Stanford University

Herbert Spencer (Stanford Encyclopedia of Philosophy)

Spencer was a social evolutionist without question but he was never
crudely social Darwinist. He was what we now refer to as a liberal
utilitarian first who traded heavily in evolutionary theory in order
to explain how our liberal utilitarian sense of justice emerges.

Though a utilitarian, Spencer took distributive justice no less
seriously than Mill. For him as for Mill, liberty and justice were
equivalent. Whereas Mill equated fundamental justice with his liberty
principle, Spencer equated justice with equal liberty, which holds
that the “liberty of each, limited by the like liberty of all,
is the rule in conformity with which society must be organized”
(Spencer, 1970: 79). Moreover, for Spencer as for Mill, liberty was
sacrosanct, insuring that his utilitarianism was equally a bona fide
form of liberalism. For both, respect for liberty also just happened
to work out for the utilitarian best all things considered.
Indefeasible liberty, properly formulated, and utility were therefore
fully compossible.

Now in Spencer’s case, especially by The Principles of Ethics
(1879–93), this compossibility rested on a complex evolutionary
moral psychology combining associationism, Lamarckian use-inheritance,
intuitionism and utility. Pleasure-producing activity has tended to
generate biologically inheritable associations between certain types
of actions, pleasurable feelings and feelings of approval. Gradually,
utilitarianism becomes
intuitive.[1]
And wherever utilitarian intuitions thrive, societies tend to be more
vibrant as well as stable. Social evolution favors cultures that
internalize utilitarian maxims intuitively. Conduct “restrained
within the required limits [stipulated by the principle of equal
freedom], calling out no antagonistic passions, favors harmonious
cooperation, profits the group, and, by implications, profits the
average of individuals.” Consequently, “groups formed of
members having this adaptation of nature” tend “to survive
and spread” (Spencer, vol. II, 1978: 43). Wherever general
utility thrives, societies thrive. General utility and cultural
stamina go hand-in-hand. And general utility thrives best where
individuals exercise and develop their faculties within the parameters
stipulated by equal freedom.

In short, like any moral intuition, equal freedom favors societies
that internalize it and, ultimately, self-consciously invoke it. And
wherever societies celebrate equal freedom as an ultimate principle of
justice, well-being flourishes and utilitarian liberalism spreads.

Spencer likewise took moral rights seriously insofar as properly
celebrating equal freedom entailed recognizing and celebrating basic
moral rights as its “corollaries.” Moral rights specify
equal freedom, making its normative requirements substantively
clearer. They stipulate our most essential sources of happiness,
namely life and liberty. Moral rights to life and liberty are
conditions of general happiness. They guarantee each individual the
opportunity to exercise his or her faculties according to his or her
own lights, which is the source of real happiness. Moral rights can’t
make us happy but merely give us the equal chance to make ourselves
happy as best we can. They consequently promote general happiness
indirectly. And since they are “corollaries” of equal
freedom, they are no less indefeasible than the principle of equal
freedom itself.

Basic moral rights, then, emerge as intuitions too though they are
more specific than our generalized intuitive appreciation of the
utilitarian prowess of equal freedom. Consequently, self-consciously
internalizing and refining our intuitive sense of equal freedom,
transforming it into a principle of practical reasoning,
simultaneously transforms our emerging normative intuitions about the
sanctity of life and liberty into stringent juridical principles. And
this is simply another way of claiming that general utility flourishes
best wherever liberal principles are seriously invoked. Moral
societies are happier societies and more vibrant and successful to
boot.

Though Spencer sometimes labels basic moral rights
“natural” rights, we should not be misled, as some
scholars have been, by this characterization. Spencer’s most sustained
and systematic discussion of moral rights occurs in the concluding
chapter, “The Great Political Superstition,” of The
Man Versus the State
(1884). There, he says that basic rights are
natural in the sense that they valorize “customs” and
“usages” that naturally arise as a way of ameliorating
social friction. Though conventional practices, only very specific
rights nevertheless effectively promote human well-being. Only those
societies, that fortuitously embrace them, flourish.

