I. Introduction

 In 1981 the Anti-Defamation League of B’nai B’rith published the model for what have
now become known as hate crimes laws.  Hate crimes laws are “penalty enhancement” laws.
They enhance the penalties for a variety of crimes (most commonly assault, homicide, arson,
trespass and vandalism) provided certain triggering conditions are met.  Some hate crimes laws
apply to specified crimes any time the perpetrator selects his victim because of her race,
ethnicity, religion, sexual orientation or other specified characteristic.  Others apply to specified
crimes only if the perpetrator selects his victim out of a special animus toward a racial, ethnic,
religious, or other specified group to which she belongs.  So, a law of the first sort but not the
second would apply to a criminal who commits one of the specified crimes against a Black
victim not out of any particular hostility toward Blacks as a racial group, and not randomly, but
because he believes the police less likely to investigate crimes committed against Blacks and
thinks judges sentencing those convicted of such crimes less likely to impose harsh sentences.

 Hate crimes laws are now common in the United States; they have even worked their way
into federal law.  Notwithstanding significant public and academic support, these laws have
been subject to strong, serious and persistent criticism.  And much of it is well-deserved, for the
arguments standardly made in favor of hate crimes laws are either unsound or weak.
Demonstrating as much is the first aim of this paper.

 Its second aim, however, is to offer a non-standard argument for hate crimes laws that is
neither unsound nor weak.  This argument I call the argument from oppression.  It captures and
expresses better than any of the standard arguments the moral basis for the sentiments or
intuitions many people have in favor of hate crimes laws, and the objections to the standard
arguments do not apply to it.  But one objection (or perhaps it is a family of objections) must still
be addressed.  This objection posits hate crimes laws (along with, perhaps, affirmative action
policies, the expansion of sexual harassment laws, slavery reparations initiatives and the like) as
the balkanizing result of an identity or interest group politics of resentment inconsistent with
progressive liberal democratic values.  This paper’s third aim is to assess whether this objection
is telling against hate crimes laws as justified by and revised suitably in light of the argument
from oppression.  At the risk of ruining an otherwise suspenseful read, a bit of foreshadowing
would seem apropos:  I argue that the objection is not telling when directed at hate crimes laws
justified by and revised suitably in light of the argument from oppression.

II.  The Standard Arguments.

 There are three standard arguments for hate crimes laws.  The first is the argument from
greater harm.  The second is the argument from more culpable mental states.  The third is the
argument from liberal democratic values.  Each of these arguments is either unsound or weak.

 A. The argument from greater harm.

 Whether crimes satisfying the trigger conditions of hate crimes laws cause greater harm,
physical or psychological, either to victims or non-victim third parties, than similar but non-hate
crimes is, of course, an empirical question.  This is often forgotten and a priori pronouncements
regarding the greater harmfulness of hate crimes are not uncommon.  Unhappily, as prima facie
plausible as these pronouncements may be, they are not supported by the available empirical
evidence.

 Crimes said to be hate crimes do often cause physical harm to their direct and immediate
victims.  But that is because they are very often assaults and homicides, not because they are
committed by perpetrators who satisfy the relevant statutory trigger conditions for hate crimes
penalty enhancements.  As a class the crimes that satisfy these conditions are no more violent or
physically harmful to their victims than the class that do not.  And, in any event, existing laws
already scale punishment to reflect the nature and degree of physical harm of assaults and
homicides to victims.

 Because there is little evidence to support the claim that crimes falling within the scope of
hate crimes laws are more physically harmful to victims (or, for that matter, to non-victim third
parties) than those otherwise similar crimes that do not, the argument is often made that these
crimes, those that fall within the scope of hate crimes laws, cause greater psychological harm to
their direct and immediate victims, or to non-victim third parties, than those that do not.  This
greater psychological harm justifies, so the argument goes, the additional punishment imposed by
hate crimes laws.

 Crimes said to be hate crimes do cause psychological harm both to their direct and
immediate victims and to non-victim third parties.  But so do virtually all crimes.  The question
is whether candidate hate crimes cause greater psychological harm than otherwise similar non-
hate crimes.  With respect to direct and immediate victims, the available empirical evidence
suggests that they do not.  Indeed, it suggests that candidate hate crimes may cause their direct
and immediate victims marginally less psychological harm than similar crimes that would not
count as hate crimes under typical hate crimes laws.

 This evidence is admittedly counter-intuitive.  There is, consequently, a temptation to
dismiss it in favor of arm-chair psychological speculation.  But the evidence ought not be
dismissed (although pending further studies it ought to be regarded as tentative).  Most candidate
hate crimes are committed against persons belonging to historically oppressed groups: Blacks,
gays, Jews, ethnic minorities and the like.  It is not implausible to suppose that many of these
victims will possess preferences and other psychological mechanisms adapted to their condition,
and that these may mitigate their subjective experience of psychological distress when made
victim to a candidate hate crime.  It is also not implausible to suppose that the nature of many
candidate hate crimes will make it clear to the victim that she was targeted because of her race, or
sexual orientation, or religion, etc., thus giving her a way of making a kind of sense of the crime
committed against her she would not be able to make were she randomly victimized.  These
suppositions, admittedly unconfirmed, are sufficiently plausible as explanations of the admittedly
counter-intuitive empirical evidence regarding the psychological harm caused by candidate hate
crimes to their direct and immediate victims that that evidence ought not be simply dismissed.

