is a multi-perspective
meta-network that constitutes a complex environment for the
coordination and collaboration of individuals. This becomes
even more complex with use of the social mobile web and
other new forms of communication that contribute
significantly to instantiating intentions [14,15,12].
Conceptualizing people as embedded in a PoSR context has
enabled us to make various observations: People are using
asynchronicity as an adaptive strategy to being time
compressed and having to do things online; people are
multiplexing messages and tasks, and there is wild variation in
whether or not there is temporal overlap between their
messages. Sometimes the messages get through and
sometimes not [16]. Fundamentally, we're concerned with
cooperation and collaboration between people, for without it,
we cannot survive.
For corporations that manage a global workforce situated in a
larger scale PoSR context, an increase in heterogeneity can
drive the evolution of process constriction (in the form of
scripts or other types of surveillance) as an adaptive strategy.
Because corporations have such a large range of diversity to
manage, processes that are linear, simple and clear appear to
help corporations control their environments and workers.
Surveillance, tracking and scripts are automated ways for them
to invoke a type of quality management or quality control.
Unfortunately, the quality lens is often turned onto their
'process' and not towards the real-world processes that drive
the outcomes of the service or work they are in the business of
providing. This is the twist: surveillance of processes mostly
reflect back on process and worker compliance and not on
problem solving, product creation or other things that the
corporations are in business for in the first place.
Companies appear to be using surveillance to both improve
performance with little regard for the human being acting
within their 'cyborg process,' and to collect data to use to
develop these processes even further. Thus they continue to
create the conditions that encourage even further commitment
to this approach.
Pre-Pub Draft > CITE Publication to be available through (ISTAS’13) also
All material © 2013 Sally A. Applin and Michael D. Fischer. All rights reserved.
Pre-Pub Draft> CITE Publication to be available through (ISTAS’13) also
The problem is that the difference between the capabilities of
a human being and those of a mechanized process is usually
quite vast. Furthermore, the intelligence that is inferred from
behavioral data usually lacks sufficient context to create a
process that can deal with the range of possible contexts that
can occur. The only solution is to attempt to reduce the range
of heterogenous contexts as well as the people acting within
these. Companies are trying to fit worker data to a process
model, rather than using the data from the workers to shape
the model and processes. The keystone being optimization of
By changing the way that they treat surveillance, corporations
might optimize their processes in a manner that benefits both
themselves and their workforce. This would involve two-way
surveillance. The former Tesco staff member said, "The guys
who made the scores were sweating buckets and throwing
stuff around the place” [8]. If they were "throwing stuff
around the place," perhaps it was inhibiting the performance
of other workers who had to clean it up, or walk around it, or
step over it or were unable to access the items they needed
because things weren't in their place anymore, or any one of
dozens of potential scenarios. If someone is wearing a device
that is measuring their output, but their workplace is not
comfortable or conducive for the success of their working
tasks, in that context the only measurement that Tesco will get
back will be that of a 'slower worker.' Having a process for
workers to report back what is wrongto exercise agency and,
more importantly, time to do that, might increase their
productivity in the long run. This approach involves changing
a management structure that would redistribute the power of
controlling at least some of the surveillance (of the work
process) to the workers.
This, in part, is what Toyota did with "The Toyota Way." The
Toyota method of manufacturing involves teaching workers
the current 'Best Practices' for manufacturing and employees
are expected to use these practices in their daily work.
Workers are measured, but the measurement is about Toyota's
process improvement, not necessarily the workers ability to
adhere to the process. Workers are encouraged to have their
own ideas about how to improve the process and to suggest
them, but not to use them until they are accepted into the main
process. When a worker's idea is accepted into the 'Best
Practices,' all workers benefit and begin to use the idea [17].
Toyota makes having an efficient process their priority and
encourages the workers to be guardians, stakeholders, and
entrepreneurs in their vision. Workers in Toyota have the
authority to stop the line if there is an error and other workers
are encouraged to help with the stopped line issue to benefit
the organization. In this way, Toyota encourages collaboration
between workers, workers and the process, and workers and
On the whole, surveillance is not a bad thing. It does help to
understand process, if one is able to account for what is going
wrong and correct it. It goes wrong, when the extension of
manufacturing machines is expected of humans, who have
their own creativity, ingenuity and agency to solve problems
and complete tasks.