Recent scholars have misinterpreted Spencer’s theory rights because,
among other reasons, they have no doubt misunderstood Spencer’s
motives for writing The Man Versus the State. The essay is a
highly polemical protest, in the name of strong rights as the best
antidote, against the dangers of incremental legislative reforms
introducing socialism surreptitiously into Britain. Its vitriolic,
anti-socialist language surely accounts for much of its sometimes
nasty social Darwinist rhetoric, which is unmatched in Spencer’s other
writings notwithstanding scattered passages in The Principles of
Ethics
and in The Principles of Sociology
(1876–96).[2]

Spencer’s “liberal” utilitarian credentials are therefore
compelling as his 1863 exchange of letters with Mill further
testifies. Between the 1861 serial publication of
Utilitarianism in Fraser’s Magazine and its 1863
publication as a book, Spencer wrote Mill, protesting that Mill
erroneously implied that he was anti-utilitarian in a footnote near
the end of the last chapter, “Of the Connection Between Justice
and Utility.” Agreeing with Benthamism that happiness is the
“ultimate” end, Spencer firmly disagrees that it should be
our “proximate” end. He next adds:

But the view for which I contend is, that Morality properly so-called
– the science of right conduct – has for its object to
determine how and why certain modes of conduct are detrimental, and
certain other modes beneficial. These good and bad results cannot be
accidental, but must be necessary consequences of the constitution of
things; and I conceive it to be the business of moral science to
deduce, from the laws of life and the conditions of existence, what
kinds of action necessarily tend to produce happiness, and what kinds
to produce unhappiness. Having done this, its deductions are to be
recognized as laws of conduct; and are to be conformed to irrespective
of a direct estimation of happiness or misery (Spencer, vol. II, 1904:
88–9).[3]

Specific types of actions, in short, necessarily always promote
general utility best over the long term though not always in the
interim. While they may not always promote it proximately, they
invariably promote it ultimately or, in other words, indirectly. These
action types constitute uncompromising, normative “laws of
conduct.” As such, they specify the parameters of equal freedom.
That is, they constitute our fundamental moral rights. We have moral
rights to these action types if we have moral rights to anything at
all.

Spencer as much as Mill, then, advocates indirect utilitarianism by
featuring robust moral rights. For both theorists, rights-oriented
utilitarianism best fosters general happiness because individuals
succeed in making themselves happiest when they develop their mental
and physical faculties by exercising them as they deem most
appropriate, which, in turn, requires extensive freedom. But since we
live socially, what we practically require is equal freedom suitably
fleshed out in terms of its moral right corollaries. Moral rights to
life and liberty secure our most vital opportunities for making
ourselves as happy as we possibly can. So if Mill remains potently
germane because his legacy to contemporary liberal utilitarian still
inspires, then we should take better account of Spencer than,
unfortunately, we currently do.

Spencer’s “liberal” utilitarianism, however, differs from
Mill’s in several respects, including principally the greater
stringency that Spencer ascribed to moral rights. Indeed, Mill
regarded this difference as the fundamental one between them. Mill
responded to Spencer’s letter professing allegiance to utilitarianism,
observing that he concurs fully with Spencer that utilitarianism must
incorporate the “widest and most general principles” that
it possibly can. However, in contrast to Spencer, Mill protests that
he “cannot admit that any of these principles are necessary, or
that the practical conclusions which can be drawn from them are even
(absolutely) universal” (Duncan, ed., 1908:
108).[4]

Spencer referred to his own brand of utilitarianism as
“rational” utilitarianism, which he claimed improved upon
Bentham’s inferior “empirical” utilitarianism. And though
he never labeled Mill a “rational” utilitarian, presumably
he regarded him as one.

