No Malibu Surfer Left Behind:
Three Tales About Market Coercion
Åsbjørn Melkevik
Harvard University
ABSTRACT: This article examines the question of private coercion in market
societies, arguing for an unconditional basic income guarantee from a classical
liberal viewpoint. It proposes three main arguments. First, classical liberals view
the purpose of government to be the reduction of coercion, both public and private.
Second, a proper understanding of the nature of coercion indicates that parties subject to certain types of hardship are being coerced. Third, where the total amount
of coercion is reduced by eliminating the hardship, the classical liberal state must
do so as to fulfill its purpose. Hence, this article argues that if the total amount of
coercion in society can be reduced by the state employing the amount of coercion
necessary to maintain an unconditional basic income guarantee, then the classical
liberal state is obligated to maintain such a guarantee by its underlying justification.
KEY WORDS: basic income, classical liberalism, Friedrich Hayek, market capitalism,
private coercion, social safety net
I. A PRINCIPLED JUSTIFICATION FOR THE BASIC
INCOME GUARANTEE
T
he extent to which the classical liberal theory of freedom as defended, for
example, by Friedrich Hayek can lead us to some original conclusions about
social justice has not been sufficiently appreciated. This article remedies such a
shortcoming. John Tomasi, a leading figure of the neoclassical liberal movement,
recently said this about Hayek advocating for state coercion to support social
service programs: “Like many others before him, Hayek leaves us to guess at his
precise reasons for advocating these familiar features of the classical liberal political
program” (2012, 22). Similarly, Jason Brennan and Tomasi (2012, 119) wrote that
“while classical liberals such as F.A. Hayek and Milton Friedman advocate a social
safety net, their advocacy seems ad hoc given their strong claim about the sanctity
of property rights.” Although it is understood that classical liberalism allows for a
social safety net, say in the form of a basic income guarantee or a negative income
tax, it is thought that such a safety net is not necessary in a classical liberal polity,
and that its justification is somewhat lacking.
This article suggests an original response to such a problem, showing why an
unconditional basic income guarantee is not only allowed but also required by the
classical liberal theory of freedom. The classical liberal justification for the state
consists of the need to reduce the amount of coercion to which individuals are
subject. Such a concern for non-coercion can be used not only to limit the scope of
the modern welfare state, as is often the case, but also, perhaps more interestingly,
©2017 Business Ethics Quarterly 27:3 (July 2017). ISSN 1052-150X
DOI: 10.1017/beq.2017.12
pp. 335–351
336
Business Ethics Quarterly
to explain why social justice is necessary from a classical liberal viewpoint. More
precisely, this article will propose three main arguments. First, classical liberals
view the purpose of government to be the reduction of coercion, both public and
private. Second, a proper understanding of the nature of coercion indicates that
parties subject to certain types of hardship are being coerced. Third, where the total
amount of coercion is reduced by eliminating the hardship, the classical liberal state
must do so to fulfill its purpose.
After establishing (section II) that private relations in market settings can be
coercive, I will examine three classical liberal theories of coercion: first (section III),
a legal approach supported by Ludwig von Mises; second (section IV), an economic
approach endorsed by Richard Epstein; and, third (section V), an autonomy-based
approach championed by Friedrich Hayek. I will argue for the latter as the only
one with which we can satisfactorily deal with the problems of private coercion.
The classical liberal concern with coercion is thought to lead to a defense of the
minimal state. For such a conclusion to be sound, however, the state has to deal
with private coercion in the labour market. Along the way, I will relate three tales
to illustrate some problems of coercion in market societies, so as to arrive in the
end at a principled argument for a basic income guarantee.
It is true, as Alistair Macleod (2008) noted, that whether a given institutional
arrangement is coercive depends on whether the arrangement is morally defensible
in regard to both the content of its rules and the process by which they have come to
be adopted. For the sake of the argument, and this is not a controversial point from
a classical liberal standpoint, I assume that market institutions are indeed morally
acceptable. Hence, this article argues that if the total amount of coercion in society
can be reduced by the state employing the amount of coercion necessary to maintain a
social safety net, then the classical liberal state is obligated to maintain such a safety
net by its underlying justification. This will permit us to offer a principled rationale
for a classical liberal unconditional basic income, which is thought to be missing by
many neoclassical liberals, and which, indeed, was never explicitly given by, say,
Hayek or Friedman. Let us begin by identifying the problem of private coercion
accurately, which is not always so simple in view of the many contradictory theories
of coercion we can find in the liberal tradition.
II. PRIVATE COERCION IN MARKET SOCIETIES:
THE TALE OF SVEINN AND HARRI
“Capitalist theory typically assumes that market transactions are voluntary and
uncoerced,” Alan Wertheimer explained (1987, 4-5), “even if they are made against a
background of economic necessity.” If we are to take further steps to reconcile market capitalism with social justice, such an assumption must be challenged. Overall,
many classical liberal proponents of market institutions have endorsed a forgiving
understanding of voluntary exchanges, establishing a distinction between coercion
proper and circumstances that limit alternatives non-coercively (Nozick 1974, 262-63).
As Wertheimer noted, however, the viability of such accounts of coercion is somewhat unclear. It becomes even murkier once we examine how market institutions
No Malibu Surfer Left Behind
337
can contribute to many problems of private coercion. Paul Gowder (2014, 331-33),
for example, recently wrote about what he calls “unfree immiserating work,” say
jobs with extremely low pay, unusually long hours, unpredictable scheduling, and
a likelihood of long-term physical degeneration. Similarly, Chris Bertram, Corey
Robin, and Alex Gourevitch (2012) have shown how there is often a “systemic denial
of freedom in private regimes of power, particularly the workplace.” Employers
sometimes invade employees’ privacy, firing them for resisting such invasion, and
they also fire workers for supporting some political candidates, for trying to join
a union, and, what is even more bizarre, for drinking in the privacy of their own
homes. These are indeed problems of coercion, as this section will now establish,
and thus they must be addressed by the state.
The usual problem with coercion is straightforward: getting one to do something
by threatening one with some evil is generally inadequate in the liberal tradition.
The use of interpersonal violence by individuals is an obvious example of coercion,
say criminal activity, and yet we should note that coercion cannot be summed up
by such violence. According to Matt Zwolinski (2015, 523), for example, it is
“relatively uncontroversial that all systems of property rights are by their very nature
coercive.” Now, although I agree with Zwolinski, we must recognize that this point
is quite controversial for classical liberals. In fact, that the capitalistic organization
of society itself includes some inherently coercive features is often categorically
denied, which is one thing this article will address.
But we may want to take a step back. A limited state is justified in the classical
liberal tradition by the fact that it protects us from our fellow citizens. Simply put,
as Friedman explained (2002, 2), the major function of government is to preserve
law and order. Enforcing criminal law is coercive by nature, but since it reduces
the amount of coercion to which individuals are subject it is nonetheless warranted.
Talking about libertarians, Bertram, Robin, and Gourevitch (2012) rightly observed
that “in fact, the idea that private action can diminish individual freedom is central
to their justification for the state, which is that some state coercion is required
to stop people from dominating, enslaving, and generally harming others.” Market
capitalism is justified in quite the same way. Such a system counters the capacity
monopolies or cartels would otherwise have to exercise coercion – for example, in
withholding services that are essential to my existence (Hayek 1978, 136). This is
uncontroversial. It is not clear, however, that some horrible condition of employment
may lead to coercion in a market capitalist framework.
Let me immediately introduce a first story, the tale of Sveinn and Harri, which
captures one facet of the problem of private coercion in market societies. Harri
means “master” in Old Norse, and Sveinn means “free servant.” Let us accordingly
imagine that Harri is a tyrannical foreman, while Sveinn has to work for Harri all day
every day, not being able to find a better job. Sveinn’s job is repetitive, dangerous,
backbreaking, while Harri is disrespectful and is looking for any excuse for firing
Sveinn. Sveinn is paid the minimum wage, but the best he can hope is for such an
awful job, which he is forced to endure because of economic desperation. He could
go work for someone other than Harri, but it would be more of the same; for example,
mining, fruit-picking, textile sweatshops, and meatpacking factories. “Raw statistical
338
Business Ethics Quarterly
data,” rightly said Gowder (2014, 330), “cannot truly capture the lousiness of these
jobs.” The question, then, is whether Sveinn is coerced. Intuitively, the answer is
yes – economic desperation has led Sveinn to a life of servitude.
A first objection could go as follows: people like Sveinn are free to starve to
death, as Gerald Cohen observed, sarcastically. Such a response, however, is too
all-encompassing. It can defeat any understanding of coercion, since starving is
always an option. Yet the concept of coercion, as Cohen explained, does not imply
that one has no alternative whatsoever. “When I am forced to do something,” said
Cohen (1983, 4), “I have no reasonable or acceptable alternative course.” Sveinn has
no reasonable alternative to immiserating work, and therefore, we could say, he is
coerced. He could become a beggar, or go work for another tyrannical foreman,
but these are not reasonable options. Classical liberals, though, will resist such
an account of coercion. We cannot, as Karl Marx thought we should, “hunt in
the morning, fish in the afternoon, rear cattle in the evening, criticise after dinner”
(1998, I.ii.17, 53). Poor people have limited options in market societies, and
yet classical liberals like Hayek will nonetheless insist that they are not truly
coerced – or, at least, if they are coerced, it is not because their reasonable
alternatives are too few or nonexistent.
There is one exception, which Hayek briefly mentioned, although it is a crucial
one. Periods of extensive unemployment are situations in which employed people
may be “at the mercy” of employers, having no alternative, and accordingly coerced.
“In periods of acute unemployment,” Hayek wrote (1978, 136-37), “the threat of
dismissal may be used to enforce actions other than those originally contracted for.
And in conditions such as those in a mining town the manager may well exercise an
entirely arbitrary and capricious tyranny over a man to whom he has taken a dislike.”
Unskilled employees are particularly vulnerable to manipulation by others in the
labour market, and therefore, from a classical liberal viewpoint, we may say that
they are vulnerable to coercion. Such an account of private coercion is consistent
with the customary cases of coercion classical liberals debated, say being mugged
or conscripted, but it also accommodates economic coercion.