 As with direct and immediate victims, there is little evidence to support the claim that
candidate hate crimes cause non-victim third parties greater psychological harm than similar
crimes that fall outside the scope of hate crimes laws.  The psychological harm of any crime to
non-victim third parties is determined primarily by the proximity and visibility of the crime to
such parties.  This, of course, does not preclude additional psychological harms to non-victim
third parties arising from the fact that the victim was selected because of or out of animosity
toward her membership in a racial, religious, ethnic or other specified group.  Or, at least it does
not preclude such additional psychological harms when the crime visibly manifests an animosity
toward a group with which the non-victim third parties strongly self-identify.  While there is no
substantial body of empirical evidence to confirm the claim that candidate hate crimes cause such
harms to non-victim third parties, the claim (that obvious gay-bashings “terrorize” the gay
community generally, etc.) is prima facie plausible.  (Of course, even if there were empirical
evidence confirming the claimed psychological harms, it would still be necessary to argue that
those harms are serious enough to justify the relevant penalty enhancements.)  It should not
surprise, then, that this argument has been made to do much of the justificatory work in the case
for hate crimes laws.  Given the absence of compelling empirical evidence establishing the
claimed psychological harms, however, it is at present a weak argument.  It is also just the sort of
argument that invites the characterization of hate crimes laws as the fruit of a  balkanizing,
identity group, politics of resentment.  But that is a matter the discussion of which must be
temporarily postponed in the interest of completing a review of the standard arguments.  The
upshot here is that the best version of the argument from greater harm is the argument from
psychological harms to non-victim third parties, and that that argument is, as things stand, not
very compelling.

 There is, however, an additional difficulty with the argument from psychological harms to
non-victim third parties that bears mentioning here.  And that is that when we think about the
mechanism through which such psychological harms are said to arise, we can see that it is not the
harms themselves that explain the moral intuition or sentiment in favor of enhanced penalties for
hate crimes.  It is rather the membership of the victim in an oppressed group, or, better, the
special moral wrong done by targeting for criminal conduct persons already disproportionately
vulnerable to harm, whether by virtue of belonging to an oppressed group or for some other
reason.

 If hate crimes cause greater psychological harm to non-victim third parties than non-hate
crimes, it is because the cause psychological harms (e.g., fear or terror) to persons who identify
strongly through group membership with the victim and thus feel themselves to be vicariously
targeted by the crime.  This sort of self-identification through group membership is not
uncommon among members of oppressed groups.  And thus it is not implausible to assert that
paradigm hate crimes cause greater psychological harms to non-victim third parties, and that
these harms justify enhancing the penalties for such crimes.  After all, paradigm hate crimes are
committed against members of oppressed groups.

 But the members of oppressed groups do not always self-identify through group
membership.  While self-identification through group membership is common within oppressed
groups, it is not necessary or inevitable.  It is usually the contingent result of strategic political
resistance to oppression.  Only through concrete political efforts have Blacks, women and gays,
for example, produced such a group-based self-identification as a political resource.  There was a
time when few members of each group would have identified themselves primarily through their
group membership.  Yet, Blacks, women and gays were each oppressed groups (and thus
disproportionately vulnerable to certain social harms) prior to and independent of members self-
identifying through group membership.   And it is this deeper more basic fact about paradigm
hate crimes, that they are selectively committed against members of oppressed groups, that
generates the moral intuitions or sentiments in favor of penalty enhancements, not the contingent
if often concommitant fact of psychological harm to non-victim third parties.

 This can be seen by considering two alternative sorts of cases.  The first involves crimes
committed selectively against the members of a group who strongly self-identify through group
membership but who are not oppressed and thus disproportionately vulnerable to social harm.
Suppose (what may in fact be true) that many or most prostitutes self-identify with other
prosititutes and that those who serially attack prostitutes specifically cause special psychological
harms (e.g., fear and terror) to non-victim third party prostitutes.  Do such crimes merit enhanced
penalties as hate crimes?  Few will have the intuition or feel the sentiment that they do.  Or,
suppose a criminal who serially and selectively attacks only DeadHeads (the legendary followers
of the Grateful Dead), a group the members of which strongly self-identify through group
membership.  Does he commit a hate crime?  Again, few if any will think he does.  In both cases
it is likely that the crimes will cause special psychological harms to non-victim third parties.
And these special harms may be noteworthy at the time of sentencing as an aggravating factor of
some import.  But they are not harms sufficient in moral weight to justify the non-discretionary
legislative imposition of enhanced penalties through hate crimes laws.  This is one important
reason why hate crimes advocates do not propose statutory language covering such cases.

 The second sort of case involves crimes committed selectively against members of a
group disproportionately vulnerable to social harm but not oppressed who do not self-identify
through group membership.  Such crimes cannot cause the relevant special psychological harms
to non-victim third parties.  But like crimes committed selectively against members of oppressed
groups they generate powerful intuitions or sentiments in favor of penalty enhancement.
Consider, for example, crimes committed selectively against the cognitively impaired.  Such
crimes generate strong moral intuitions or sentiments in favor of penalty enhancement.  But not
because of psychological harms to non-victim third parties, but rather because of the
disproportionate vulnerability to harm of the cognitively impaired and the additional moral
wrong done by those who specifically target them for criminal conduct.  In terms of the moral
logic in favor of penalty enhancement, these cases present a closer analogy to paradigm hate
crimes than do those selectively committed against the members of non-oppressed groups who
nevertheless strongly self-identify through group-membership.

 Crimes committed selectively against Jews in the United States present an interesting
difficulty here.  Jews self-identify through group membership for both religious reasons and
reasons rooted in resistance to a history of oppression.  It is quite plausible to assert that crimes
committed selectively against Jews in the United States cause greater psychological harm to non-
victim third parties than similar crimes not so selectively committed.  But there is a question as to
whether we should assimilate such crimes to the class of crimes selectively committed against,
say, Blacks and other groups currently oppressed (the members of which may or may not self-
identify strongly through group membership) or to the class of crimes selectively committed
against, say, DeadHeads and other groups the members of which self-identify through, but are
not oppressed by virtue of, group-membership.  This question turns, I think, on the empirical
question of whether Jews constitute an oppressed group today in the United States.  If they do,
then crimes committed selectively against Jews belong in the same category as those committed
selectively against Blacks.  If they do not, then they belong in the same category as crimes
committed against the members of other non-oppressed groups that nevertheless strongly self-
identify through group membership.