One should not underestimate what “rational”
utilitarianism implied for Spencer metaethically. In identifying
himself as a “rational” utilitarian, Spencer distanced
himself decidedly from social Darwinism, showing why Moore’s infamous
judgment was misplaced. Responding to T. H. Huxley’s accusation that
he conflated good with “survival of the fittest,” Spencer
insisted that “fittest” and “best” were not
equivalent. He agreed with Huxley that though ethics can be
evolutionarily explained, ethics nevertheless preempts normal struggle
for existence with the arrival of humans. Humans invest evolution with
an “ethical check,” making human evolution qualitatively
different from non-human evolution. “Rational”
utilitarianism constitutes the most advanced form of “ethical
check[ing]” insofar as it specifies the “equitable limits
to his [the individual’s] activities, and of the restraints which must
be imposed upon him” in his interactions with others (Spencer,
vol. I, 1901:
125–28).[5]
In short, once we begin systematizing our inchoate utilitarian
intuitions with the principle of equal freedom and its derivative
moral rights, we begin “check[ing]” evolutionary struggle
for survival with unprecedented skill and subtlety. We
self-consciously invest our utilitarianism with stringent liberal
principles in order to advance our well-being as never before.

Now Henry Sidgwick seems to have understood what Spencer meant by
“rational” utilitarianism better than most, although
Sidgwick didn’t get Spencer entirely right either. Sidgwick engaged
Spencer critically on numerous occasions. The concluding of Book II of
The Methods of Ethics (1907), entitled “Deductive
Hedonism,” is a sustained though veiled criticism of
Spencer.[6]

For Sidgwick, Spencer’s utilitarianism was merely seemingly deductive
even though it purported to be more scientific and rigorously rational
than “empirical” utilitarianism. However, deductive
hedonism fails because, contrary to what deductive hedonists like
Spencer think, no general science of the causes of pleasure and pain
exists, insuring that we will never succeed in formulating universal,
indefeasible moral rules for promoting happiness. Moreover, Spencer
only makes matters worse for himself in claiming that we can
nevertheless formulate indefeasible moral rules for hypothetically
perfectly moral human beings. First of all, in Sidgwick’s view, since
we can’t possibly imagine what perfectly moral humans would look like,
we could never possibly deduce an ideal moral code of
“absolute” ethics for them. Secondly, even if we could
somehow conceptualize such a code, it would nevertheless provide
inadequate normative guidance to humans as we find them with all their
actual desires, emotions and irrational
proclivities.[7]
For Sidgwick, all we have is utilitarian common-sense, which we can,
and should, try to refine and systematize according the demands of our
changing
circumstances.[8]

Sidgwick, then, faulted Spencer for deceiving himself in thinking that
he had successfully made “empirical” utilitarianism more
rigorous by making it deductive and therefore “rational.”
Rather, Spencer was simply offering just another variety of
“empirical” utilitarianism instead. Nevertheless,
Spencer’s version of “empirical” utilitarianism was much
closer to Sidgwick’s than Sidgwick recognized. Spencer not only
shadowed Mill substantively but Sidgwick methodologically.

In the preface to the sixth edition of The Methods of Ethics
(1901), Sidgwick writes that as he became increasingly aware of the
shortcomings of utilitarian calculation, he became ever more sensitive
to the utilitarian efficacy of common sense “on the ground of
the general presumption which evolution afforded that moral sentiments
and opinions would point to conduct conducive to general
happiness…” (Sidgwick, 1907: xxiii). In other words,
common sense morality is a generally reliable, right-making decision
procedure because social evolution has privileged the emergence of
general happiness-generating moral sentiments. And whenever common
sense fails us with conflicting or foggy guidance, we have little
choice but to engage in order-restoring, utilitarian calculation. The
latter works hand-in-glove with the former, forever refining and
systematizing it.