“Coactus tamen voluit,” once said Max Weber (1978, 334), that is, “although
coerced, it was still his will.” Such a phrase is an oxymoron, and the reason why
so many understandings of coercion fail to grasp the evils of economic coercion is
because they do not see the contradiction in such a phrase. One is coerced when
one must act according to another person’s will rather than one’s own. As Arthur
Ripstein explains (2009, 107-08), according to Kant, “consent is fundamental to a
system of equal freedom,” as it expresses “each person’s entitlement to be his or her
own master.” Alternatively, Michael Weinstein wrote that “coercion as compulsion
produces the condition of slavery, in which the actions of the coerced become means
to the realization of another’s plan” (1972, 66). Simply put, when someone attempts
to mug you, when the state enacts a system of conscription, and when your boss
threatens to fire you if you join a union, then you are not acting as your own
master anymore. Such an understanding of coercion has been championed by
Hayek (1978, 13), and it is a key component of the classical liberal rationale
for market capitalism:
No Malibu Surfer Left Behind
339
The question of how many courses of action are open to a person is, of course, very
important. But it is a different question from that of how far in acting he can follow his
own plans and intentions, to what extent the pattern of his conduct is of his own design,
directed toward ends for which he has been persistently striving rather than toward
necessities created by others in order to make him do what they want. Whether he is free
or not does not depend on the range of choice but on whether he can expect to shape his
course of action in accordance with his present intentions, or whether somebody else
has power so to manipulate the conditions as to make him act according to that person’s
will rather than his own. Freedom thus presupposes that the individual has some assured
private sphere, that there is some set of circumstances in his environment with which
others cannot interfere.
The key point is that one is coerced to the extent that one is made to submit to the
will of others, and one is vulnerable to coercion to the extent that one is vulnerable to
manipulation by others. Having no reasonable alternative does not define coercion,
but it can certainly make one vulnerable to coercion. Market capitalism provides
people with an assured private sphere, and thus it reduces the amount of coercion
in society, or so classical liberals have argued. Economic necessity, however, often
provides opportunities for coercion. Having to do awful work for very low wages
with no reasonable alternative but starvation may place one at the mercy of one’s
employer. The manager may then become a “private tyrant,” requiring employees to
submit to his every caprice, say random drug testing, as Gourevitch notes (2016, 18).
Such an account of coercion is central to the classical liberal rationale, as it can be
used to explain its noted preference for market capitalism, limited government, and
decentralization. But it can also be used, as this article will demonstrate, to vindicate
a basic income guarantee.
Now, a second objection could go as follows: As Hayek said (1978, 120), most
employed people are not coerced in a market society because they act according
to their own will. “To do the bidding of others is for the employed the condition
of achieving his purpose. Yet, though he may find this at times highly irksome, in
normal conditions he is not unfree in the sense of being coerced” (Hayek 1978, 120).
Such a thought had already been introduced by Adam Smith (1776, III.iv.12) in his
discussion of the tradesman who “derives his subsistence from the employment,
not of one, but of a hundred or a thousand different customers. Though in some
measure obliged to them all, therefore, he is not absolutely dependent upon any one
of them.” As Philip Pettit mentioned (2006, 142), markets reduce dependency and
domination – or, if you will, it reduces the amount of coercion. Markets decentralize
power. They are characterized by a large number of buyers and sellers, none of
whom may manipulate specific people. Hence, Robert Taylor (2013) recently argued
for a celebratory attitude toward markets. Economic coercion in the labour market,
we could then say, should be addressed by more competition, that is, by a flexible
labour market, not a more robust safety net.
Classical liberals will thoroughly agree with Taylor, as, indeed, the rationale for
market capitalism rests on the fact that competition reduces the total amount of
coercion. Yet this is not enough from a classical liberal viewpoint. As I explained
elsewhere (Melkevik 2016b), although market societies will make most people
340
Business Ethics Quarterly
vastly better-off, they will also make a few people worse-off, that is, some people
will inevitably fall through the cracks of market institutions. This is a problem
since, as Zwolinski notes (2015, 523), the classical liberal theory “focuses on individuals, not on aggregates.” Though markets reduce the total amount of coercion,
they may not do so for everyone equally. There is thus still a problem of coercion,
if only because we know that some people like Sveinn will have to submit to their
tyrannical supervisor. Before we go on to argue such a point, however, we have
to consider three alternate understandings of coercion favoured by some classical
liberals and libertarians.
III. THE STATE AS AN APPARATUS OF COERCION:
THE TALE OF EDWIN AND SÓLI
Let us now consider two alternative theories of coercion championed by classical liberals like Ludwig von Mises as well as by libertarians like Robert Nozick,
focusing not on manipulation but on the state as an apparatus of coercion. Such
theories respond, for example, to Robert Lee Hale, a leading figure of the first
law-and-economics movement, who criticized laissez-faire liberalism. Simply put,
Hale’s challenge to laissez-faire (1923, 470) goes as follows: “The systems advocated by professed upholders of laissez-faire are in reality permeated with coercive
restrictions of individual freedom.” Private coercion is ubiquitous and unavoidable
for the working class, Hale claimed, as private coercion is actually backed by the
law of property in market systems. Or, as Zwolinski said, property rights are by
their very nature coercive.
Suppose a worker wants to eat a bag of peanuts, said Hale, and let me add
that this worker is named Sóli, which fittingly means “poor man” in Old Norse.
The bag costs five cents, which Sóli does not have. Of course, Sóli may not use
violence to get the peanuts from Edwin, name which means “rich friend,” as
forcible dispossession is one of the most obvious cases of coercion. This, however,
is not the real problem of coercion in the tale of Edwin and Sóli. According to
Hale, the problem is as follows. Edwin may remove at his discretion the legal
duty under which Sóli is in regard to the bag of peanuts. Provided Sóli really
wants to eat peanuts, he may then have to obey the will of Edwin. For example, Sóli may agree to pay five cents for the legal permission to eat a bag of
peanuts, or, in other cases, Sóli may have to conform to the terms of Edwin in
ways which are much more unpleasant, say working at disagreeable toil for a
slight wage. “Unless, then, the non-owner can produce his own food,” said Hale
(1923, 473), “the law compels him to starve if he has no wages, and compels
him to go without wages unless he obeys the behests of some employer. It is
the law that coerces him to wage-work under penalty of starvation – unless he
can produce food.” Can Sóli produce his own food, then asked Hale? Probably
not. The law of property forbids Sóli to cultivate any particular piece of ground
unless he happens to be its owner. Hence, as Hale thought, Sóli is truly coerced
in a market capitalist system, and it is the law of property that enables such
private coercion.
No Malibu Surfer Left Behind
341
Many classical liberals and libertarians disagree. “A person’s choice among
differing degrees of unpalatable alternatives,” Nozick explained (1974, 263f),
“is not rendered nonvoluntary by the fact that others voluntarily chose and acted
within their rights in a way that did not provide him with a more palatable alternative.”
One is not coerced to the extent that the circumstances that limit one’s alternative
are not themselves coercive in nature, for instance when the circumstances are just
other people exercising their rights. Nozick assumes, for example, that Edwin cannot
be said to coerce Sóli in acting within his rights, and that if all people act within
their rights with respect to Sóli, then, no matter how poor Sóli’s condition, Sóli is
not coerced. Therefore, following Nozick, libertarians have established a distinction
between coercion and circumstances that limit alternatives non-coercively, and
market exchanges are thought to be non-coercive inasmuch as such exchanges stay
within the bounds of the law. Such a theory, however, leads us to some conclusions
that cannot be reconciled with the broader classical liberal philosophy.
First, we would have to say that the legal owner of a spring in a desert would
not coerce people dying of thirst by asking an unreasonable price for his water,
inasmuch as he is simply exercising his rights over the spring. This case, though,
is the textbook example of a coercive monopoly. In the classical liberal rationale, it
is used to explain why anti-trust actions may be necessary to break up monopolies,
or it is used to explain why property rights must be arranged as to prevent such
monopolies to arise in the first place.
Second, we would also have to say that there is no market coercion involved
when, for example, Edwin threatens to fire Sóli for no good reason. As Gourevitch
explains (2016, 18), in the United States employers have legal powers called “managerial prerogatives,” which, within some broad boundaries, permit them to engage
in a master and servant relationship with their employees. For instance, “managers
may fire employees whose political views or sexual orientation or even physical
appearance they disapprove of” (Gourevitch 2016, 18). Were such powers exercised
by the state, surely classical liberals would condemn them as coercive.
The general problem, as Hayek explained (1978, 139), is that “in defining coercion we cannot take for granted the arrangements intended to prevent it.” The rights
employers have are such arrangements, and therefore we cannot say that Edwin does
not coerce Sóli in asking him to work at disagreeable toil for a slight wage because
he is permitted to do so per property and contract law. The same goes for the oasis
case. It means that one cannot identify people’s rights by the criterion “minimize
coercion,” and that people’s rights cannot define non-coercion in a given system. If
the background institutions of markets can be shown to lead to economic desperation
or manipulation by others, then we should not conclude that there is no problem of
coercion inasmuch these institutions have appropriately defined people’s rights. It
probably means that we should have additional institutions, say a safety net, which
will further minimize coercion in markets.1
Now, we may think, without authorized power to use violence, an institution cannot
be called coercive, such that only the state and some agencies can coerce. Such a
theory was championed by Ludwig von Mises. That is, private organizations within
market societies are not coercive, as they are not expressly authorized to resort to
342
Business Ethics Quarterly
violent action. More precisely, in a review of Hale’s book Freedom through Law,
Mises (1953, 410) wrote:
A state or a government is an apparatus of coercion and compulsion. Within the territory
that it controls, it prevents all agencies, except those that it expressly authorizes to do so,
from resorting to violent action. A government has the power to enforce its commands
by beating people into submission or by threatening them with such action. An institution
that lacks this power is never called a government.