 What these cases show, I think, is that even if, or where, paradigm hate crimes cause
greater psychological harm to non-victim third parties, it is not the greater psychological harms
that justifies the penalty enhancements imposed by hate crimes laws.  It is rather the special
moral wrong done by those who selectively target members of oppressed groups.  But this
argument must wait for further development.  There are other more standard arguments still
waiting to be examined first.

 B. The argument from more culpable mental states.

 The criminal law in the United States already scales punishment according to whether an
offender acts intentionally, knowingly, recklessly or negligently.  Thus, if hate crimes laws are to
be justified by appeal to the more culpable mental states of those who commit crimes satisfying
their trigger conditions, it must be that those who commit such crimes act from a more culpable
mental state than those who commit crimes otherwise identical but outside the scope of such
laws.

 This is straightforwardly implausible with respect to those hate crimes laws that apply to
perpetrators who simply select their victims because of their race, or religion, or ethnicity, or
sexual orientation, etc.  Recall that hate crimes laws of this sort would apply to a criminal who
selects only Protestant victims because he believes only Protestants will in fact enjoy the sort of
salvation that might redeem their suffering here on earth.  He selects his victims because of their
religion.  But does he act from a mental state more culpable than the criminal who randomly
selects his  victims just because he likes the experience of anonymous power, or the criminal who
carefully selects his victims based on who is most likely to suffer the greatest from his crime?
Obviously not.

 To be sure, it may be that some criminals who commit crimes against victims selected
because of their race, or religion, or ethnicity, etc., act from mental states more culpable than
those who commit otherwise identical crimes against victims selected for other reasons.  But this
is not true of all criminals who commit crimes against victims selected for such reasons.  It
follows that where the mental state of a criminal who selects his victim because of her race, or
religion, or ethnicity is in fact more culpable than it would be had he selected his victim for some
other reason, the additional culpability must be a function of something beyond the fact that he
selected his victim because of her race, or religion, or ethnicity, etc.  What that something else is
I will turn to later.  For now, it is enough to note that the argument from more culpable mental
states fails as an argument for those hate crimes laws that enhance penalties whenever the
criminal selects his victim because of her race, or religion, or ethnicity, or sexual orientation, etc.

 The argument from more culpable mental states is perhaps more plausible as a
justification for those hate crimes laws that apply to perpetrators who select their victims out of
one or another specified group-based animosity.  After all, to act from racist, anti-Semitic,
homophobic, or xenophobic hatred or animosity is clearly to act from a very culpable mental
state, indeed a mental state more culpable than many still significantly culpable alternatives.  And
this greater culpability can be explained by reference to the evil of targeting individuals solely
out of a group-based animosity.

 But is it obvious that the mental state of a criminal who selects his victims out of racist
animosity is more culpable than that of the criminal who selects his victims out of a desire to see
the weak suffer, or to impose the greatest harm possible on a non-victim third party, or to display
his superiority to the ordinary run of humanity, or to salve an ego all to easily bruised.  That is
less clear.  Indeed, once adequate attention is given to the range of vicious and evil reasons that
lead persons to commit crimes against others, it is not obvious that racist, anti-Semitic,
homophobic or xenophobic reasons are significantly more vicious or evil than other familiar
reasons criminals have for selecting victims which the law largely ignores when it comes to
scaling punishment for crimes.   In fact, the whole idea of correctly scaling criminal punishments
to reflect the culpability of the reasons for which the criminal selected his victim looks to be
beyond the reach of ordinary human abilities once the full range of reasons that lead people to
commit crimes against particular victims is fully in view.  It is, perhaps, no accident that for the
most part the criminal law has limited its inquiry into the culpability of mental states to whether
the act in question was done intentionally, knowingly, recklessly or negligently.

 Like the argument from psychological harms to non-victim third parties, the argument
from the greater culpability of racist, anti-Semitic, homophobic, xenophobic or similar mental
states has been pressed into active and regular service by proponents of hate crimes laws.  But it
too is a weak argument.  The problem is not that an empirical premise remains unconfirmed, but
rather that an axiological premise – that to select a victim out of racist animosity, say, is worse,
ceteris paribus, than to select a victim for virtually any other reason – appears either dubious or
unjustifiable through argument available to ordinary human intellect.  And this weakness renders
the argument from more culpable mental states vulnerable, like the argument from psychological
harms to non-victim third parties, to the charge that it arises out of and affirms an undesirable
balkanizing, identity group, politics of resentment.   What else, the objectors ask, could explain
the vigor and confidence with which proponents of this argument assert the greater culpability of
selecting a victim because he’s Black, or gay, or Jewish, or Latino as compared to selecting a
victim for any number of other reasons?  This charge aside, the upshot here is that the argument
from more culpable mental states is not a very compelling argument for either of the two sorts of
hate crimes laws in force in the United States today.

 C.  The argument from liberal democratic values.

 Hate crimes laws are sometimes defended on the grounds that they are needed to give
adequate public and symbolic expression to the fundamental liberal democratic values of
nondiscrimination and tolerance.  Crimes committed against particular victims because of or out
of a group-based animosity for their race, or religion, or ethnicity or sexual orientation violate
these values in a manner and to a degree calling for special public condemnation.  Or so the
argument goes.