We've suggested that agency in many organizations is limited
for those in the lower rungs of the hierarchy and that when
workers in these organizations do have agency, it is limited in
scope. Things that may appear to be surveillance in an
organization to a worker, might, if viewed through a different
lens be a less well-thought-out management task with serious
effects on staff that aren't consciously considered. If we gave
the benefit of the doubt to Tesco, we could ask if whether or
not their attempts to gather data on behavior and gamify the
process is just a method to increase output while giving
themselves feedback about their processes in real time. This
might place it into a manufacturing category where the
workers are considered to be individual cyborg processes and
are optimized as Tesco might optimize their other machines.
Perhaps Tesco is trying to squeeze every second out of
workers to improve their company performance.
This last point is important, for the differences in worker
agency by being forced to wear devices in processes that are
so sensitive to linear time, can have the result that they lose
most of the time they might have. This includes anything that
could benefit from their agency either within or outside the
scope of their jobs. In particular, detailed process requirements
limit the possibility of workers employing asynchronous
means of balancing work requirements to achieve a better mix
of outcomes. In a time when workers are gaining more
experience at being asynchronous in their lives outside the
workplace, [16] this limits the application of these new skills
in the workplace.
Even the consumer can become entwined in the corporation's
'script process' as a 'required feedback loop,' and not
necessarily a human customer. At the Starwood chain of
hotels, when a service is performed after a request such as
asking for more towels or a some other service, the staff is
instructed to call the hotel room and ask for feedback about
the process in the form of something such as, "Did you get the
towels ok? Was everything to your satisfaction?" [18]. This is
a form of quality management that puts the feedback of
process of Starwood well ahead in priority to Starwood than of
the time of the person in the hotel room, who had to make the
initial request, then wait for the disruption to have the request
fulfilled, then to be disrupted further by a call, named the
'Service Recovery,' asking for an evaluation of how the
process did or did not address the initial request. It doesn't stop
there, post-checkout feedback forms, Twitter requests, email
surveys are all continuing demands for process feedback. The
entire stay then becomes a vehicle for the processes of
Starwood evaluating its processes.
From a management perspective, with a group of
heterogenous, diverse workers, devising a 'common' language
Pre-Pub Draft > CITE Publication to be available through (ISTAS’13) also
All material © 2013 Sally A. Applin and Michael D. Fischer. All rights reserved.
Pre-Pub Draft> CITE Publication to be available through (ISTAS’13) also
such as a process that is surveyed seems a smart way to keep
the workers focused. Unfortunately, it also limits their agency
and the surveillance techniques only further isolate and
reinforce their heterogeneity. The corporations see the workers
as a homogenous group focused on a task, but the workers
remain diverse and are executing a program that does not
work for each individual.
One possible scenario is to avoid the plain 'transparency'
argument of companies. It isn't just about the fact that they do
this, its also about the effect this has on their workers as
human beings and members of society. What is challenging
for both workers and companies is the cohort problem: with
more processes in place, workers are becoming less proficient
in applying agency to solve problems. Corporations have to
adapt to this situation that they in part created, by creating
more detailed processes. This is twofold, the corporations
continue to lack a workforce that knows how to solve
problems 'in the wild' and they also lack increasingly any
mentors for workers to learn from, as the older, more educated
agency enabled worker cohorts are either laid off or retire.
There must be a balance between quality control and quality
of life. Allowing for agency and bi-directional surveillance
might address this.
Thing Theory [13] is an agent based network model that
utilizes knowledge of a local environment, gained via its
relationship to both people in the environment and to system-
subagents and meta-agents in other systems, who inform the
Thing-agent of needed knowledge. Thing Theory is loosely
informed in part on the notion of the character "Thing" from
"The Addams Family," a 1960's television show based on a
comic of the same name created by Charles Addams. Thing is
characterized as a disembodied hand (and forearm) that has
been with the family for many years and is described as both a
'family retainer' and 'friend'. It inhabits a series of tabletop
boxes in different rooms of the house [19] that could be
compared to a type of roughly cobbled physical network.