As Barbara Fried noted (2001, 47), moreover, Mises clearly implied that an institution that lacks this power can never be called coercive. Edwin does not have the
authority to beat Sóli into submission, as Sóli may very well look elsewhere if he
wants a bag of peanuts, and therefore, Mises would say, the relationship between
Edwin and Sóli is not coercive. Edwin lacks authorized power, or the monopoly of
violence to put it in a Weberian way.
Such an account, however, misses several important points. First, it is not
because the state has a monopoly on legitimate force that it also holds a monopoly
on coercion. In fact, one may remember that Max Weber (1978, 313-9, 334-6,
729-31) established a distinction between “legal coercion” monopolized by
the state, “physical coercion” typical of a robber, “psychological coercion” as
exercised by a church, and the “indirectly coercive force of the market,” the
latter of which is what is missing in many classical liberal theories. Second,
and more importantly, the monopoly the state has on force is not total. People
can legitimately use force in self-defence. If Sóli tries to grab Edwin’s bag
of peanuts, Edwin is presumably authorized by the state to use some force in
resisting this invasion of his property rights. Hence, what the state has is not a
monopoly on legitimate use of force, but a monopoly on the authorization of
use of force, or, if you will, a monopoly on authoritative decisions about who
can and cannot use force. For instance, the state authorizes homeowners to use
some force in order to defend their property.
These two accounts of coercion proposed by Nozick and Mises, therefore, fail to
fit within the broader classical liberal rationale for market capitalism. A system of
property rights is indeed coercive, or, in other words, the law allows some people
to coerce others. “In protecting property the government is doing something quite
apart from merely keeping the peace. It is exerting coercion,” said Hale (1923, 472).
To be more precise, the state reduces the total amount of coercion in society by
employing the amount of coercion necessary to maintain property and contract law.
The same rationale is used more generally for limited government – though coercive,
criminal law reduces coercion. Hence, market capitalism has never been a system of
laissez-faire for classical liberals, as, say, Aaron Director (1964, 2) and Jacob Viner
(1960) explained. It is rather a carefully designed system whose purpose is to limit
coercion in society. That Nozick and Mises do not acknowledge the coercive nature
of markets should make classical liberals wary of their understandings of coercion.
These theories are non-generalizable, or ad hoc if you will, such that they cannot be
used to shed light on many classical liberal objectives, save for vindicating limited
government.
No Malibu Surfer Left Behind
343
IV. ON MALIBU SURFERS AND BEACH BUMS:
THE TALE OF VALDR AND LÓNI
The previously discussed theories of coercion do not satisfactorily capture the classical liberal objective of reducing the total amount of coercion in society. Whereas
Nozick’s theory defines coercion through the arrangements intended to prevent it,
namely a system of rights, Mises refuses to even consider the possibility of private
coercion in market societies. Overall, it therefore seems like classical liberals should
favour a theory of coercion as manipulation, following Hayek. That is, one is free
to the extent that one can expect to shape one’s course of action in accordance with
one’s present intentions. Now, the problem is the following: economic desperation
as well as periods of extensive unemployment can lead to coercion in market societies, even though, of course, state redistribution will be coercive as well. Classical
liberals, therefore, are facing a dilemma. A system of property rights will create
specific sets of circumstances that will make some people vulnerable to coercion,
being at the mercy of tyrannical employers, and yet the state cannot address such a
problem without having recourse to some coercion of its own.
Broadly speaking, in a market society, there are only two ways in which the state
can ameliorate poverty. First, and this is the usual strategy favoured by progressives,
we may alter the proportion of the gain received by each individual member, such
that we take from some to give to others. Second, we may increase the total social
gains possible from political organization, or, as Epstein explains (1985, 4) with a
tale of two pies, government action may move a society from a small pie to a larger
pie, while preserving the relative entitlement among the members of the group.
A rising tide, we are told, lifts all boats. Classical liberals normally favour this second Pareto improvement strategy, which is thought to be non-coercive, and which
is certainly the best way to tackle poverty in the long run. But we could also ask
the following question: What about the short run for people like Sóli and Sveinn
who are faced with immediate problems of private coercion?
Let me introduce a third story, which I call the tale of Valdr and Lóni. The conveniently named Valdr, which means “ruler,” is in fact the ruler of Malibu, while
Lóni is a Malibu surfer, whose name fittingly means “the lazy one” in Old Norse.
Malibu is peopled with somewhat affluent residents, though Lóni is certainly not
one of them. He is what we might call derogatively a “beach bum,” someone who is
resistant to work. It just so happens that Lóni likes to spend his days surfing, that is,
he has a high preference for leisure. Now, say Valdr is embarrassed by the inequality
between Lóni and everyone else in Malibu, and therefore he enacts a new municipal
tax – the Malibu surfer tax. There are about 13,000 people in Malibu, each of whom
now must pay $2 each year, which is then given to Lóni, hypothetically the only
poor individual in Malibu, as a basic income guarantee. This is an obvious case of
coercion for classical liberals – it is forced redistribution. Though I will later argue
for an even more comprehensive guarantee to counter the evils of private coercion,
we must first understand why state coercion has always been such a problem for
classical liberals, so that we may, later on, compare the evils of redistributive and
private coercion.
344
Business Ethics Quarterly
The problem is then the following. A free market economy will make everyone
better-off in the long run, including Sóli and Lóni, though it will also create the
conditions under which Sóli may have to agree to work for Edwin at disagreeable
toil for a slight wage in the short run. Meanwhile, Lóni will surf all day, as long as
there are still waves on Zuma Beach. “In the sweat of thy face,” says the Book
of Genesis (3:19), “shalt thou eat bread” – and yet Lóni could not care less about
working, even if it means he has to fast until someone like Valdr is compassionate
enough to present him a bit of bread free of charge.
To be more precise, we are now faced with three problems. First, if the state alters
the proportion of the gain received by each individual member, it is state coercion.
Second, if the state does nothing, then poor people like Sóli will have to endure
some cases of private coercion, perhaps forced to do some immiserating work, thus
becoming vulnerable to manipulation by others. Third, as Epstein noted (2002, 147),
state coercion will benefit some individuals greatly, like lazy Lóni, while others will
benefit only a small bit. Classical liberals may then prefer the evils of private coercion to the evils of state coercion, as the latter carries additional risks; for example,
state coercion may promote idleness and skew the distribution of gains. After all,
“The strongest of all the arguments against the interference of the public with purely
personal conduct,” said John Stuart Mill, “is that when it does interfere, the odds
are that it interferes wrongly, and in the wrong place” (2008, 92). Many classical
liberals have indeed been distrustful of state intervention, but, I want to suggest,
such a position is inconsistent with their concern for non-coercion.
Overall, we could think, classical liberals favour the following theory of coercion.
One coerces when one does not preserve the relative entitlement of the people in
a given society, as determined through a certain baseline – for example, the state
of nature or laissez-faire. “The implicit normative limit upon the use of political
power,” said Epstein (1985, 4), “is that it should preserve the relative entitlements
among the members of the group, both in the formation of the social order and in
its ongoing operation.” Such a theory permits us to explain why a limited government is needed to address private coercion – as, for example, the mugger disrupts
the relative entitlements of the people he robs. It also explains the classical liberal
distrust of progressive taxation, although as I have shown elsewhere (Melkevik
2016a), such a distrust may be misplaced. More importantly, such an account of
coercion echoes the intuition behind the famous “enough and as good” proviso John
Locke has championed (2003, 112, II.v.27):
Whatsoever then he removes out of the state that nature hath provided, and left it in,
he hath mixed his labour with, and joined to it something that is his own, and thereby
makes it his property. It being by him removed from the common state nature hath
placed it in, it hath by this labour something annexed to it that excludes the common
right of other men. For this labour being the unquestionable property of the labourer, no
man but he can have a right to what that is once joined to, at least where there is enough,
and as good, left in common for others.
Such a proviso, as Nozick understood it (1974, 175), “is meant to ensure that the
situation of others is not worsened.” Similarly, as Zwolinski notes (2015, 519), the
No Malibu Surfer Left Behind
345
proviso says that, “Natural resources should be used, in some sense, for the benefit
of all persons, not just a privileged few.” Now, although we may agree that market
capitalism makes most people better-off, as it dramatically increases the stock of
goods that can be acquired through trade, the problem is that some people will not
benefit from such a coercive system of property and contract law. Markets do not
benefit “everyone.” Hence, that people like Sóli cannot afford to buy a bag of peanut
without having to submit to the will of Edwin is a problem for classical liberals,
or at least it should be. It is true that people will develop legitimate expectations
around their property rights, permitting them to shape their course of action in
accordance with their intentions. But a given system of property rights can only
be used as a rough benchmark to decide what is or what is not coercive in a given
society, since such a system of rights cannot have any normative authority over our
understanding of coercion.
Redistribution will indeed alter the pro rata division within a given society, but
that may not be such a problem if the goal is indeed to reduce the total amount of
coercion. It is unlikely that a single mechanism, say market capitalism, will be able
to fulfill on its own such a classical liberal objective, and thus we may have to accept
some redistribution as a way to compensate for the shortcomings of market capitalism
that cannot address some forms of economic coercion in the labour market. In this
case, redistribution supplements market institutions, as we accept some coercion
to reduce a greater amount of coercion.
We should remember that private coercion is a problem in market societies
hic et nunc, that is, here and now, for a few people living destitute lives. Valdr is
certainly coercing Malibu residents with his Malibu surfer tax, but, I argue, he is also
preventing another kind of coercion. The crucial lesson from a classical liberal point
of view is that market capitalism will radically reduce the total amount of coercion
in society, which will benefit most people, though not everyone, unfortunately.
A few poor people like Sveinn and Sóli will have to submit to the unsavory character
of people like Harri and Edwin, that is, some tyrannical foreman or exploitative
businessperson. Since, as Zwolinski notes, the classical liberal tradition is individualistic, not aggregative, we therefore have to have a mechanism that ensures that
the total amount of coercion is reduced for everyone, not just for most people.
A basic income guarantee is such a mechanism, and, as I will suggest, such a guarantee should aid everyone, including people like Lóni, as lazy as they may be, if
we are to be consistent with the classical liberal theory of liberty.