 There are many difficulties with this argument.  The first is that it is weak as an argument
for hate crimes laws that apply just in case the criminal selects his victim because of but not
necessarily out of animosity toward her race, religion, ethnicity, etc.  The criminal who decides
just to assault Catholics, not because he has any group-based animosity toward them, but rather
because he wants to systematize his victims in some fashion and attacking only Catholics enables
him to do so, does not violate the liberal democratic values of nondiscrimination and tolerance in
any significant way, or at least he violates those values no more so than does any other criminal
guilty of assault.  So, the argument is best taken as an argument for hate crimes laws that apply
just in case the criminal selects his victim out of one or another specified group-based animosity.

 But even so taken the argument is not strong.  Even if a criminal who so selects his victim
violates the liberal democratic values of nondiscrimination and tolerance, and even if that
violation calls for a special and visible public condemnation, it does not follow that that
condemnation must take the form of enhanced criminal penalties.  A special concern with and
condemnation of such conduct may be publicly and visibly expressed in a variety of ways within
and through public political culture.  Given the seriousness of enhancing criminal penalties (a
limitation of liberty after all), the argument from liberal democratic values is weak as an
argument for hate crimes laws unless it can be shown that enhancing penalties is the only or the
best way publicly and visibly to express a special concern with and condemnation of such
conduct.

 But suppose this can be shown.  Is the argument from liberal democratic values then a
strong argument for hate crimes laws, or at least those laws triggered when a victim is selected
out of a specified group-based animosity?  Well, yes and no.  It all depends on what we mean
when we speak of the liberal democratic values of nondiscrimination and tolerance.

 Suppose we mean that persons ought not impose avoidable harms on others for irrational,
irrelevant or indefensible reasons.  Employers ought not deny jobs to otherwise qualified persons
solely because of their race.  Children ought not exclude from their circle of friends perfectly
kind and fun but very heavy or bespectacled peers.  And the like.  If this is what we mean when
we speak of the values of nondiscrimination and tolerance, then the argument from liberal
democratic values justifies hate crimes laws applicable not only to criminals who select their
victims because they are Black or gay or Jewish, but also to criminals who select their victims
because they are fat, or skinny, or ugly, or exceedingly beautiful, or socially awkward, or socially
adept, and so on, including all irrational, irrelevant or indefensible reasons for which a criminal
might select his victim.  But if this is what we mean then it is hard to see how to limit hate crimes
laws to any manageable list of specified group-based animosities the selecting of a victim from
which will trigger the laws’ application.  And if the list of group-based animosities triggering the
application of hate crimes laws is to include hatred of the fat, the skinny, the ugly, the tattooed,
the homeless, the socially awkward, the excessively wealthy, etc., then what is the point of
having hate crimes laws?  Why not simply enhance the penalties for all crimes of the targeted
type, e.g., all assaults, all vandalisms?

 Suppose what we mean by the values of nondiscrimination and tolerance is instead that
persons ought not impose avoidable harms on others because of some attribute or trait that either
cannot be changed or can be changed only at an unreasonable cost.  Employers, again, ought not
deny jobs to otherwise qualified persons solely because of their race, or their religion, or
ethnicity.  And children, again, ought not exclude from their circle of friends perfectly kind and
fun but very heavy or bespectacled peers.  But, again, if this is what we mean by the values of
nondiscrimination and tolerance, the list of specified group-based animosities the selecting of a
victim from which will trigger the enhancing of criminal penalties will be long indeed.  Those
who target the tall, or the exceedingly intelligent, or the stutterers ought to be subject, on this
understanding of nondiscrimination and tolerance values, to the penalty enhancements of hate
crimes laws.

 If hate crimes laws of the sort proponents advocate are to be justified by the argument
from liberal democratic values, then the values of nondiscrimination and tolerance must be tied
specially to race, religion, ethnicity, sexual orientation and the like.  After all, the sentiments and
intuitions felt by the supporters of hate crimes laws are aroused by lynchings and cross-burnings,
gay-bashings, vandalisms of Jewish businesses or cemeteries, arsons in Latino neighborhoods
and the like.  These are the paradigmatic hate crimes that call for enhanced penalties.  Not attacks
against the fat, the skinny, the tattooed, the shabbily dressed, the ugly, or the glamorous, even
when they are specifically targeted out of a generalized animosity for such persons.

 So, suppose what we mean when we speak of the values of nondiscrimination and
tolerance is just that persons ought not impose avoidable harms on others because of their race,
religion, sexual orientation, ethnicity and the like.  If this is what we mean, then we might be able
to argue from these liberal democratic values to hate crimes laws of the typical sort.  The
question is whether this is what we do, or should, mean by these values.

 The advantage of this account of nondiscrimination and tolerance values, or at least of
these values as they are invoked as part of a justification for hate crimes laws, is that it captures
the idea that there is something special about race, ethnicity, religion, sexual orientation and the
like and thus something special about imposing avoidable harms on others because of their race,
or ethnicity, etc.  If I refuse to date tall people or to purchase goods from persons who bear
tattoos, I irrationally and indefensibly impose an avoidable harm on others, perhaps even because
of a characteristic or trait that cannot be easily altered.  But it would be a stretch at best to say
that I violate the fundamental liberal democratic values of nondiscrimination and tolerance, at
least insofar as we are talking about those values as violated by paradigmatic hate crimes.  To be
sure, my conduct may be rightly criticized, even perhaps morally criticized through a loose use of
the language of nondiscrimination and tolerance.  But the basis of that criticism cannot really be
that I violate the values of nondiscrimination and tolerance insofar as we take those values to be
essential to a just and stable liberal democratic order.  But if we substitute Blacks for tall people
or Jews for persons who bear tattoos, then the picture changes, at least for those of us living in
the United States, with its history.