Thing communicates with the family by gestures, sign
language, writing out notes, or tapping out messages in Morse
code. Thing serves the family by accessing a portal in
contextual proximity to what is needed or desired at the
precise moment required, in the precise room or context
needed. Thing is not only a ubiquitous agent, but an
anticipatory one that migrates within the environment. The
sensing, response and location-awareness of Thing is a useful
aspirational model for a network agent in a location-aware
Smart Environment.
Our Thing-agent functions as a communal surveillance agent.
It develops relationships and shapes knowledge and responses
for different actors within the system that it serves. The Thing-
agent collects information on all systems, and could work as a
model for our two way (or multi-way) surveillance. Not only
could management track workers, but workers could track
management to see when management needs their project
completed so they can more effectively prioritize their
time. A Thing-agent can serve a real need to revise and design
processes so that they are not so brittle, fragile, or
If we examine the Toyota model, Toyota keeps track of their
production in terms of subsystems, rather than individual
people. The people function within groups that are part of
interlinking processes that serve the goal of producing
Toyota's products. This model of viewing groups of people as
interlinking processes between subsystems gives people some
space to use their talents and invoke knowledge of problem
solving as it pertains to the tasks that they do.
In [13] we discuss possible approaches for implementing
Thing theory in location-aware environments. A Thing-agent
at a minimum is a means to inform users of the information
about processes within a smart environment in pragmatic
terms that make sense to those users. Ideally with a high level
interactive interface that adapts to the POV of each user.
A Thing-agent will include a representation of the pragmatic
contexts in which processes are expressed. This is done such
that the Thing-agent facilitates choices by user, rather than
forcing them into specific choices.
In such a situation there are many complex contingencies that
can arise. A Thing-agent might employ non-monotonic multi-
agent simulations incorporating specifications for each process
and that relates sensor-based information to present scenarios
of possible ways processes and users might interact with each
other. On this basis a Thing-agent can offer available choices
to the user in different contexts, and provide feedback in terms
of what is likely to occur should the user make a given choice,
thus a basis for 'fine-tuning' how they proceed.
There are existing logics and associated software architectures
for developing such multi-agent simulations in a manner to
support decision making. Deontic logic is a good candidate for
a useful semantics that supports constructing simulations of
the type needed [20, 21]. Casto and Maibaum [22] present a
deontic logic suitable for representing the interrelations of
users with agency, and [23] discusses at length representing
and reasoning with agency in deontic logic.
A Thing-agent mediated multi-agent simulation can be useful
for designing new management/worker cooperation and
knowledge exchange. Such an exchange could leverage
'simulation services' provided by a Thing-agent to support
their decision making, and possibly improve the joint
understanding of dynamic production environments. These
simulation services would direct attention once again to
outcomes rather than processes, since all parties could see
different possibilities emerging from their actions in
interaction with others. They could then alter course as needed
to improve probabilities of a good outcome dynamically. This
approach to process control would be far more adaptive and
would favor outcomes over processes while leaving room for
workers to deploy their individual skills towards achieving
these outcomes. Exploiting a Thing-agent to ensure that the
information that needs to be shared to support agency in the
Pre-Pub Draft > CITE Publication to be available through (ISTAS’13) also
All material © 2013 Sally A. Applin and Michael D. Fischer. All rights reserved.
Pre-Pub Draft> CITE Publication to be available through (ISTAS’13) also
network of workers and managers is available to maximize the
capacity for cooperation, and avoid the problems that can arise
from PoSR networks such as missed or misunderstood
messages, or connected individuation [24]. In a workplace
environment, Thing-agents could be employed within a
system of restricted privacy based on tasks, rather than
tracking people's behavior.
Not everyone in the social environment of work fully knows
the roles and responsibilities of what it is like to do each
other's jobs. A Thing-agent type model could assist the
process. In this way, two or multi-way surveillance may
become simply respect and cooperation to complete tasks
rather than a measuring stick that erodes the agency of humans
within organizations.
1. N. Greenfieldboyce, "Indie truckers; keep big brother out of my cab,"
NPR, April 20, 2011. [Online]. Available:
brother-out-of-my-cab [Accessed: 27 Feb. 2013].