Now Spencer’s “empirical” utilitarianism works much the
same way even though Spencer obfuscated these similarities by
spuriously distinguishing between “empirical” and
supposedly superior, “rational” utilitarianism. Much like
Sidgwick, Spencer holds that our common sense moral judgments derive
their intuitive force from their proven utility-promoting power
inherited from one generation to the next. Contrary to what
“empirical” utilitarians like Bentham have mistakenly
maintained, we never make utilitarian calculations in an
intuition-free vacuum. Promoting utility is never simply a matter of
choosing options, especially when much is at stake, by calculating and
critically comparing utilities. Rather, the emergence of utilitarian
practical reasoning begins wherever our moral intuitions breakdown.
Moral science tests and refines our moral intuitions, which often
prove “necessarily vague” and contradictory. In order to
“make guidance by them adequate to all requirements, their
dictates have to be interpreted and made definite by science; to which
end there must be analysis of those conditions to complete living
which they respond to, and from converse with which they have
arisen.” Such analysis invariably entails recognizing the
happiness of “each and all, as the end to be achieved by
fulfillment of these conditions” (Spencer, vol. I, 1978:
204).

“Empirical” utilitarianism is “unconsciously
made” out of the “accumulated results of past human
experience,” eventually giving way to “rational”
utilitarianism which is “determined by the intellect”
(Spencer, 1969: 279 ff.). The latter, moreover, “implies
guidance by the general conclusions which analysis of experience
yields,” calculating the “distant effects” on lives
“at large” (Spencer, 1981: 162–5).

In sum, “rational” utilitarianism is critical and
empirical rather than deductive. It resolutely though judiciously
embraces indefeasible moral rights as necessary conditions of general
happiness, making utilitarianism rigorously and uncompromisingly
liberal. And it was also evolutionary, much like Sidgwick’s. For both
Spencer and Sidgwick, utilitarian practical reasoning exposes, refines
and systematizes our underlying moral intuitions, which have thus far
evolved in spite of their under-appreciated utility. Whereas Spencer
labeled this progress towards “rational” utilitarianism,
Sidgwick more appropriately called this “progress in the
direction of a closer approximation to a perfectly enlightened
[empirical] Utilitarianism” (Sidgwick, 1907: 455).

Notwithstanding the undervalued similarities between their respective
versions of evolutionary utilitarianism, Spencer and Sidgwick
nevertheless parted company in two fundamental respects. First,
whereas for Spencer, “rational” utilitarianism refines
“empirical” utilitarianism by converging on indefeasible
moral rights, for Sidgwick, systematization never ceases. Rather,
systematizing common sense continues indefinitely in order to keep
pace with the vicissitudes of our social circumstances. The best
utilitarian strategy requires flexibility and not the cramping
rigidity of unyielding rights. In effect, Spencer’s utilitarianism was
too dogmatically liberal for Sidgwick’s more tempered political
tastes.

Second, Spencer was a Lamarckian while Sidgwick was not. For Spencer,
moral faculty exercise hones each individual’s moral intuitions. Being
biologically (and not just culturally) inheritable, these intuitions
become increasingly authoritative in succeeding generations, favoring
those cultures wherever moral common sense becomes more uncompromising
all things being equal. Eventually, members of favored societies begin
consciously recognizing, and further deliberately refining, the
utility-generating potency of their inherited moral intuitions.
“Rational,” scientific utilitarianism slowly replaces
common sense, “empirical” utilitarianism as we learn the
incomparable value of equal freedom and its derivative moral rights as
everyday utilitarian decision
procedures.[9]

Notwithstanding their differences, Spencer was nonetheless as much a
utilitarian as Sidgwick, which the latter fully recognized though we
should hesitate labeling Spencer a classical utilitarian as we now
label Sidgwick. Moreover, Sidgwick was hardly alone at the turn of the
twentieth-century in depicting Spencer as fundamentally utilitarian.
Even scholars in Germany at that time read Spencer as a utilitarian.
For instance, A. G. Sinclair viewed him as a utilitarian worth
comparing with Sidgwick. In his 1907 Der Utilitarismus bei
Sidgwick und Spencer
, Sinclair concludes “Daher ist er
[Spencer], wie wir schon gesagt haben, ein evolutionistischer Hedonist
und nicht ein ethischer Evolutionist,” which we can translate as
“Therefore he (Spencer) is, as we have already seen, an
evolutionary hedonist and not an ethical evolutionist”
(Sinclair, 1907: 49). So however much we have fallen into the
erroneous habit of regarding Spencer as little invested with
19th-century utilitarianism, he was not received that way at all by
his immediate contemporaries both in England and in continental
Europe.