 It is tempting here to say that what the core values of nondiscrimination and tolerance
demand is that we ignore race, religion, ethnicity or sexual orientation in our interactions with
others.  That we be color-blind, etc., when it comes to social life, even in our criminal conduct,
should we endeavor such conduct.  But this is a view that many if not most proponents of hate
crimes laws will have reason to reject.  It implies that there is no significant moral difference
between a White criminal who selects his victims because they are Black and a Black criminal
who selects his victims because they are White, or between a Protestant criminal who selects his
victims because they are Jewish and a Jewish criminal who selects his victims because they are
Protestant.  But many if not most proponents of hate crimes laws begin with the intuition that
there is a moral difference, even if there are also moral similarities, between these cases.  In each
example, the prior but not the latter case reflects, expresses and arguably serves to reconstitute
existing historical patterns of structural, group-based oppression.   To treat the two cases as if
they are not morally distinct in any significant way is to fail to attend to the realities of long-
standing, structural, group-based oppression at which hate crimes laws are aimed as a partial
remedial social response.  It is to view the issue of hate crimes from the point of view of the non-
oppressed.

 It is racially motivated crimes against Blacks, not racially motivated crimes generally,
religiously motivated crimes against non-Christians, not religiously motivated crimes generally,
that generate the intuitions and sentiments many feel in favor of hate crimes laws.  It is crimes
against members of already oppressed groups precisely because they are members of already
oppressed groups that call for enhanced penalties.  Given the realities and history of structural,
group-based oppression, a racially motivated murder committed by a Black man against a White
man is, for the purposes of punishment, not very different from any other murder, or at least not
very different from a murder motivated by a hatred of brunettes committed by a blonde.  But a
racially motivated murder committed by a White man against a Black man is different from other
murders.  It connects with those realities and presses that history forward in ways other murders
cannot.

 At their core and as affirmed and protected by law, the values of nondiscrimination and
tolerance must be understood in terms of race, religion, ethnicity, sexual orientation and the like.
There is indeed something special, morally speaking, about these specific kinds of social
groupings: they point us toward some of the most pressing historical cases of structural group-
based oppression that call for a remedial social response.  What we mean, then, when we speak
of the fundamental values of nondiscrimination and tolerance is not that race, religion, ethnicity
or sexual orientation ought always to be ignored in social life, but rather that social life ought to
be reorganized so as to eliminate the real, specific, long-lived, structural group-based oppression
of Blacks and other non-Whites, Jews and other non-Christians, gays, lesbians and other non-
heterosexuals, and non-European ethnic groups.  These are the great, structural and evil instances
of discrimination and intolerance we want to end, not the more general fact that we notice and
sometimes are moved in our social interactions by the race, religion, ethnicity or sexual
orientation (or height, IQ, or handsomeness, for that matter) of others.  The intuition or sentiment
that racially motivated assaults on Blacks, for example, deserve more punishment than other
assaults in the United States arises out of a strong desire to end a real, particular case of structural
group-based oppression and an awareness of how such crimes affirms and threatens to
reconstitute that very oppression.

 The argument from liberal democratic values, then, is strongest if the values of
nondiscrimination and tolerance are understood not just to be specially connected and limited to
race, religion, ethnicity, sexual orientation and the like, but to be so connected and limited in a
particular way – proscribing the imposition of avoidable harms on Blacks because they are Black,
or Jews because they are Jews, etc., where the imposition of those harms reflects, expresses or
serves to reconstitute the real, historical oppression of Blacks, Jews, etc.  Of course, this version
of the argument, like the strongest versions of the arguments from greater harm and from more
culpable mental states, invites the objection that those who advocate hate crimes laws are
engaging in a balkanizing, identity-group, politics of resentment.

 There is, consequently, a temptation to retreat to the position that while it is not true that a
murder is always just a murder, it is true that a racially motivated murder is always just a racially
motivated murder.  All racially motivated murders violate the liberal democratic values of
nondiscrimination and tolerance equally, and thus if any call for enhanced penalties, all call
equally for the same enhanced penalties.  This version of the argument from liberal democratic
values is weak, however.  It presupposes a basic social structure within which Blacks and
Whites, gays and straights, non-Christians and Christians are situated or positioned
symmetrically as groups.  But this presupposition is false.  Indeed, it is from an awareness that
this presupposition is false that the moral intuitions and sentiments that most strongly favor hate
crimes laws arise.  Even if this presupposition were true, however, this version of the argument
from liberal democratic values would still be weak insofar as it offers no account of why
selecting a victim on the basis of her race, religion, ethnicity or sexual orientation is so bad as to
merit enhanced penalties (rather than some other sort of special public condemnation) as
compared to selecting a victim on the basis of her height, weight, IQ or occupation.  Only one
such account is plausible, of course, and that is that some of the most dramatic and damaging
cases of oppression in the United States have involved race (Blacks and other non-Whites),
religion (Jews and other non-Christians), ethnicity (non-European) and sexual orientation (gays,
lesbians and other non-heterosexuals).  But to admit this is to admit that not all racially motivated
crimes, for example, are morally equal, for not all racially motivated crimes reflect, express or
potentially reconstitute such oppression.  Oppression is always asymmetric.  So, racially
motivated attacks by Whites on Blacks connect with historic and ongoing oppression in ways that
racially motivated attacks by Blacks on Whites never could.

 The upshot then is two-fold.  The strongest version of the argument from liberal
democratic values invites the same balkanization objection as the strongest versions of the other
standard arguments for hate crimes laws.  And the strongest version of the argument from liberal
democratic values rides piggy back on, indeed is perhaps best understood as an imperfect
articulation of, a deeper argument from oppression.  It is that argument to which we turn now.