V. FROM PRIVATE COERCION IN MARKETS TO THE
BASIC INCOME GUARANTEE
The classical liberal theory of freedom follows Kant’s intuition according to which
people have a strong interest in having no other person determine their own purposes
(Ripstein 2009, 111). Using such an account of coercion, I would now propose three
conclusions, each linked to one of the three tales I recounted, so that we may see
how the classical liberal theory of freedom justifies an unconditional basic income
guarantee, without assuming any altruistic motive.
346
Business Ethics Quarterly
First, concerning the tale of Edwin and Sóli, there is private coercion in market
societies, inasmuch as poor individuals often have no real possibility for autonomyexercising choice. Though some people may be ready to consent, for example, to
the conditions of sweatshop labour, it does not follow that such consent is also
an expression of self-mastery. Following Justice Holmes, we may say that “the
fact that a choice was made according to interest does not exclude duress” (Union
Pac. R. Co. v. Public Service Commission of Missouri, 248 U.S. 67, 70 [1918]). Of
course, we have to be careful with such an assumption, since not everyone can be a
professional baseball player. Most courses of action are only open for so many people, and yet recall that a limited course of action does not make one unfree according
to Hayek. The question is rather about having a private sphere with which no one
can interfere. Extreme poverty, we could then say, cultivates private coercion, as it
makes it that much easier for someone else to manipulate one. Economic desperation forces some people to submit to the will of others. In this case, that there are
many well-off employers who would offer lousy jobs does not make the would-be
employees any more free. Sóli’s consent to work for Edwin at disagreeable toil for
a slight wage is not necessarily an autonomy-exercising decision.
Hence, classical liberals should follow Gowder (2014, 306) in saying that the most
plausible version of economic liberty requires the state to “guarantee, if possible
that no one will be driven by economic desperation to engage in immiserating work,
which may impair rather than facilitate his or her pursuit of the good.” Simply put, a
classical liberal polity should be concerned with the pressing issue of occupational
choice. For many unskilled workers, entering the labour market may be a coercive
experience, in which they can be made to agree, through a binding agreement, to
submit to the will of an employer so not to starve.
Second, concerning the tale of Sveinn and Harri, and building on our first conclusion, we must consider an additional side of private coercion, namely what could
be called the problem of “little tyrants.” Though Sóli may face coercion, having to
go work for Edwin if he wants to eat, we must acknowledge that people like Sveinn
also face coercion within the workplace. Such a point was well-understood by
Nozick (1997, 16), as he gave the following example of workplace coercion: “You
threaten to get me fired from my job if I do A, and I refrain from doing A because
of this threat,” and therefore “I was coerced into not doing A.” In some industries,
say the worst service-sector jobs, workers are regularly faced with such instances of
private coercion, which should perhaps make us question the justice of the doctrine
of employment-at-will, too often enabling “little tyrants” within a given company.
Save for some protections against gender and race discrimination, employees may
then be fired for good or bad reasons, as well as for no reason at all, which may
enable bizarre and undignified cases of private coercion. For example, employees
are sometimes forbidden to pee, or commanded to pee, or asked to dress in some
unusual ways. As Bertram, Robin, and Gourevitch noted (2012), employers are then
granted powers forbidden to the state – they are free to punish without due process.
The tale of Edwin and Sóli demonstrated that the capitalistic organization of
society itself includes some inherently coercive features, forcing people to enter the
labour market, less they starve. The tale of Sveinn and Harri, conversely, shows how
No Malibu Surfer Left Behind
347
people often have to face cases of coercion inside the workplace. Overall, then, there
are problems of private coercion both inside and outside the workplace, which must
be addressed by the state. A basic income guarantee is one such way to address the
problems of private coercion. It permits people, first, to resist having to integrate
the labour market under coercive circumstances, and, second, to refuse to go along
with the caprice of tyrannical foremen. If no one is to be left behind, then classical
liberals must embrace such a guarantee. This conclusion can be accepted by most,
if not all, classical liberals. For example, we may remember that Hayek has argued
for an assured minimum income (2013, 249), that is, a non-market safety net, or “a
floor below which nobody need to descend.” The basic income, however, if made
unconditional, will lead us to the Malibu surfer problem we examined.
Third, then, concerning the tale of Valdr and Lóni, the Malibu surfer problem is
not as serious an objection to a social safety net as may be thought. Some classical
liberals like Epstein may very well complain against giving a 22-year-old $10,000
per year for living at home with his parents and hanging out with his friends (2002,
166), and yet, we should ask, what is the alternative? Are we to give money only to
the “deserving poor,” like the Old Poor Law did? As a matter of fact, Epstein thinks
so (167), as he wrote that “no gift of assistance should be made in unconditional
fashion. Some investigation should be made of the condition of the recipient,
and some restrictions should be imposed on the way in which the gift, whether
in cash or kind, is used.” If we are to reduce the total amount of coercion for
everyone, and not just for most people, our third conclusion must be that some
gift of assistance should be made unconditionally, without which the so-called
“underserving” poor would be vulnerable to private coercion. Also, we would
then have to invade the private life of all those who would receive state assistance, which is not something that can be welcomed by classical liberals. Will
an unconditional guarantee lead to a life of self-indulgence for some people
like Lóni and idle teenagers? Yes, it will. But the problems of private coercion
Sóli and Sveinn are facing are serious enough that we must be ready to shoulder
such a cost. As undeserving as a few people may be, an unconditional safety
net safeguards market societies from the evils of private coercion, permitting
people to avoid intolerable works, be they degrading or unbearably tedious
(Gowder 2014, 306).
“For the ‘underserving poor,’ those responsible for their own fate,” once stated
the Charity Organisation Society, “abuse of the relief system could only be curbed
by ‘making it as distasteful as possible to the applicants; that is, by insisting (as a
general rule) on a labour test or residence in a workshop.’” After approvingly quoting
that passage, typical of the Benthamite rhetoric used in the Poor Law Amendment
Act 1834, Rothbard (1978, 148) explained that the libertarian position calls for
“the complete abolition of governmental welfare and reliance on private charitable
aid, based as it necessarily will be on helping the ‘deserving poor’ on the road to
independence as rapidly as possible.” Some classical liberals have indeed defended
a similar position as they argued against the welfare state. The main problem was
perhaps better captured by Jeremy Bentham (1843, III.ii.1, 401) when he explained
the following: “In this country, under the existing poor laws, every man has a right to
348
Business Ethics Quarterly
be maintained, in the character of a pauper, at the public charge: under which right
he is in fact, with a very few exceptions, (amounting not to one perhaps in fifty,)
maintained in idleness.” Simplifying greatly, welfare benefits putatively encourage
laziness, and social justice is consequently not to be understood as a reciprocal
relationship among different parties. It rather becomes an institutional protection
of idleness.
Such a problem, following John Rawls (1988, 257), is now colloquially known
as the Malibu surfer problem, that is, the welfare state is thought to lead to a life
of self-indulgence for some people who consume more than their fair share of
public resources. To use state coercion to support these idle persons, it is therefore
thought, is wrong. This article has disproved such a thought from a classical liberal
viewpoint. It is not viable to merely help the deserving poor. Leaving aside the
intractable problems that would accompany any attempt to pick up those who are
indeed “deserving,” we need to consider how market institutions can be coercive,
either directly or indirectly, which, of course, needs to be remedied if we accept the
classical liberal rationale.
Consider what may be called the “exploitation objection,” which says that it is
wrong to live off the labour of others (Rawls 2001, 179). Stuart White proposed three
responses to this objection, but for classical liberals only one of these answers suffices
to justify an assured income. The “balance of fairness response,” explains White
(2006, 4f), goes as follows: “Introducing a basic income will, on balance, prevent
greater injustice than it might create.” For classical liberals, the greater injustice is
private coercion, such as having one’s privacy invaded by one’s employer, having
to obey the impulsive will of your manager, or having to do demeaning sweatshop
labour so not to go hungry.
As White noted, such a response concedes that the basic income guarantee
will introduce some injustice. It will disrupt the relative entitlement of the
people, making some people worse off so as to make others better off. Such
redistribution is coercive. It is also most probably exploitative for Lóni to live
off the labour of Sóli and Sveinn. But, overall, the argument is that the evils of
private coercion are greater than those accompanying a comprehensive social
safety net, and the evils to be removed by such a safety net are greater than
those that accompany the governmental interference necessary to carry out such
a net. If we follow Hayek in defining coercion as being made to act according
to someone else’s will rather than one’s own, then it seems the above examples
of private coercion are worse than being forced by your government to pay a
few more tax dollars.
Such a “balance of fairness” argument shows why classical liberalism should
indeed be committed to an unconditional basic income guarantee. A different liberal
tradition that would not be so concerned with coercion, however, may have to use
additional arguments to justify such a guarantee. But the classical liberal concern
for non-coercion is strong enough that no additional argument is needed, provided
that we can show that the total amount of coercion in society can be reduced for
everyone by the state employing the amount of coercion necessary to maintain such
a basic income guarantee.
No Malibu Surfer Left Behind
349
We have now established that the classical liberal state is obligated to maintain a basic income guarantee by its underlying justification. Yet one could want
additional reasons to support an unconditional guarantee rather than a conditional
one. Let me propose the following thought. The purpose of the basic income
guarantee is to enhance everyone’s capacity to refuse to do immiserating work,
which makes people vulnerable to coercion. To make such a guarantee conditional
on willingness to work, therefore, would be self-defeating. That is, if we assess
the people’s willingness or ability to work before we provide the basic income,
then we essentially defeat the initial objective of the guarantee, which is to create
a private sphere so as to counter the coercive nature of the labour market. In this
case, coherence requires of classical liberals that they endorse an unconditional
guarantee, because their individualistic theory cannot justify leaving anyone
behind, and because they cannot invade the private life of all those who need or
would need state assistance.
It is regrettably common to say that the classical liberal tradition is consequentialist and focused on aggregate happiness (Gray 1982, 50), especially following
John Rawls (2000, 366) and Samuel Freeman (2007, 45). Yet to do so is almost
certainly to misread classical liberalism. After all, Hayek did not call his book The
Constitution of Happiness but The Constitution of Liberty, and, similarly, Friedman
did not title his book Capitalism and Happiness but Capitalism and Freedom. The
classical liberal justification for the state consists in the need to reduce the amount of
coercion to which individuals are subject, not to make people happier, and therefore
the above argument for an unconditional basic income guarantee can explain both
the wider features of a classical liberal polity as well as the principled commitment
to that guarantee – which was never explicitly given by Hayek, and which is often
thought to be missing by neoclassical liberals.