2. SpeedGauge. "Driver behavior & speed control for fleets in North
America, Europe & Australasia," Available online: [Accessed: 27 Feb. 2013].
3. Federal Motor Carrier Safety Administration. Department of
Transportation, "Electronic on-board recorders for hours-of-service
compliance; Final Rule," Federal Register, Rules and Regulations, vol.
75:64, pp. 17207-17252. [Online]. Available:
HOS-Compliance.aspx [Accessed: 27 Feb. 2013].
4. I. Rubinstein and N.Good, "Privacy by design: a counterfactual analysis
of Google and Facebook privacy incidents," Berkeley Technology Law
Journal, NYU School of Law, Public Law Research Paper No. 12-43,
August, 2012. [Online]. Available: or [Accessed: 27 Feb. 2013].
5. Motorola Solutions, "WT4000 wearable terminal," 2013. [Online].
EN,XF-EN [Accessed: 27 Feb. 2013].
6. Motorola, "With picking errors down and productivity upBen E.
Keith is seeing a brighter future," Case Study, 2011, p. 2. [Online].
torola_benkeith_sae_10443151.pdf [Accessed: 27 Feb. 2013].
7. H. Baumann, "Order picking supported by mobile computing," Doctoral
Dissertation. University of Bremen, 2013. Available: [Accessed 27 Feb.
8. K. Rawlinson, "Tesco accused of using electronic armbands to monitor
its staff," UK Independent. [Online]. Available:
using-electronic-armbands-to-monitor-its-staff-8493952.html [Accessed:
27 Feb. 2013].
9. J. M. Broder, "Stalled out on Tesla's electronic highway," The New
York Times, Feb. 10, 2013. [Online]. Available:
highway.html?ref=johnmbroder&_r=0 [Accessed: 27 Feb. 2013].
10. E. Musk, "A most peculiar test drive," Tesla Motors, Feb. 13, 2013.
[Online]. Available:
test-drive Accessed: [Accessed: 27 Feb. 2013].
11. J.M. Broder, "That Tesla data: what it says and what It doesn't," The
New York Times, Feb. 14, 2013. [Online]. Available:
says-and-what-it-doesnt/?ref=johnmbroder [Accessed: 27 Feb. 2013].
12. S.A. Applin and M.D. Fischer, "Applied agency: resolving multiplexed
communication in automobiles," in Adj. Proc. AutomotiveUI '12, In
cooperation with ACM SIGCHI, 2012, pp. 159-163.
13. S.A. Applin and M.D. Fischer, "Thing theory: connecting people to
location-aware smart environments," In Proc. LAMDa’13, DFKI, March
14. S.A. Applin and M.D. Fischer, "A cultural perspective on mixed, dual
and blended reality. In Proc. LAMDa’11, DFKI, (2011), pp. 14.
15. S.A. Applin and M.D. Fischer, "PolySocial Reality: prospects for
extending user capabilities beyond mixed, dual and blended reality." In
Proc. LAMDa’12, DFKI, Feb. 2012, pp. 1-4.
16. S.A. Applin, M.D. Fischer and G. Becker, "PolySocial Reality and the
enspirited world." Presentation at SXSW 2012, Austin, Texas, USA,
Mar. 2012.
17. J.K. Liker, The Toyota Way: Fourteen Management Secrets from the
World's Greatest Manufacturer, McGraw-Hill, 2004.
18. S.A. Applin, Personal communication with Starwood Hotels. Feb.,
19. The Addams Family, Lurch Learns to Dance S01E13, (06:31). [Online].
Available: from YouTube via
Anabrans., [Accessed:
22 Jan. 2013].
20. M. Fischer, "Cultural agents: a community of minds," Engineering
Societies in the Agents World VI. Dikenelli, O., Gleizes, M. and Ricci,
A. (eds). Lecture Notes in Computer Science 3963, Springer-Verlag,
2006, pp. 259-274.
21. M Ryan, J Fiadeiro, T Maibaum, "Sharing actions and attributes in
modal action logic." Theoretical Aspects of Computer Software. 1991,
pp. 569-593.