III. The Argument from Oppression.

 The argument from oppression goes like this:

P1:  Structural group-based oppression is a fact of history and contemporary life in the
          United States;
P2:  Those who belong to oppressed groups are disproportionately and systemically
          more vulnerable to a wide range of structurally and socially produced but
          avoidable harms;
P3:  Those who by virtue of their social positioning vis a vis the vulnerable members
          of an oppressed group enjoy a special capacity to protect them from the harms to
          which they are vulnerable, or to help to reorder social life so as to minimize (and
          eventually eliminate) their structurally produced,  group-based vulnerability, have
          a special moral obligation to do so;
P4:  Those who intentionally select a member of an oppressed group, because they are
          a member of that group (but not necessarily out of a group-based animus), as their
          victim for a crime the harm of which reflects, expresses or contributes
          distinctively to the social reconstitution of the disproportionate vulnerabilities
          from which members of that group suffer, and who do so notwithstanding their
          being positioned socially such that they enjoy a special capacity to protect or aid
          the members of that group, fail to live up to a special moral obligation they owe to
          their victim;
P5:  It is the violation of this special moral obligation that makes paradigmatic hate
          crimes morally worse than otherwise similar crimes;
P6:  In general levels of criminal punishment should be scaled to reflect degrees of
          moral wrongfulness;
C:  The additional moral wrongfulness of paradigmatic hate crimes justifies the extra
          punishment imposed by hate crimes laws (of a suitably revised sort).

 This argument differs from the standard arguments in important ways.  While it appeals
to the place of hate crimes in a social order that systemically works a certain kind of harm on
members of oppressed groups, it is not an argument from the greater harm caused by hate crimes
to direct and immediate victims or non-victim third parties.  And while it appeals to the greater
moral culpability of hate crimes offenders, it is not an argument from the greater culpability of
motives grounded in one or another group-based animosity.  And finally, while it rests ultimately
on and aims at the vindication of liberal democratic values, it casts itself fundamentally as an
argument from structural group-based oppression, not from the values of nondiscrimination and
tolerance as values applicable to interactions between individuals assumed to by symmetrically
positioned as group members  by and within their basic social structure.

 Central to the argument from oppression are the ideas of oppressed groups,
disproportionate vulnerabilities, and special moral obligations to protect or aid.  Explicating
some of these ideas may help to avoid misunderstanding.  And so it is to that task I turn now.

 Oppressed groups are first social groups.  Social groups are more than mere aggregates.
Insurance companies may aggregate persons with one or another genetic trait for actuarial
purposes.  But taken together persons so aggregated do not constitute a social group.  To
constitute a social group members must stand in determinate relations with one another
constituted through their interactions with one another and with those at the margins or outside
the group.  There are many different kinds of social groups, including associations, cultural or
identity groups, and structural groups.  Oppressed groups are necessarily structural groups,
although as such they sometimes overlap with cultural or identity groups, or with associations.

 Associations are social groups within which the determinate relations group members
stand in with respect to themselves and to those at the margins or outside the group are a function
of the group’s shared aims, purposes or ends (e.g., families, the Catholic Church).  Cultural or
identity groups are social groups within which the relevant relations are a function of how group
members construct their identity or self-understanding at its most basic levels (e.g., Chicanos,
Southerners).  Cultural or identity groups need not share any aim, purpose or end, although
sometimes they do.  But they must share something by way of tradition, history, language, social
practice or cultural forms, and the like sufficient for members to constitute themselves as an “us”
or “we” to which they belong.  Structural groups are social groups within which the determinate
relations group members stand in with respect to themselves and those at the margins or outside
the group are a function of how they are positioned by the basic social structure of their society
when it comes to access to fundamental goods and resources.  Structural groups are produced
through the institutional mechanisms through which authority, power, labor and production,
desire and sexuality, prestige and the like are socially constituted and organized.  Oppressed
groups are structural groups the members of which are systematically disadvantaged in the
distribution of or their access to the basic goods and resources needed to develop and exercise
capacities for self-expression, self-development, and self-determination.  The members of
oppressed groups (e.g., African-Americans in the United States) are thus disproportionately
vulnerable to a wide and serious range of socially produced and avoidable harms (among the
most serious, poverty, illiteracy, economic marginalization, violence, incarceration, avoidable
illness, early mortality, and the like).  Describing accurately these disproportionate vulnerabilities
and marking out the many and interrelated ways in which they arise out of a basic social structure
that systematically disadvantages some but not others is a key task of oppression theory.

 The members of structural groups need share no common aim, purpose or end.  And they
need neither find nor construct their identity or self-understanding at basic levels through appeal
to their membership in the group.  Indeed, they may (and unhappily sometimes do) belong to the
group without even knowing it.  The existence of and membership within structural groups is a
function of how persons are objectively positioned socially relative to one another within and
through their basic social structure.  The differentiation of structural groups is distinct, then, from
the differentiation of associationist groups and cultural or identity groups, for in the latter cases
group members must affirm their membership to be group members.  In the United States,
Blacks, gays, Jews and various ethnic groups, as well as women, have been and remain to various
degrees oppressed structural groups.  This social fact does not depend at all on whether Blacks,
gays, Jews, etc., think of themselves as a group.  Of course, it also does not preclude them
constituting associationist groups or cultural or identity groups more or less identical in terms of
their membership.