Not only was Spencer less than a “social Darwinist” as we
have come to understand social Darwinism, but he was also less
unambiguously libertarian as some, such as Eric Mack and Tibor Machan,
have made him out to be. Not only his underlying utilitarianism but
also the distinction, which he never forswears, between “rights
properly so-called” and “political” rights, makes it
problematic to read him as what we would call a
‘libertarian’.

Whereas “rights properly so-called” are authentic
specifications of equal freedom, “political rights” are
not. They are interim devices conditional on our moral imperfection.
Insofar as we remain morally imperfect requiring government
enforcement of moral rights proper, political rights insure that
government nevertheless remains mostly benign, never unduly violating
moral rights proper themselves. The “right to ignore the
state” and the right of universal suffrage are two essential
political rights for Spencer. In Social Statics, Spencer says
“we cannot choose but admit the right of the citizen to adopt a
condition of voluntary outlawry.” Every citizen is “free
to drop connection with the state – to relinquish its protection
and to refuse paying for its support” (Spencer, 1970: 185). For
Spencer, this right helps restrict government to protecting proper
moral rights because it allows citizens to take their business
elsewhere when it doesn’t.

However, Spencer eventually repudiated this mere political
right.  For instance, in his 1894 An Autobiography, he
insists that since citizens “cannot avoid benefiting by the
social order which government maintains,” they have no right to
opt out from its protection (Spencer, 1904, vol. 1: 362). They may not
legitimately take their business elsewhere whenever they feel that
their fundamental moral rights are being ill-protected. Because he
eventually repudiated the “right to ignore the state,” we
should refrain from uncritically regarding him as anticipating
Nozick.

Spencer’s commitment to the right of universal suffrage likewise wanes
in his later writings. Whereas in Social Statics, he regards
universal suffrage as a dependable means of preventing government from
overreaching its duty of sticking to protecting moral rights proper,
by the later Principles of Ethics he concludes that universal
suffrage fails to do this effectively and so he abandons his support
of it. He later concluded that universal suffrage threatened respect
for moral rights more than it protected them.  Universal
suffrage, especially when extended to women, encouraged
“over-legislation,” allowing government to take up
responsibilities which were none of its business.

Spencer, then, was more than willing to modify political rights in
keeping with his changing assessment of how well they secured basic
moral rights on whose sanctity promoting happiness depended. The more
he became convinced that certain political rights were accordingly
counterproductive, the more readily he forsook them and the less
democratic, if not patently libertarian, he became.

Likewise, Spencer’s declining enthusiasm for land nationalization
(which Hillel Steiner has recently found so inspiring), coupled with
growing doubts that it followed as a corollary from the principle of
equal freedom, testify to his waning
radicalism.[10]
According to Spencer in Social Statics, denying every
citizen the right to use of the earth equally was a “crime
inferior only in wickedness to the crime of taking away their lives or
personal liberties” (Spencer: 1970, 182.)  Private land
ownership was incompatible with equal freedom because it denied most
citizens equal access to the earth’s surface on which faculty exercise
and happiness ultimately depended. However, by The Principles of
Ethics
, Spencer abandoned advocating comprehensive land
nationalization, much to Henry George’s ire. George, an American, had
previously regarded Spencer as a formidable ally in his crusade to
abolish private land tenure.

Now Spencer’s repudiation of the moral right to use the earth and the
political right to ignore the state, as well as the political right of
universal suffrage, undermines his distinction between rational and
empirical utilitarianism. In forswearing the right to use the earth
— because he subsequently became convinced that land
nationalization undermined, rather than promoted general utility
— Spencer betrays just how much of a traditional empirical
utilitarian he was. He abandoned land nationalization not because he
concluded that the right to use the earth did not follow deductively
from the principle of equal freedom. Rather, he abandoned land reform
simply because he became convinced that it was an empirically
counterproductive strategy for promoting utility.