In conclusion, following the theory of coercion championed by Hayek, this article
has provided a thoroughly principled classical liberal argument for an unconditional
basic income guarantee. Of course, classical liberals should try to find institutional
ways in which we may lower the cost of such assistance, but the assistance itself
cannot be questioned. As Friedman noted (2002, 192), any redistributive measure to
ameliorate poverty will reduce the incentives poor people have to help themselves.
This is a problem, and yet it is nowhere as serious as the problems of private coercion
in a market society. “In the general course of human nature,” said Alexander
Hamilton (Hamilton, Madison, and Jay 1999, 471), “a power over a man’s subsistence amounts to a power over his will.” If we are to accept the classical liberal
rationale for market capitalism, this is simply not something that can be allowed:
Sóli and Sveinn should always be able to act according to their own will, never
according to whim of their boss.
ACKNOWLEDGEMENTS
I am grateful to Andrew Lister, three anonymous reviewers, as well as to associate
editor Jeffrey Moriarty and editor in chief Bruce Barry, for enlightening comments
and sound advice.
350
Business Ethics Quarterly
NOTE
1. Classical liberals, moreover, usually endorse the doctrine of employment-at-will, permitting both
employers and employees to terminate the employment relationship for any reason. They must, however,
address the coercive consequences of such a doctrine in market societies (Payne v. Western & Atlantic
Railroad Co., 81 Tenn. 507, 518-19 [1884]). A complete government monopoly of employment, such
as we could find in socialist states, is coercive from a classical liberal viewpoint. In such cases, there
is no possibility to work for another employer, and therefore, Hayek noted (1978, 137), the employer
would “possess unlimited powers of coercion.” Yet we may note that if the law grants all employers
in a market society discretionary powers, for example to act as capricious tyrants, then such a system
is most probably coercive as well. Employees may go work elsewhere, but, wherever they go, they
will have to deal with the same power of coercion. That markets are decentralized is not a guarantee
of non-coercion if the rules permit manipulation by others, and if they lead to economic desperation
for at least some people.
REFERENCES
Bentham, Jeremy. 1843. “Outline of a Work Entitled Pauper Management Improved.” In
The Work of Jeremy Bentham Vol. VIII, 369–439. Edinburg: William Tait.
Bertram, Chris, Corey Robin, and Alex Gourevitch. 2012. “Let It Bleed: Libertarianism
and the Workplace.” Crooked Timber, July 1. http://crookedtimber.org/2012/07/01/
let-it-bleed-libertarianism-and-the-workplace/.
Brennan, Jason, and John Tomasi. 2012. “Classical Liberalism.” In Oxford Handbook
of Political Philosophy, edited by David Estlund, 115-32. New York: Oxford
University Press.
Cohen, Gerald A. 1983. “The Structure of Proletarian Unfreedom.” Philosophy and Public
Affairs 12: 3–33.
Director, Aaron. 1964. “The Parity of the Economic Market Place.” Journal of Law and
Economics 7: 1–10.
Epstein, Richard. 1985. Takings. Cambridge: Harvard University Press.
———. 2002. “Can Anyone Beat the Flat Tax?” Social Philosophy and Policy 19: 140–71.
Freeman, Samuel. 2007. Rawls. New York: Routledge.
Fried, Barbara. 2001. The Progressive Assault on Laissez Faire. Cambridge: Harvard
University Press.
Friedman, Milton. 2002. Capitalism and Freedom. Chicago: University of Chicago Press.
Gourevitch, Alex. 2016. “The Limits of a Basic Income: Means and Ends of Workplace
Democracy.” Basic Income Studies 11: 17–28.
Gowder, Paul. 2014. “Market Unfreedom.” Critical Review 26: 306–47.
Gray, John N. 1982. “F. A. Hayek and the Rebirth of Classical Liberalism.” Literature of
Liberty 5: 19–66.
Hale, Robert L. 1923. “Coercion and Distribution in a Supposedly Non-Coercive State.”
Political Science Quarterly 38: 470–94.
Hamilton, Alexander, James Madison, and John Jay. 1999. The Federalist Papers. Edited
by Clinton Rossiter. New York: Signet Classic.
Hayek, Friedrich. 1978. The Constitution of Liberty. Chicago: University of Chicago
Press.
———. 2013. Law, Legislation and Liberty. London: Routledge.
Locke, John. 2003. Two Treatises of Government. New Haven: Yale University Press.
Macleod, Alistair M. 2008. “Coercion, Justice, and Democracy.” In Coercion and the State,
edited by David A. Reidy and Walter J. Riker, 63-76. Berlin: Springer.
No Malibu Surfer Left Behind
351
Marx, Karl, and Friedrich Engels. 1998. The German Ideology. New York: Prometheus Books.
Melkevik, Åsbjørn. 2016a. “No Progressive Taxation Without Discrimination? On the
Generality of the Law in the Classical Liberal Tradition.” Constitutional Political
Economy 27: 418–434.
———. 2016b. “Four Concepts of Rules: A Theory of Rule Egalitarianism.” European
Journal of Political Theory. doi: 10.1177/1474885116653366.
Mill, John S. 2008. On Liberty. Oxford: Oxford University Press.
Mises, Ludwig von. 1953. “Review of Freedom through Law.” The Freeman 410, March 9.
Nozick, Robert. 1974. Anarchy, State, and Utopia. New York: Basic Books.
———. 1997. “Coercion.” In Socratic Puzzles. Cambridge: Harvard University Press.
Pettit, Philip. 2006. “Freedom in the Market.” Politics, Philosophy, and Economics
5: 131–149.
Rawls, John. 1988. “The Priority of Right and the Ideas of the Good.” Philosophy and
Public Affairs 17: 251–76.
———. 2000. Lectures on the History of Moral Philosophy. Cambridge: Harvard University
Press.
———. 2001. Justice as Fairness: A Restatement. Cambridge: Harvard University Press.
Ripstein, Arthur. 2009. Force and Freedom: Kant’s Legal and Political Philosophy.
Cambridge: Harvard University Press.
Rothbard, Murray. 1978. For a New Liberty: The Libertarian Manifesto. New York: Collier
Books.
Smith, Adam. (1776) 1976. An Inquiry into the Nature and Causes of the Wealth of
Nations. Edited by R. H. Campbell, A. S. Skinner, and W. B. Todd. Oxford: Oxford
University Press.
Taylor, Robert S. 2013. “Market Freedom as Antipower.” American Political Science
Review 107: 593–602.
Tomasi, John. 2012. Free Market Fairness. Princeton: Princeton University Press.
Viner, Jacob. 1960. “The Intellectual History of Laissez Faire.” The Journal of Law and
Economics 3: 45–69.
Weber, Max. 1978. Economy and Society. Berkeley: University of California Press.
Weinstein, Michael A. 1972. “Coercion, Space, and the Modes of Human Domination.”
In NOMOS XIV: Coercion, edited by J. Roland Pennock and John W. Chapman,
63-80. Chicago: Atherton.
Wertheimer, Alan. 1987. Coercion. Princeton: Princeton University Press.
White, Stuart. 2006. “Reconsidering the Exploitation Objection to Basic Income.” Basic
Income Studies 1: 1–17.
Zwolinski, Matt. 2015. “Property Rights, Coercion, and the Welfare State: The Libertarian
Case for a Basic Income for All.” The Independent Review 19: 515–529.
Copyright of Business Ethics Quarterly is the property of Society for Business Ethics and its
content may not be copied or emailed to multiple sites or posted to a listserv without the
copyright holder's express written permission. However, users may print, download, or email
articles for individual use.
Ethics Inf Technol (2015) 17:219–229
DOI 10.1007/s10676-015-9377-6
ORIGINAL PAPER
The tragedy of the master: automation, vulnerability, and distance
Mark Coeckelbergh1
Published online: 3 September 2015
! Springer Science+Business Media Dordrecht 2015
Abstract Responding to long-standing warnings that
robots and AI will enslave humans, I argue that the main
problem we face is not that automation might turn us into
slaves but, rather, that we remain masters. First I construct
an argument concerning what I call ‘the tragedy of the
master’: using the master–slave dialectic, I argue that
automation technologies threaten to make us vulnerable,
alienated, and automated masters. I elaborate the implications for power, knowledge, and experience. Then I critically discuss and question this argument but also the very
thinking in terms of masters and slaves, which fuels both
arguments. I question the discourse about slavery and
object to the assumptions made about human–technology
relations. However, I also show that the discussion about
masters and slaves attends us to issues with human–human
relations, in particular to the social consequences of
automation such as power issues and the problem of the
relation between automation and (un)employment. Finally,
I reflect on how we can respond to our predicament, to ‘the
tragedy of the master’.
Keywords Automation ! Artificial intelligence !
Robotics ! AI ! Ethics of AI ! Philosophy of AI ! Master–
slave dialectic
Introduction
Echoing a long tradition going back at least to Aristotle’s
comments in the Nicomachean Ethics about the slave as a
‘living tool’ and the tool as a ‘lifeless slave’ (Aristotle,
1161b2-7), thinking about technology in terms of masters
and slaves is more popular than ever. Current discussions
about automation, robotics, and AI in the robotics/AI
community and beyond seem to be am especially fertile
ground for such a thinking. For instance, there have been
many recent warnings about AI or robots taking over,
indeed becoming our masters rather than the other way
around. Consider for instance claims about AI in the media
by Hawking,1 Musk,2 Wozniak,3 and Rees.4 Sometimes the
future of AI, automation, and robotics is explicitly discussed in terms of ‘‘masters’’ and ‘‘slaves’’. For instance,
echoing a well-known theme in the history of thinking
about machines, Wallach compares technology with ‘a
dangerous master’ which gets out of control (Wallach
2015) and Bryson has argued that ‘robots should be slaves’:
we should not humanize them or make them into moral
agents, but rather as ‘tools we use’ (Bryson 2010).