 It bears emphasizing that the existence of and membership within structural groups
generally, or oppressed groups more particularly, is a function of reiterated patterns of social
relations and interactions over time.  Three things follow from this.  First, membership within
structural groups is something that may come in degrees, depending on the degree to which one
is, over time and in general, implicated structurally in a web of systematically advantageous or
disadvantageous social relations.  So, particular persons may be more or less Black with respect
to their membership in the structural group called Blacks.  Second, membership within, indeed
the existence of, a structural group may change over time.  So, membership within and even the
existence of Jews as a structural group in the United States has and continues to undergo
significant change.  Third membership in a structural group is logically independent of
membership in associationist groups and cultural or identity groups.  So, a particular individual
may be marginally (perhaps not even) Black with respect to Blacks as a structural group but be
centrally Black with respect to Blacks as a cultural or identity group, or vice versa.

 Individual members of oppressed groups do not all suffer the same particular, individual
socially produced and avoidable harms by virtue of their social positioning.  But they are all
disproportionately vulnerable to them.  And for many, this vulnerability will itself constitute an
actualized harm, a sort of psychological weight rooted in a deep sense of anxiety, insecurity, or
powerlessness.

 To be vulnerable is to be in a distinctively or especially precarious position, to be exposed
or at risk to an unusual degree to some injury or harm.  To be vulnerable is not the same as
simply belonging to a class the members of which merely satisfy some precondition for a
particular harm.  Only women get ovarian cancer.  So, only women are at risk of ovarian cancer.
It doesn’t follow that to be a woman is to be vulnerable to ovarian cancer.  Similarly, only the
employed can lose their jobs during an economic recession.  It doesn’t follow that to be
employed is to be vulnerable to unemployment during a recession.

 Vulnerabilities can arise from many sources.  Some women are vulnerable to ovarian
cancer by virtue of their genetic endowment.  So, nature is one source of vulnerabilities.  The
disproportionate vulnerabilities that mark oppressed groups as oppressed groups arise from the
“normal” functioning of the basic social structure of the society in which they exist.  So, given
the structure of labor markets and of authority within most employment contexts today, Blacks
are more vulnerable to unemployment during an economic slowdown than are Whites.  They are
more likely to be laid off, and more likely to suffer adversely from being laid off.  And given the
historical exclusion of Blacks from political processes (something which continues informally
today), Blacks are more vulnerable to the legislative sacrifice of their interests for the “common
good.”  They are more likely to have their interests ignored during ordinary legislative processes,
and then to be accused of and marginalized for divisiveness for asserting their interests once
ordinary legislative processes are complete.  The multiplication of these disproportionate
vulnerabilities, ranging across wide ranges of social life, produce what Marilyn Frye has called
the “bird cage” effect: Blacks and other oppressed groups find themselves caged by an
intersecting network of constraints arising out of the “normal” operation of the basic social
structure that limit or undermine their self-development, self-expression and self-determination.

 Disproportionate vulnerabilities are commonly understood to impose special obligations
on those especially well-placed to prevent harm to, to protect, or to aid the vulnerable, and these
special obligations are often legally enforced.  So, adults have various special obligations to
children.  Providers of various services have various special obligations to the elderly, as do
providers of medical services to the terminally ill.  Those not cognitively impaired have special
obligations to those who are cognitively impaired.  These obligations arise as moral obligations
in each case out of the asymmetry of the respective parties’ social relationship or relative
positionings.  They are given legal backing because (although not exclusively because -
consequentialist considerations will have a role to play here) of the moral gravity of their
violation.  Similarly, the special obligations violated in paradigmatic hate crimes arise as moral
obligations out of the asymmetry of the respective parties’ social relationship or positioning
(their group-based relationship or positioning within or through the “normal” functioning of the
basic social structure) and are given legal backing because (although, again, not exclusively
because - consequentialist considerations will again have a role to play here) of the moral gravity
of their violation.  The moral intuition or sentiment at work here in the case of paradigmatic hate
crimes belongs to the same family as that at work when we judge that stealing from a blind man
is morally worse, and thus deserving of or at least eligible for greater punishment, than stealing
from a sighted man.

 A full defense of the argument from oppression would require much more than the
foregoing.  However, I hope the foregoing sufficient to give a clear enough sense of how the
argument goes and how it differs from and improves on the standard arguments to support a
concluding discussion of the claim that hate crimes laws arise out of and express a balkanizing,
identity-group politics of resentment.

IV.  Hate Crimes Laws: Progressive Politics or Balkanization

 Objections to hate crimes laws and the politics from which they arise cannot be assessed
apart from assumptions about the justification, aim and shape of hate crimes laws.  Suppose,
then, that the argument from oppression is taken as the sound justification for (suitably revised)
hate crimes laws.  What then are we to make of the objection that hate crimes laws arise out of
and affirm a balkanizing, identity-group politics of resentment?

 This objection has been raised in various forms by numerous critics.  In their book
critical of existing hate crimes laws and the politics from which they seem to have emerged,
Jacobs and Potter argue that rewriting particular criminal laws to take into account the racial,
religious, ethnic, etc., identities of offenders and victims will undermine the criminal law’s
potential for bolstering social solidarity by redefining crime as a problem of intergroup conflict
and encouraging citizens to think of themselves first not as citizens but as members of identity-
groups, ideally victimized and besieged, and ready to exact revenge on their oppressors.  This
is, of course, not one but many objections.  Those who make the sort of objection just sketched
typically make one or more of the three following claims:

1.  Hate crimes laws arise out of and encourage a politics of resentment that is
     psychologically harmful to those who engage in it.
2.  Hate crimes laws arise out of and encourage a balkanizing, interest-group politics
     hostile to the common good and at odds with republican and liberal democratic
     (small “r”; small “d”)  values.
3.  Hate crimes laws presuppose a false degree of internal unity within the relevant
     social groups (Blacks, gays, Jews, etc.) and thus encourage these groups to police
     their own membership in oppressive ways.