Even more obviously, by repudiating political rights like the
“right to ignore the state” and universal suffrage rights,
he similarly divulged just how much empirical utilitarian
considerations trumped all else in his practical reasoning. Not only
was Spencer not a committed or consistent libertarian, but he was not
much of rational utilitarian either. In the end, Spencer was mostly,
to repeat, what we would now call a liberal utilitarian who, much like
Mill, tried to combine strong rights with utility though, in Spencer’s
case, he regarded moral rights as indefeasible.

Allan Gibbard has suggested that, for Sidgwick, in refining and
systematizing common sense, we transform “unconscious
utilitarianism” into “conscious utilitarianism.” We
“apply scientific techniques of felicific assessment to further
the achievement of the old, unconscious goal” (Gibbard in Miller
and Williams, eds., 1982: 72). Spencer’s “liberal”
utilitarianism was comparable moral science. Sidgwick, however, aimed
simply at “progress in the direction of a closer approximation
to a perfectly enlightened Utilitarianism” (Sidgwick, 1907:
455). Spencer, by contrast, had more grandiose aspirations for
repairing utilitarianism. Merely moving towards “perfectly
enlightened Utilitarianism” was scientifically under ambitious.
Fully “enlightened” utilitarianism was conceptually
accessible and perhaps even politically practicable. And Spencer had
discovered its secret, namely indefeasible moral rights.

Spencer, then, merits greater esteem if for no other reason than that
Sidgwick, besides Mill, took him so seriously as a fellow utilitarian
worthy of his critical attention. Unfortunately, contemporary
intellectual history has been less kind, preferring a more convenient
and simplistic narrative of the liberal canon that excludes him.

Spencer’s “liberal” utilitarianism was bolder and arguably
more unstable than either Mill or Sidgwick’s. He followed Mill
investing utilitarianism with robust moral rights hoping to keep it
ethically appealing without forgoing its systemic coherence. While the
principle of utility retreats to the background as a standard of
overall normative assessment, moral rights serve as everyday sources
of direct moral obligation, making Spencer no less an indirect
utilitarian than Mill. But Spencer’s indirect utilitarianism is more
volatile, more logically precarious, because Spencer burdened rights
with indefeasibility while Mill made them stringent but nevertheless
overridable depending on the magnitude of the utility at stake. For
Spencer, we never compromise basic rights let the heavens fall. But
for Mill, the prospect of collapsing heavens would easily justify
appealing directly to the principle of utility at the expense of
respect for moral rights.

Now, critics of utilitarianism from William Whewell (1794–1866)
to David Lyons more recently have taken Mill and subsequent liberal
utilitarians to task for trying to have their utilitarian cake and eat
their liberalism too. As Lyons argues with great effect, by imposing
liberal juridical constraints on the pursuit of general utility, Mill
introduces as a second normative criterion with independent
“moral force” compromising his utilitarianism. He risks
embracing value pluralism if not abandoning utilitarianism altogether.
And if Mill’s liberal version of utilitarianism is just value
pluralism in disguise, then he still faces the further dilemma of how
to arbitrate conflicts between utility and rights. If utility trumps
rights only when enough of it is at stake, we must still ask how much
enough is enough? And any systematic answer we might give simply
injects another normative criterion into the problematic logic of our
liberal utilitarian stew since we have now introduced a third higher
criterion that legislates conflicts between the moral force of the
principle of utility and the moral force of
rights.[11]

If these dilemmas hold for Mill’s utilitarianism, then the
implications are both better and worse for Spencer. Though for Mill,
utility always trumps rights when enough of the former is in jeopardy,
with Spencer, fundamental rights always trump utility no matter how
much of the latter is imperiled. Hence, Spencer does not need to
introduce surreptitiously supplemental criteria for adjudicating
conflicts between utility and rights because rights are indefeasible,
never giving way to the demands of utility or disutility no matter how
immediate and no matter how promising or how catastrophic. In short,
for Spencer, basic moral rights always carry the greater, practical
(if not formal) moral force. Liberalism always supersedes
utilitarianism in practice no matter how insistently Spencer feigns
loyalty to the latter.