In this paper, I do not simply deny that machines are
going to take over (by contra-predicting that this will not
happen), assume that this is a possibility (e.g. by endorsing
Hawking or by claiming that instead we should develop
robots that are slaves, as Bryson did), or point out that this
1
http://www.bbc.co.uk/news/technology-30290540.
http://www.theguardian.com/technology/2014/oct/27/elon-muskartificial-intelligence-ai-biggest-existential-threat.
3
http://www.afr.com/technology/apple-cofounder-steve-wozniak-onthe-apple-watch-electric-cars-and-the-surpassing-of-humanity-2015
0323-1m3xxk.
4
Martin Rees, ‘‘Meet your replacement’’, The Daily Telegraph, 16
May 2015.
2
& Mark Coeckelbergh
mark.coeckelbergh@dmu.ac.uk
1
Faculty of Technology, Centre for Computing and Social
Responsibility, De Montfort University, 5.79 Gateway
House, The Gateway, Leicester LE1 9BH, UK
123
220
M. Coeckelbergh
Masters, slaves, and automation technology
a struggle for recognition: ‘each seeks the death of the
other’ and this ‘involves the staking of its own life’ (Hegel
1807, p. 113). The result of this struggle to death is that one
masters the other; this is the ‘‘master’’, the victor in the
mortal struggle. The ‘‘slave’’ then works for the master.
This, however, makes recognition impossible since the
slave6 (qua slave) cannot give it to the master since the
slave lacks ‘independent consciousness’ (p. 117). Moreover, whereas the master’s relation to nature is mediated by
the slave, the slave has a more direct relation to nature
through work: the slave makes products for the master.
Having created products, the slave ‘realizes that it is precisely in this work wherein he seemed to have only an
alienated existence that he acquires a mind of his own’ (p.
119).
Hegel’s master–slave dialectic is more complex and
must be interpreted in the light of the entire book, that
is, in the light of the history of the spirit Hegel narrates.
For the purpose of this paper, I will be content with this
brief version as a working version of the dialectic used
for the construction of the argument in this section,
retaining in particular (1) the insight that the master, who
has turned the other into a slave, then becomes entirely
dependent on the slave and (2) Hegel’s interesting suggestion that masters and slaves have a different relation
to nature: the one mediated, the other unmediated
(through work).
Now what does this dialectic imply if applied to the
issue of automation and, more generally, to the human–
technology relation? In line with previously mentioned
claims recent claims about AI and robots taking over, it
might be interpreted to imply that there is a struggle
between humans and technology, in which technology
becomes our master. There seems to be a reversal of roles:
first technology was our slave, as Aristotle said and as it
was ‘‘meant’’ to be (the idea being that technology is
developed in order to serve humans), but now the machine
becomes our master. The robots will be our new lords and
masters, we will be the slaves.
However, this is not really what the master–slave
dialectic says, even in this brief working version. The slave
does not become the master; the master remains master, but
is what we could call a troubled master, who lacks true
independence—being dependent on the master for recognition and subsistence—but at the same time cannot step
out of the master role. One could call this ‘‘tragic’’, since
on the one hand it is the result of the will and action of the
In the Phenomenology of the Spirit (1807) Hegel describes
how self-conscious develops between two beings. There is
6
kind of warnings distract from other, more urgent problems
(as I have done in my Wired article5). Instead I first revise
the thesis that robots will enslave us—claiming that it’s
exactly the other way around: the problem is not that we
become slaves but that we remain masters—and then
question the very terms in which the problem is framed.
Let me explain.
In the first section I do not defend talking about robots
and AI in terms of masters and slaves, but ask: if we talk
about our relation to technology in this way, then how may
this conceptualize the human–machine relation? What does
automation imply for ‘‘masters’’ and ‘‘slaves’’? What follows for the role of humans? What follows for machines?
Who is master and who is slave? What are the social
implications? I construct the argument on what I call ‘the
tragedy of the master’: automation technology developed
to serve us renders us vulnerable, creates distance to
material reality, and constrains our actions to those compatible with the technology. In the second section, I then
present a critique of the argument constructed in the previous section. In particular, I question the terms in which
this discussion is framed and criticize some assumptions.
First I point to the concept of ‘‘slave’’, which is seldom
unpacked in the present discussions, to moral–historical
meanings attached to the term ‘‘slaves’’ (and indeed
question what the very term means). Then I criticize the
assumptions about human–technology relations implied in
the discussion. I also ask whether alternative (automation?)
technologies might avoid the problems indicated in this
paper and whether we might move beyond master–slave
thinking. In this way the paper aims to contribute to
thinking about automation, AI, and robotics, but also to
philosophy of technology in general.
Section I: Automation technology and the tragedy
of the master
In this section I construct the thesis that, contrary to the
received view, the main problem with automation is not
that we might become the slaves of technology but that we
remain its (troubled) masters. Let me start with Hegel’s
influential so-called ‘‘master–slave dialectic’’ to support
this claim and, on the way, re-cast some of the arguments
in my work about technology, vulnerability (Coeckelbergh
2013a) and distance (e.g. Coeckelbergh 2013b, 2015a, b).
5
http://www.wired.com/2014/12/armageddon-is-not-the-ai-problem/.
123
Note that ‘‘master’’ and ‘‘servant’’ may be a better translation of the
German Herr and Knecht. However, here I will continue to use the
terms ‘‘master’’ and ‘‘slave’’ in order to remain in line with the
popular discussion.
The tragedy of the master: automation, vulnerability, and distance
master to become the master, but on the other hand once
the master position is achieved the highly desired independence and freedom turns into a new kind of dependence
and bondage.
This happens with technology, and in particular with
automation technology: we created technology to be our
slave. We gain from this since we do no longer have to do
the ‘‘dirty’’ work and have more comfort, but we pay a
price: we become dependent on the technology. In particular, re-interpreting what I have said about vulnerability
and technology (Coeckelbergh 2013a, b), the master
becomes a dependent and vulnerable master whose vulnerability is transformed in specific ways by the tools used.
Furthermore, as masters of the machines we become
alienated from nature since the work is done by the
machine. Let me further explain and elaborate these claims,
and add a related but third claim about what automation
does to us.
Automation technologies render us (1) vulnerable,
(2) alienated, and (3) automated
First, automation renders us vulnerable since we transferred work to the machine: if something happens to the
automation system we are lost. In this sense, our vulnerability is transformed by technology in a specific way: we
are no longer or at least not mainly ‘‘directly’’ vulnerable
to natural processes but we become vulnerable to the
machine. As we delegate work to the machine and the
machine comes to mediate our being-vulnerable to nature,
we are rendered vulnerable to the machine, to the technology. This creates the paradox or irony that we desired
to become less vulnerable and more independent from
nature, but the very tools we use for this purpose create
new vulnerabilities and dependencies (Coeckelbergh
2013a, b). We wanted to be masters of nature and in our
struggle with nature we put our lives at risk in order to
achieve that mastery, but once achieved, then as masters
of technology we become more dependent than ever—now
on technology. We thus find ourselves in the Hegelian
tragic situation of the master: it is through our own
actions, aimed at freedom, that we achieve (a new kind of)
bondage.
Second, automation renders us alienated from material
reality, since we now work in ways that no longer require
intense and direct engagement with that material reality;
the machine mediates between us and material–physical
reality, between us and nature. It thereby creates a gap, a
distance, between us and nature. Now in so far as technologies mediate our relations to the world this always
happens to some degree (for instance a shovel is ‘‘between’’ a gardener and the soil) and hence there is always
some distance, but some technology may increase the
221
distance.7 Automation, in particular, worsens the problem
concerning distance to nature if and in so far the action is
entirely delegated to the machine. This is different from
mere mediation of what remains largely human action, as
for example when we use a hammer or a shovel, or even a
tele-operated robot.8 Whereas previously the technology
was an aid to human action (for example the hammer is an
aid; the agency remains with the human), now the action
itself is entirely in the hands of the machine. This changes
the relation to nature: it creates more distance. Agency is
transferred from humans to machines in the sense that they
now do the work (however, they lack the ability to give
recognition to the human). The machines are now artificial
agents, which, like humans, work upon the world but are—
in contrast to us—in direct contact with nature. Thus, here
the point is not that these artificial agents will become our
masters, but rather that even if this does not happen, we
have a problem qua masters. In order to master nature, we
have created entities that were meant to be our artificial
slaves, but this has created distance to nature. Only the
entities that serve us engage in material–physical activities,
mix themselves with nature. Only the machines are in close
touch with nature; we are their alienated and dependent
masters.
Note that this point differs from Marx, who is of course
a famous interpreter of Hegel: it is neither about loss selfexpression, as in the early Marx (1844), nor about capital,
as in the Marx of Capital (1867). The emphasis here is on
the alienation from nature, not so much on the alienation
from the work, the products of the work, the other workers,
7
For instance, in my work on technology and distance, I have argued
that electronic information and communication technologies (ICTs)
tend to increase the distance—literally, by enabling communication
over long distances, but also metaphorically, creating social and
moral distance [see for instance my claims about financial technologies and distance in Money Machines (Coeckelbergh 2015a) and
about nature and ICTs in Coeckelbergh (2015b)]. However, here I
will focus on the issue of automation.
8
There is also use of automation technology that does not fully
delegate the action to the machine. In such cases, the machine plays
the role of mediator between humans and their environment. For
instance, a surgeon may use a robot to remotely operate on a patient.
In such cases, there is at least less distance to nature than in cases of
full delegation of agency: there is technological mediation, but the
action remains in the hands of the surgeon to an important degree. It
could even be argued that at least some teleoperated systems decrease
or bridge the distance between us and nature, in particular when they
enable us to remotely manipulate and experience what could not be
manipulated or experienced without the system. Consider for instance
exploration of another planet by means of a tele-operated robot.
However, in spite of this bridging of physical and experiential
distance, there may remain some feeling of alienation as it is still an
indirect way of relating to nature. The humans operating the machine
and experiencing nature in this way are also highly dependent on the
technology (see also the first point again).