22. P. Castro, and T. Maibaum, "A complete and compact propositional
deontic logic." Theoretical Aspects of ComputingICTAC. Springer
Berlin/Heidelberg. 2007, pp. 109-123
23. J. Horty, Agency and Deontic Logic. Oxford University Press, Oxford,
24. S.A. Applin and M.D. Fischer, Connected Individuation. 2011. [Online].
Available: from
[Accessed: 22 May, 2013].
... As long as there are pedestrians or other VRUs in the system, vehicles that are not automated, and computer programs with errors, the cooperation of those participating in traffic will be subject to the agency of humans and the errors of automation. Even in the most fail-safe and secure systems, human creativity will find ways to intentionally discover workarounds within systems to meet their own needs, while machines will run programs that have unexpected outcomes when interpreted by humans and contribute to conflict [4], [5]. ...
... This would create an unbalanced system by which human drivers would have cognitive load increased as they take on the additional labor of "babysitting" semiautomated and automated vehicles on the roads, enabling those manufacturers to claim successes that would be due, in part, to human drivers enabling their success [3], [6], [20]. This could lead to accidents in situations when humans selectively (rather than always) assist automation on the road by yielding [4], [5]. ...
For more than a century, driver?vehicle interaction research has been mainly focused on interfaces that assist drivers as they interact with their vehicles with very limited attention paid to interacting with people and their agents moving within the surrounding commons. With the emergence of semiautomated and self-driving vehicles in combination with mobile phones, internal vehicular interfaces will gradually transition to (nondriving) passengers and their vehicles. These newly transitioning interfaces share (or will require) similar specifications to the ones currently used to control many consumer electronics. As mobile devices continue to grow in usage, people are creating secondary interface experiences that are related to but may not be attached to vehicles. It is suggested that future automotive IT infrastructure and interaction specifications should follow from general human?computer interaction guidelines, with the understanding that there will be a high heterogeneity to manage in these contexts. A critical point of vehicle interface is between self-driving cars, manually steered vehicles, and other, so-called vulnerable road users (VRUs), whose attention may also be divided between the road and their mobile phones. When humans, as passengers in semiautomated or automated vehicles, are no longer directly responsible for human?human/interpersonal communication and negotiation, new interactive paradigms must be considered and designed to protect people in manually controlled cars and VRUs within shared road space. Thus, automotive user interface research must also formally extend outward to include intervehicular interaction rather than being limited to the mainly intravehicular environment, as is the present case. In this article, we will examine broader issues of user interface (UI) consistency in vehicular contexts and present recommendations for future automotive UI research.
... A definition of agency based on the notion of non-deterministic behaviours [1] fails to recognise the increasing variety and complexity of human-machine networks 1 (HMNs) [2], the intention of technology designers [3], and active intervention by bots within social networks [4,5]. The concept of agency is particularly problematic in human-machine interactions [6]. ...
Conference Paper
Human-machine networks affect many aspects of our lives: from sharing experiences with family and friends, knowledge creation and distance learning, and managing utility bills or providing feedback on retail items, to more specialised networks providing decision support to human operators and the delivery of health care via a network of clinicians, family, friends, and both physical and virtual social robots. Such networks rely on increasingly sophisticated machine algorithms, e.g., to recommend friends or purchases, to track our online activities in order to optimise the services available, and assessing risk to help maintain or even enhance people's health. Users are being offered ever increasing power and reach through these networks by machines which have to support and allow users to be able to achieve goals such as maintaining contact, making better decisions, and monitoring their health. As such, this comes down to a synergy between human and machine agency in which one is dependent in complex ways on the other. With that agency questions arise about trust, risk and regulation, as well as social influence and potential for computer-mediated self-efficacy. In this paper, we explore these constructs and their relationships and present a model based on review of the literature which seeks to identify the various dependencies between them.
... Such fears reflect a wider trend of increasing distrust of businesses, governments, and organizations, which is fueled by publicized high-profile leaks of data abuse such as wiretapping by the U.S. NSA. Employers may begin to utilize bodily sensor networks to facilitate employee monitoring, benchmarking, and performance relate benefits [58]. ...