Naturally, one can salvage this kind of utilitarianism’s authenticity
by implausibly contending that indefeasible moral rights always
(meaning literally without exception) work out for the utilitarian
best over both the short and long-terms. As Wayne Sumner correctly
suggests, “absolute rights are not an impossible output for a
consequentialist methodology” (Sumner, 1987: 211). While this
maneuver would certainly rescue the logical integrity of Spencer’s
liberal version of utilitarianism, it does so at the cost of
considerable common sense credibility. And even if it were
miraculously true that respecting rights without exception just
happened to maximize long-term utility, empirically demonstrating this
truth would certainly prove challenging at best. Moreover,
notwithstanding this maneuver’s practical plausibility, it would
nevertheless seem to cause utilitarianism to retire a “residual
position” that is indeed hardly “worth calling
utilitarianism” (Williams in Smart and Williams, 1973: 135).

Whether Spencer actually envisioned his utilitarianism this way is
unclear. In any case, insofar as he also held that social evolution
was tending towards human moral perfectibility, he could afford to
worry less and less about whether rights-based utilitarianism was a
plausible philosophical enterprise. Increasing moral perfectibility
makes secondary decision procedures like basic moral rights
unnecessary as a utility-promoting strategy. Why bother with promoting
general utility indirectly once we have learned to promote it directly
with certainty of success? Why bother with substitute sources of
stand-in obligation when, thanks to having become moral saints, act
utilitarianism will fortunately always do? But moral perfectibility’s
unlikelihood is no less plausible than the likelihood of fanatical
respect for basic moral rights always working out for the utilitarian
best.[12]
In any case, just as the latter strategy causes utilitarianism to
retire completely for practical purposes, so the former strategy
amounts to liberalism entirely retiring in turn. Hence, Mill’s version
of “liberal” utilitarianism must be deemed more compelling
and promising for those of us who remain stubbornly drawn to this
problematical philosophical enterprise.

Spencer’s rights-based utilitarianism nonetheless has much to
recommend for it despite its unconventional features and implausible
implications. Even more than Mill, he suggests how liberal
utilitarians could attempt to moderate utilitarianism in other ways,
enabling it to retain a certain measure of considerable ethical
appeal. Spencer’s utilitarianism wears its liberalism not only by
constraining the pursuit of utility externally by deploying robust
moral rights with palpable independent moral force. It also, and more
successfully, shows how utilitarians can liberalize their
utilitarianism by building internal constraints into their maximizing
aims. If, following Spencer, we make our maximizing goal
distribution-sensitive by including everyone’s happiness within it so
that each individual obtains his or her fair share, then we have
salvaged some kind of consequentialist authenticity while
simultaneously securing individual integrity too. We have salvaged
utilitarianism as a happiness-promoting, if not a
happiness-maximizing, consequentialism. Because everyone is “to
count for one, nobody for more than one” not just as a resource
for generating utility but also as deserving to experience a share of
it, no one may be sacrificed callously without limit for the good of
the
rest.[13]
No one may be treated as a means only but must be treated as an end
as well.

Spencer’s utilitarianism also has much to recommend for it simply for
its much undervalued importance in the development of modern
liberalism. If Mill and Sidgwick are critical to making sense of our
liberal canon, then Spencer is no less critical. If both are crucial
for coming to terms with Rawls particularly, and consequently with
post-Rawlsianism generally, as I strongly believe both are, then
Spencer surely deserves better from recent intellectual history.
Intellectual history is one of the many important narratives we tell
and retell ourselves. What a shame when we succumb to scholarly
laziness in constructing these narratives just because such laziness
both facilitates meeting the pedagogical challenges of teaching the
liberal tradition and answering our need for a coherent philosophical
identity.

Follow us