123
222
etc., and the point is also independent from claims about
capitalism.
Third, automation automates us, ‘‘masters’’, since once
a practice is automated we have to adapt our behavior and
indeed the entire practice to the machine. Although we do
not become the slaves of the machine in the sense that
they ‘‘command’’ us what to do, it is in our very role as
‘‘masters’’ of the machine that we adapt to what the
machine does and can do.9 Again this creates the paradox
that we desired independence and mastery and used
technology to achieve this, but the result is that, in so far
as we become automated, we are turned into machines
ourselves, that is, we are turned into the most unfree and
dependent entities we can imagine. Since Descartes
machines have always played the role of the negative of
the human: we do not want to be machines, which means
here that we do not want to be totally unfree; instead we
want to be free, autonomous, and self-directing. And in a
Hegelian context we should add: we want to be selfconscious beings. In so far as we become more machinelike, we lack the conditions for this freedom and selfconsciousness. Machines cannot do what the slave does
through work: the slave sees herself in the products of her
work, but the machine has no possibility to achieve that
self-consciousness. Moreover, it is clear that in our relation with the machine we cannot get recognition at all.
Whereas in the case of the human slave it at least makes
sense to discuss the question, the machine as slave takes
away every possibility for recognition, since only humans
can give such recognition.
To conclude, automation raises the question of what we
may call ‘‘the tragedy of the master’’: there is a risk that the
automation technology we developed and use to serve us
renders us vulnerable and dependent in new ways, creates
distance between us and material reality, and ‘‘automates’’
us in the sense that we have to adapt our practices what
automation technology does and can do [what I have called
‘‘the automation of the social’’ (Coeckelbergh 2014)]. I
have argued above, however, that if we take inspiration
from Hegel’s master–slave dialectic we should not conceptualize these risks as implying that the machines
become our masters. Instead, I have argued that we remain
troubled and tragic masters, who become vulnerable and
dependent on the machine, alienated from nature, and
automated by the machine.
Note that the vulnerability and the alienation mentioned
are still the vulnerability and alienation of the master, not
of the slave–machine. The slave–machine has other vulnerabilities as a machine and as a slave. As a machine it
9
See again my claims about ‘‘the automation of the social’’ in my
recent keynote addresses at the Robophilosophy conference and at
AISB 2015, 21 April 2015 (Coeckelbergh 2014).
123
M. Coeckelbergh
lacks the being-vulnerable of the human master. As I have
argued (Coeckelbergh 2013a, b), humans always experience their vulnerability, it is ‘lived’ vulnerability. As a
slave, the machine is dependent on the master in the sense
that the master uses the slave, can switch off the slave, can
be violent towards the slave etc. Note again, however, that
this is different from the threat of death in the case of the
human slave, again because machine vulnerability is categorically different from human vulnerability: only
humans and other living beings can die, only humans can
anticipate their death and, more generally, compared to
machines only humans are-at-risk (Coeckelbergh 2013a,
b). Machines cannot experience, let alone experience risk
and vulnerability. Thus, the vulnerability of the master is of
a specific kind: it is a human kind of vulnerability. Furthermore, the vulnerability of the master is related to her
alienated position: it is made possible by not working, with
‘‘working’’ understood as a mixing of one’s labour with
nature.10 It is made possible by being dependent on the
slave for the material economy, which alienates the master
but also makes her vulnerable. If the machine stops functioning and hence stops working, she is lost, since she no
longer knows how to work. This brings me to my next
point.
Vulnerability and alienation are also related to different
kinds of knowledge: the human master lacks the embodied
know-how of the slave, a know-how which can only be
developed through work, through mixing with nature.
Now when automation takes over work, not only the
human master loses this connection and this knowledge.
The human slave, who now uses a machine to do the
work and becomes an operator rather than a worker, also
loses this kind of knowledge, since when automation is in
place, the machine is the only entity which is in direct
and active contact with its environment (without, of
course, having experience and therefore it lacks true
engagement and involvement). There is a loss of knowhow. Furthermore, whereas before the slave could gain
self-consciousness through work, now the slave loses this
possibility. Therefore, both the human master and the
human slave become alienated. Let me further discuss the
issue of knowledge.
10
One may object that the master still has the work of ensuring that
the automation continues to work. However, in so far as this means
that the master functions as an operator who monitors and supervises,
this does not count as ‘‘work’’ according to the definition ‘mixing
one’s labour with nature’. It only counts as ‘‘work’’ when the master
repairs the machine. This is indeed a less alienated relation and
involves a different kind of knowledge: know-how. I will say more
about this below.
The tragedy of the master: automation, vulnerability, and distance
Automation, knowledge and its relation to power,
dependence, and alienation
When automation is introduced in industry, health care,
and the military, for instance, the slave–operators and the
master–managers are distant from their environment. In so
far as the machine takes over production, care, and military
action, both the managers (commanders, etc.) and the
‘‘workers’’ (who are no longer workers but operators,
including care-operators and soldiers–operators) become
alienated from the things and people they used to work
with. This distance also has implications for knowledge.
The automation shift creates a problem in terms of loss of
know-how: although at first the operators still have knowhow derived from their previous, hands-on work, as
Bainbridge (1983) suggested the next generations of
operators lack this practical knowledge: ‘Unfortunately,
physical skills deteriorate when they are not used … This
means that a formerly experienced operator who has been
monitoring an automated process may now be an inexperienced one’ (Bainbridge 1983, p. 775) Thus, operators
lack the knowledge to supervise the machines if they have
never ‘‘worked’’ themselves. This means that they become
even more dependent on the machines. Similarly, if
something goes wrong supervisors ‘will not be able to takeover’ since they also lack the relevant know-how and
manual skill (p. 776).
Perhaps those people called ‘‘workers’’ can achieve a
less alienated relation when they are no longer operators or
supervisors but instead become experienced repairers of
the machines: this requires know-how and skill, which are
developed through work. As they work on the machine in
order to enable the machine to continue its work, the
repairers are less alienated than the operators and supervisors. Mixing their labour with nature, it seems that they
are workers again. They become less dependent on the
machine—indeed less vulnerable—and regain some power
over those who do not know how to repair. However, it
remains unclear how ‘‘direct’’ their relation to nature is
since they mainly relate to the machine: Is this a direct
relation to nature, and if so, how direct is the contact with
the (rest of the) environment? Moreover, to the extent that
repair is also automated, they lack this opportunity; then
there is more distance again.
Yet while operators and supervisors are still relatively
‘‘close’’ to both the machine and material reality and may
have some know-how (e.g. through repairing the machines), the master–manager is the most alienated agent of all,
being more distant from nature and from others. Distant
from both the machines and from those who operate them
(and those who directly supervise the operators), the master
has no chance to do ‘‘work’’ in the sense of mixing with
and transforming physical and material reality. (The same
223
applies to the master–owner, who does not work but merely
owns.) The master–manager is also the most vulnerable and
dependent of all: she is dependent on the human workers–
slaves and the machine workers–slaves. She has formal
power, and power to order and prescribe to the human
slaves and the machine slaves. But if something goes
wrong with the machine or if the human slaves–workers
refuse to operate the machines, the master is in trouble
since she lacks the know-how of both kinds of slaves. And
because of this vulnerability, she is already troubled even
before something happens. As I have argued previously,
human vulnerability is not so much about an ‘‘objective’’
state of affairs or probability, but also about the experience
of vulnerability, including imagining what might happen
(Coeckelbergh 2013a, b). We are not vulnerable in the way
an object or machine is vulnerable; as human beings we
actively and consciously experience and relate to our
vulnerability.
Note that the human ‘‘worker–slave’’ should not only be
understood as referring to workers in a factory, in an
industrial setting. The term also refers to ‘‘information’’
workers in the ‘‘information society’’, who are themselves
entirely dependent on their machines—especially in the
case of increasing automation—and on the masters–managers. But again these information workers and these
managers of people and information are also highly
dependent and vulnerable on the machines they use. And
for the ‘‘master’’ this means two orders of vulnerability and
dependence: a direct dependence on the management
machines she uses such as computers (first order dependence one could say) and an indirect (second order)
dependence on the machines the slave–worker uses (industrial machines, machines related to the infrastructure
which makes possible the computing, etc.).
Note again how this picture is perhaps more complex
and ambiguous, and in any case different, than the Marxian
analysis of political economy. According to the Marx of
the Economic and Philosophic Manuscripts of 1844,
workers become alienated from their labour and from the
products of their labour. Because of division of labour,
workers could no longer determine the purpose of the
work. Workers thus become automatized when they do
‘very one-sided and machine-like type of labour’ and are
thus being ‘reduced to a machine’ (Marx 1844). This is in
line with what I said, but (1) I emphasize alienation from
the environment and the vulnerability and dependency that
come with automation, (2) I noted that workers do not
become automatized when they repair the machines, and
(3) the argument I constructed differs from Marx since it
connects this automation not only with slavery but also
with mastery: in relation to technology, most workers are
also masters—not only servants of the capitalists and
slaves of their machines, as one may think on the basis of a
123
224
(superficial?) reading of Marx. And the managers have the
same problem as the workers when it comes to dependence
on technology.
Furthermore, outside the sphere of production, the
sphere of consumption can also be understood in this way.
Consumers must be seen as spoiled masters who pay a high
price for their mastery and their comfort and convenience:
they get others to do the work and they merely consume,
but they are totally dependent on those who have mastered
nature for them, and they are highly alienated from the
products they consume and from the people and machines
which made them. For instance, in so far as food production is automated, the farmer (as master of the machine) is
already alienated from nature, but the consumer’s distance
from nature is even greater since the consumer does not
even see where the food is produced and has no ‘‘work’’
relation with it whatsoever. Moreover, when farming
becomes automated, the farmer is also alienated since her
role shifts from that of a ‘‘worker’’ (with know-how, with a
more direct relation to nature) to that of an ‘‘operator’’ and
‘‘manager’’—that is, a vulnerable and dependent master.
The farmer becomes highly dependent on her machines.