Full-text available
There is a long history of medically-prescribed implantable devices used to assist in treating chronic health conditions. Examples include pacemakers [1], cochlear and retinal implants [2], insulin pumps [3], and deep brain stimulation implants for relief of Parkinson's disease tremors and seizures [4] to name but a few. There is also a growing trend towards wearable consumer electronics for a myriad of applications, including interactive haptic environments [5], healthcare [6], data communications [7], wearable interfaces [8], and people tracking [9]. Artificial Intelligence (AI) continues to make remarkable progress in controlled settings and promises to bring intelligence to everything we interact with [10].
... Such fears reflect a wider trend of increasing distrust of businesses, governments, and organizations, which is fueled by publicized high-profile leaks of data abuse such as wiretapping by the U.S. NSA. Employers may begin to utilize bodily sensor networks to facilitate employee monitoring, benchmarking, and performance relate benefits [58]. ...
Information and communication technologies (ICTs) play a vital role in the dissemination of information to farmers. ICT tools are also essential for harnessing and capturing local farmer knowledge. Innovative use of ICT tools recognizes the importance of two-way communication, and ICT can be used more efficiently to achieve the goals of agricultural extension. Beyond connectivity ICT offers security and mobility to owners, and requires that users have basic literacy. In addition to voice communication ICT tools allow for transfer of data, which can be used in context for applications and in agriculture. New ways of managing knowledge have emerged across developing countries beyond the traditional farmer-extension systems.
... Companies may use sensor proximity in public places to datamine preferences, etc. and use this information to bombard individuals with personalized advertising, as is currently commonplace on the internet with constant consumer pop-ups and adverts. Employers may begin to utilize bodily sensor networks to facilitate employee monitoring, benchmarking and performance relate benefits [50]. ...
Conference Paper
Full-text available
This paper presents a SWOT analysis for the emerging and futuristic field of non-medical body-implantable devices. This area will begin to materialize as one of the next big themes in future personal computing and offers huge rewards to society if implemented correctly. The technology boasts many strengths which are applicable to a variety markets including entertainment, social networking, personal safety, security, consumerism, communications, healthcare, convenience and human body enhancement. Such subcutaneous sensor technology releases citizens from the multitude of portable computing devices, keys, wallets, passes, etc. However, the technology would be a target for hackers and would likely result in more violent robberies and forceful ID removal. Additionally, adverse health effects, device and battery safety and reliability, and co-existence with medically prescribed implants are issues developers must solve before the technology could excel. External emerging technologies such as Cloud computing, IoT, and NFC support development and potential success of implantable systems and combines to help address issues of personal safety, terrorism, people tracking and identification, e-payments, and long-term fitness profiling. Threats to the technology's uptake include societal fears on such aspects as adverse health effects, dehumanisation, breaches of human rights, conservatism, social privacy, and religious objections. With this technology potentially beginning to enter the mainstream in the next 5–10 years considerable effort is required to develop legislation, policies, procedures, device and network security, and convince the general public this technology is the next logical step in personal computing.
... The agency (e.g. freedom to make choices from DRAFT DRAFT DRAFT available options) of the remaining human staff becomes increasingly suppressed as corporate scripts increasingly direct these employees to behave more like the machines [10]. A Starbucks card automatic refill is a 'magical' idea until the algorithm transfers funds to a third party. ...
Conference Paper
Full-text available
We examine the current state of 'mixed-use' new technologies integration with legacy systems, and whether the human assistance required to complete tasks and processes could function as a training ground for future smart systems, or whether increasing 'co-dependence with' or 'training of' algorithmic systems, enhancing task completion and inadvertently educating systems in human behaviour and intelligence, will simply subsume people into the algorithmic landscape.
Companies increasingly equip employees with smart watches to, e.g., support them in carrying out their work. Smart watches can however collect data about them and reveal sensitive information. This may result in limiting the acceptance of these devices by employees, despite their potential helpfulness. In this paper, we therefore analyze factors that influence employees’ willingness to share smart watch captured private data. In more detail, we investigate employees’ technological knowledge about data collection and processing and the associated risks, their technical affinity, their smart watch ownership and usage, and their legislation knowledge about respective laws. To this end, we have conducted an online survey with more than 1,000 full-time employees. Our findings suggest that employees are aware of the risk associated with smart watches but partially have incorrect knowledge about legal frameworks. Moreover, more than one-third of the participants own a personal smart watch and have a certain technological affinity. However, our results reveal different impacts from these factors on employees’ willingness to share data with their employers.
Nearly every week, a technology company is introducing a new surveillance technology, varying from applying facial recognition to observing and cataloguing behaviours of the public in the Commons and private spaces, to listening and recording what we say, or mapping what we do, where we go, and who we're with—or as much of these facets of our lives as can be accessed. As such, the general public writ-large has had to wrestle with the colonization of publicly funded space, and the outcomes to each of our personal lives as a result of the massive harvesting and storing of our data, and the potential machine learning and processing applied to that data. Facebook, once content to harvest our data through its website, cookies, and apps on mobile phones and computers, has now planned to follow us more deeply into the Commons by developing new mapping technology combined with smart camera equipped Augmented Reality (AR) eyeglasses, that will track, render and record the Commons—and us with it. The resulting data will privately benefit Facebook's continued goal to expand its worldwide reach and growth. In this paper, we examine the ethical implications of Facebook's Project Aria research pilot through the perspectives of Responsible Innovation, comparing both existing understandings of Responsible Research and Innovation and Facebook's own Responsible Innovation Principles; we contextualise Project Aria within the Commons through applying current social multi-dimensional communications theory to understand the extensive socio-technological implications of Project Aria within society and culture; and we address the potentially serious consequences of the Facebook Project Aria experiment, inspiring countless other companies to shift their focus to compete with Project Aria, or beat it to the consumer marketplace.
Conference Paper
Full-text available
The technology industry has evolved over the years with a development lens increasingly focused on end users and usage cases. Indeed, for the past decade or more, personas (the designer-created profiles of end users) have become stand-ins for various usage cases and user models. With regard to location aware software and mobile applications, the usage of Dual Reality and Mixed Reality as metaphors have functioned in a similar vein. Just as personas are not people, Mixed and Dual Reality do not fully represent or address the complex usage cases developing as more peo- ple do more things, with more software at more times and in more spaces than ever before. This new complex appli- cation ecosystem presents greater opportunities and chal- lenges for application design. We discuss ways that devel- opers can use PolySocial Reality (PoSR) to represent a more complete complex structural model of individuals interacting within multiple environments.
Conference Paper
Full-text available
This paper is an anthropological perspective on the impact of Dual, Mixed Reality and 'PolySocial Reality' (PoSR) on Location Awareness and other applications in Smart Environments. We intend to initiate a friendly inter- disciplinary discussion on the interaction aspects and cultural implications surrounding these new forms of integrated technologies. Anthropologists examine human group behavior in the context of culture. Because the LAMDa workshop is addressing 'group behavior,' we will contribute to the workshop with our understanding of humans, culture and group behavior and to learn from others what type of group behavior is expected as new technologies and their subsequent experiences are created for human use. When we discuss human group behavior, we refer to the definition of a social group: a collection of humans who repeatedly interact within a system. Humans can, and do switch context between environments and blend traces of one into the other in a socially unconscious manner often seemingly simultaneously. We propose that the cultures and behaviors of humans are increasingly actively permeating Internet and network- based applications, as well as those that are geolocal. With the future Internet of things, Dual Reality and Mixed Reality, the opportunity for humans to extend their own blended reality, as well as to create new ones is unfolding. That said, because humans interact within groups, the multiplexing of their mutual blended realities rapidly creates a PoSR. Sorting out the relationships between realities as well as between synchronous and asynchronous time and geolocal space can create a condition where realities are simultaneous and the idea of 'x' can be perceived as equaling 'not x.' We explore this new type of interoperability between virtual and physical, ideational and material, representations and objects and culture.
Conference Paper
Full-text available
As architects, entrepreneurs, hoteliers, healthcare profes- sionals and others embrace the Internet of Things (IoT) and the Smart Environment paradigms, developers will bear the brunt of constructing the IT relationships within these, making sense of the big data produced as a result, and managing the relationships between people and technolo- gies. We explore how PolySocial Reality (PoSR), a frame- work for representing how people, devices and communi- cation technologies interrelate, can be applied to develop- ing use cases within integrated IoT and Smart Environment paradigms, giving special consideration to the nature of location-aware messaging from sensors, and the resultant data collection. Based on this discussion, we suggest ways to enable more robust messaging, and eliminate redundant messages by enlisting a social awareness of software agents applied in carefully considered contexts.