The consumer’s position, then, is one of double dependence: dependence on the farmer and, indirectly, dependence on the machines (and on the animals which were first
turned into slaves and have then been turned into machines). Separated from nature by two orders of alienation,
consumers are perhaps the most tragic masters in contemporary society. Thinking that they are the ultimate
masters who have reached the pinnacle of freedom (an
illusion communicated by advertisement), they are instead
the most dependent and vulnerable of all. They are like
very young children, but with one important difference:
young children (young children qua pre-workers) can still
gain self-consciousness through work as they grow up;
consumers (people qua consumers) have lost this
possibility.
Table 1 represents some of the analysis made here.
Conclusion
To conclude this section, it turns out that automation
technology—at least in so far as it takes over the human
action—does not render us slaves, but, rather, makes us
into vulnerable and dependent masters. Consumers, owners, and managers, but also most ‘‘workers’’ such as
operators and supervisors find themselves in a position that
is characterized by alienation from nature, lack of knowhow, and high dependence on the machines used to lessen
our dependence on nature. New vulnerabilities related to
our dependence on machines is the price we pay for our
comfortable mastery. We wanted to become more free in
relation to nature and we did not want to work, but instead
123
M. Coeckelbergh
of becoming more free and very powerful, we became
highly dependent on the slaves–machines that do the work
for us (and on those that know how to repair the machines).
It seems that if we continue to rely on automation, we will
remain in this tragic position.
Section II: Critical discussion: slaves, human–
technology relations, and moving beyond master–
slave thinking
After having articulated this argument inspired by the
master–slave dialectic, however, let me indicated some
problems and limitations of this argument and approach
with regard to assumptions it makes about (1) the term
‘‘slaves’’ and (2) about human–technology relations. At the
end of the section, however, I will indicate what we can
nevertheless learn from this way of thinking. I will also ask
the question what can kind of technologies might avoid the
problems discussed in Section I and how we can move
beyond master–slave thinking.
Criticism of arguments in Section I: Problematic
assumptions about slaves and human–technology
relations
First, the term ‘‘slave’’ is not very clear in the discussion
and needs unpacking. For instance, what is the difference
between a servant and a slave? Historically, a slave would
be owned by the master. But is this difference relevant to
the present discussion, and if so why and how? In any case,
more work is needed on the concept. Consider the very
term ‘‘slave’’. It is a term which is of course very morally–
historically loaded. As Bryson acknowledges, it is associated with ‘historical slave trade’, with ‘racism’, and with
‘endemic cruelty’ and ‘dehumanization’ (Bryson 2010).
Bryson then attempts to take distance from these associations by using the term ‘servants’ and by saying that robots
are servants ‘without being a person’. But this raises again
the question concerning the difference between ‘servant’
and ‘slave’, and makes us wonder if ‘servant’ is so much
less morally loaded. And indeed we may insist that when
we use the term ‘slaves’ as a metaphor and apply it to
machines we do not say anything about humans. However,
it is doubtful if this purification operation is successful.
When we continue to use the term ‘‘slave’’, the problematic
meanings of the term do not disappear entirely. The ghost
of historical slavery continues to show up. Do we really
want to let in these meanings by talking about ‘‘slaves’’,
even if, rationally speaking, we are indeed talking about
machines and not about humans? Moreover, it remains
doubtful if authors who make the comparison—from
Aristotle to Bryson—are right about their definition of, and
The tragedy of the master: automation, vulnerability, and distance
225
Table 1 Implications of automation for knowledge, dependency, and alienation
Human
master
Kind of knowledge
Power, dependence, vulnerability
Alienation
Lacks embodied know-how of the slave
Power but dependent on recognition by
slaves–workers, Which renders them
vulnerable
No direct relation to environment,
technology, humans
Lacks know-how of the machine
Highest degree of alienation
As a manager dependent on the workers and
machines
As a user dependent on the machines
As a consumer dependent on the machine
and on human masters (and slaves?)
Becomes ‘‘automated’’ by the machine,
risks to become a co-machine or metamachine
Human
slave
Had know-how when craftsperson but when
automation sets in becomes operator:
know-how ‘‘transferred’’ to machine
(involving loss of embodied knowledge,
which becomes formalized—thus not a
real transfer)
(May be able to regain know-how and more
direct relation through becoming repairer)
Machine
slave
(slave–
machine)
Only entity with direct relation to nature?
Entity with ‘‘know-how’’ (but not the same
as human, embodied knowledge)
As worker dependent on human manager
As a user or operator dependent on the
machines, becomes ‘‘automated’’, but also
some power over master who has even
less know-how
(As repairer more power over machine and
some power over master–manager,
master–operators, and others who lack
know-how)
Dependent on humans but also ‘‘power’’
over humans in the sense that it renders
them dependent and vulnerable
But: cannot give recognition (and cannot
receive recognition)
assumptions about, (historical) slaves. In particular, in so
far as it is assumed that historical slaves were ‘tools’, their
humanity and their agency is denied. Historical slaves were
human beings who experienced, interpreted, and resisted
their slavery. They were never ‘mere tools’ (Aristotle) or
even mere ‘servants’ (Bryson). By using the ‘‘slave’’
metaphor, then, we may affiliate our arguments too closely
with a kind of thinking—a dehumanizing and morally–
politically problematic kind of thinking—we should avoid.
Second, this way of thinking in terms of masters and
slaves has a number of significant drawbacks when it
comes to thinking about human–technology relations. First,
it reduces the range of human–technology relations to one
particular one (master–slave). There are many more possibilities of human–technology relations. For example,
rather than having its own kind of agency, technology may
extend and enhance human agency. There are also different
roles for, say, robots. The precise ways in which automation impacts on human–technology relations need further
study and cannot and should not be reduced to a ‘‘master–
slave’’ narrative. Second—and this is a more fundamental
critique—this argument is in danger of assuming that
technology is something ‘‘external’’ to the human. If we set
up ‘‘technology’’ or ‘‘the machine’’ as an external agent
Had less alienated relation to
environment, but machine
capitalism took it away, becomes
co-machine or meta-machine
(May regain less alienated relation if
moving to repair)
Only entity in direct ‘‘contact’’ with
environment?
(But not lived relation to nature, not
lived work)
which is either master or slave, then we assume that the
technology has little to do with the human and that human
subjectivity and practice is not really changed by technology. But quite the contrary is the case: technology is
made and used by humans, and shapes our subjectivity and
practices. It should not be assumed, for instance, that the
‘‘work’’ remains the same when it is automated. For
instance, the heating and cooling of a building is very
different when it takes the form of automatic air conditioning instead of older practices of regulating temperature
of a building, including heating (e.g. by means of a stove or
fireplace) and cooling (e.g. opening windows or keeping
them closed). Even if and when machines are given
‘‘agency’’, this agency can only be understood in relation to
the human. It is delegated agency, and delegation is a
human process. There is not ‘‘the machine’’ apart from the
human.
Now one could deal with the latter problems in two
ways: (a) by acknowledging the limited range and by
restricting the scope of the argument to automation, which
does typically involve a phenomenology of technological
agency, where technology comes to be experienced as
standing apart from us and sometimes appears as a quasiother—whatever other human–technology relations may be
123
226
involved—and (2) by moving towards a less dualistic view
of vulnerability, alienation, and what I have called ‘‘the
automation of the social’’. Let me briefly explain what I
mean.
First, in Human Being @ Risk (2013) I have conceptualized our relation to risk as being-at-risk and as beingvulnerable; technology may well be experienced as external but at the same time we are always related to it and
entangled with it as worldly beings who again and again
put ourselves at risk, render ourselves vulnerable, by using
technology. Part of this vulnerability also means that we
are changed and shaped by the tools we use (and of course
by our relationships with other humans). Second, technology creates new ways of alienation—yet in contrast to a
superficial reading of the master–slave dialectic I emphasize in this article that this also applies to the master—but
there is also a sense in which we are always in the world;
this kind of basis is not up to change, whatever else technology may do to us. Surely each technology may create
new or additional forms of alienation. Automation creates
its own kinds alienation, as explained in this article. But
this does not release us from our worldly condition, a
condition which Heidegger described as being-in-the-world
(Heidegger 1927). For knowledge and experience, this
implies that while there is certainly a sense in which
automation alienates us from nature, we always retain some
relation to it since, as embodied and environmental beings,
we can never be entirely alienated. We are always in the
world and in our environment; we are always already
environmental (Coeckelbergh 2015b). Third, once we
realize that technology is not so external, we can take
responsibility for it; although we should acknowledge that
we have never full control over it. With the later Heidegger
we may see modern technology, including automation
technology, as a condition which befalls us and which we
cannot simply alter. Maybe it is a kind of destiny (see for
instance Heidegger 1977).
A remaining problem, however, is that by focusing on
human–technology relations we might be blind to how
technologies such as automation, AI, and robots mediate
human–human relations. For this purpose, ‘‘master–slave’’
conceptualizations and the concept of alienation Marx
talked about might be still relevant. More generally, social
and political theory are relevant here. Automation is not
only about a kind of power struggle between humans and
technology (if this is an adequate way of thinking about
humans and technology at all; in any case there are more
types of relations—see my previous comment); it is also
about social relations between humans. Is automation in
the form of robotics and AI used to increase inequality and
injustice? By whom is it used, will it be used? If the use
and development of automation grows, who will gain and
who will lose? What will be the societal consequences?
123
M. Coeckelbergh
How does automation change and make possible different
power relations? These kinds of social and political issues
should not be neglected, and a focus on the tragedy of the
master in relation to technology may distract from these
human–human issues and from the precise role (automation) technology plays in them.
Consider, for instance, the issue of power relations in the
light of specialization and centralization. When knowledge
becomes specialized and centralized, this changes the
power relations between humans: experts get more power
(sometimes on the basis of know-that, sometimes because
of their know-how) and also certain institutions such as
corporations become more powerful since they possess the
knowledge/expertise and the machines. This process of
specialization and centralization depends on the machine,
but at the same time humans also create these new social
structures and cultures, and then use machines to implement and materialize these. Mumford suggested in Technics and Civlization (1934) that before the machine
mechanized us, humans aspired to mechanize human
labour. He claims that
Men had become mechanical before they perfected
complica...
Purchase answer to see full
attachment