Conference Paper
Full-text available
Distributed systems may be specified in Structured Modal Action Logic by decomposing them into agents which interact by sharing attributes (memory) as well as actions. In the formalism we describe, specification texts denote theories, and theories denote the set of semantic structures which satisfy them. The semantic structures are Kripke models, as is usual for modal logic. The possible worlds in a Kripke model are the states of the agent, and there is a separate relation on the set of states for each action term. Agents potentially share actions as well as attributes in a way controlled by locality annotations in the specification texts. These become locality axioms in the logical theories the texts denote. These locality axioms provide a refined way of circumscribing the effects of actions. Safety and liveness conditions are expressed (implicitly) by deontic axioms, which impose obligations and deny permissions on actions. We show that denotic defaults exist so that the specifier need not explicitly grant permissions or avoid obligations in situations where normative behaviour is not an issue.
In this dissertation I present the results of a newly developed mobile computing solution with reasonable investment costs that supports the picking process in a high density picking environment with multiple orders. The developed solution is presented on a head-mounted display (HMD). It has a graphical user interface that displays graphical representations of the shelves to pick from. Results show that in a high density picking environment, this solution is faster than paper-pick lists and pick-by-voice and virtually eliminates errors. Using color helps to identify the correct row and some evidence suggests that symbols and partial images as well as context feedback can further improve the error rate. Testing on an assembly line of an automobile manufacturer where normally pick-by-light was used showed some difficulty in user acceptance for HMDs. A tablet-PC mounted on the pick cart was well accepted in this study and may provide similar benefits and performance.
Conference Paper
We explore the transition to greater sociability concerning automobiles in the context of communications being more fully integrated into their systems.In particular, we examine the consequences of the emergent network that results from interaction between cars, drivers, passengers, pedestrians and locations.
Regulators here and abroad have embraced “privacy by design” as a critical element of their ongoing revision of current privacy laws. The underlying idea is to “build in” privacy (in the form of Fair Information Practices or FIPs) when creating software products and services. But FIPs are not self-executing. Rather, privacy by design requires the translation of FIPs into engineering and usability principles and practices. The best way to ensure that software includes the broad goals of privacy as described in the FIPs and any related corporate privacy guidelines is by including it in the definition of software “requirements.” And a main component of making a specification or requirement for software design is to make it concrete, specific and preferably associated with a metric. Equally important is developing software interfaces and other visual elements that are focused around end-user goals, needs, wants and constraints. The Article offers the first comprehensive analysis of engineering and usability principles specifically relevant to privacy. Based on the relevant technical literature, it derives a small number of relevant principles and illustrates them by reference to ten recent privacy incidents involving Google and Facebook. The Article concludes that all ten privacy incidents might have been avoided by the application of these privacy engineering and usability principles. Further, we suggest that the main challenge to effective privacy by design is not the lack of design guidelines. Rather, it is that business concerns often compete with and overshadow privacy concerns. Hence the solution lies in providing firms with much clearer guidance about applicable design principles and how best to incorporate them into their software development processes. Greater guidance is also needed for how to balance privacy with business interests, and there must be oversight mechanisms as well.
Develops deontic logic against the background of a rigorous theory of agency in branching, or indeterministic, time. It is often assumed that the notion of what an agent ought to do can be identified with that of what it ought to be that the agent does. The book provides a framework in which this assumption can be formulated precisely and shown to be mistaken. In its place, it offers an alternative account of what agents ought to do that relies on an analogy between action in indeterministic time and choice under uncertainty, as it is studied in decision theory. This alternative account is then related to issues involving conditional obligation, group obligation, act utilitarianism, and rule utilitarianism.
Conference Paper
In this paper we present a propositional deontic logic, with the goal of using it to specify fault-tolerant systems, and an axiomatization of it. We prove several results about this logic: completeness, soundness, compactness and decidability. The main technique used during the completeness proof is based on standard techniques for modal logics, but it has some new characteristics introduced for dealing with this logic. In addition, the logic provides several operators which appear useful for use in practice, in particular to model fault-tolerant systems and to reason about their fault tolerance properties.