As directed at existing hate crimes laws and the standard justifications given for them (the
arguments from greater harm, more culpable mental states, and liberal democratic values), these
claims are, at least prima facie, not without some bite.  But the case against existing hate crimes
laws and the standard justifications given for them does not depend on the strength of these
objections, for the standard justifications are themselves either unsound or weak, as demonstrated
above.  So, rather than assess the force of these objections against existing hate crimes laws and
their standard justifications, I want to assess their force against hate crimes laws justified by (and
suitably revised in light of) the argument from oppression.

 A. A politics of resentment?

 Several critics from the left have argued that hate crimes initiatives (like perhaps slavery
reparations or affirmative action initiatives) bear an unhappy relationship to the politics of
resentment.  The politics of resentment is the politics of those who suffer from and seek to
anesthetize the deep hurt of finding themselves in a social world publicly regarded as legitimate
or just but within which they are systematically and significantly unable to satisfy their desires.  It
arises first out of a righteousness or righteous indignation, produced affectively, that overwhelms
and thus blocks the subjective experience of weakness, hurt and suffering.  That righteous
indignation is then given direction through the identification of some agent or agents as the cause
of that weakness, hurt or suffering.  The weakness, hurt or suffering is then displaced by the
inflicting on the causally responsible agent or agents a harm.  The aim of this displacement is not
to alter the balance of power in the social world, however.  It is merely to ease the pain of one’s
own condition.  Thus, for it to work, the harm inflicted on the external source of one’s suffering
must carry itself some public mark of legitimacy.  The paradigm harm inflicted, then, is
punishment, for punishment is what revenge is called by those who hypocritically seek to cloak
their acts in the garb of legitimacy or justice. Of course, the imposition of such a harm
(revenge/punishment) does not and was not intended to eliminate the source of the hurt and
suffering; it does not alter the balance of power in the social world.  Those who find themselves
systematically and significantly unable to satisfy their desires will continue to so find themselves;
and by passing their act of revenge off as legitimate punishment, they will have reaffirmed the
very norms that legitimate their suffering.  But they will have dulled their pain.

 Hate crimes laws, reparations for slavery and other group-based initiatives are often said
to arise out of and reflect such a politics of resentment.  They are the fruits of a politics of
revenge cloaking itself in the language of legitimacy and justice.  But they are poisoned fruits.
They do nothing to change the balance of power in the social world, and they invest those who
suffer the most under existing social conditions in their own suffering.

 This objection misses the mark, however.  Hate crimes laws, reparations initiatives,
affirmative action proposals and the like arise not out of a politics of resentment rooted in the
experience of suffering, but a politics of anger rooted in the knowledge of and struggle against
injustice.  The aim is not to inflict a harm on some agent or agents held causally responsible for
the suffering of those who suffer, but rather to reorder the basic social structure to eliminate the
structural and institutional bases of oppression.  Of course, this is, according to the critics, just
what proponents of hate crimes laws may be expected to say in their defense.

 To be sure, the politics of victimhood and resentment has sometimes accompanied hate
crimes and similar initiatives.  But for the most part, those actually pushing politically for these
initiatives are neither deeply invested in their own suffering, nor acting from a self-defeating
resentment.  They are angry, mobilized, and increasingly politically empowered.  Their aim: the
chance to participate and advance their own interests in a basic social structure free of group-
based oppression.  Ultimately, whether or not theirs is a disabling and disfiguring politics of
resentment cannot be determined a priori.  Time and social experience will settle the matter.  To
date, however, there is little empirical evidence to support this objection.

 B.  A balkanizing, interest-group politics?

 Critics from both the right and left have complained that hate crimes laws and similar
initiatives arise out of and give priority not to a politics of the common good but rather a divisive
interest-group politics.  In the inevitable clash over which groups are to enjoy which gains, the
argument goes, the res publica is lost.

 But if proponents of hate crimes laws and similar initiatives are correct about the realities
of oppression, the res publica was never in hand in the first place.  What divides proponents from
these critics of hate crimes laws, then, is not a commitment to a republican and liberal democratic
politics of the common good and treating like cases alike; both share that commitment.  What
divides them is their evaluation of existing social conditions.  Proponents of hate crimes laws and
similar initiatives argue that group-based oppression is a significant feature of these conditions,
and that there is, accordingly, no neutral, universal point of view from which all citizens may
engage in democratic deliberation over the common good or what in fact constitutes a like case.
To insist that some murders or assaults deserve greater punishment than others as a result of the
social positioning of the victims and offenders is not to subordinate the common good or the
demands of formal justice to the particular interests of one’s own group.  It is rather to insist that
social position and point of view matters and thus ought not be set aside in public deliberations
over what the common good requires or what counts as a like case for the purposes of formal
justice.  To deny this is either to deny the reality of oppression or to falsely and covertly
universalize the social position and particular point of view of the nonoppressed.

 Critics sometimes argue that social position and point of view did not, and rightly did not,
matter for other marginalized groups at earlier points in U.S. history.  In relatively short order,
numerous immigrant groups, for example, assimilated into mainstream politics, rarely
challenging the unstated point of view from which public democratic deliberation over the
common good and the demands of justice proceeded.  Why should matters be any different for
Blacks, gays, Jews and the like?

 The answer is that marginalized cultural groups may be assimilated into mainstream
democratic deliberation without any fundamental challenge to the unstated point of view from
which that deliberation proceeds because, or so long as, they are not also oppressed structural
groups.  Oppressed structural groups cannot participate, or cannot participate effectively, in
democratic deliberation over what the common good or justice requires without calling that point
of view into question, without refusing to adopt a point of view from which their oppression is
invisible or obscured.   And so their politics appears to be, indeed is, divisive at some level.  But
there is no other path to justice and the common good consistent with democratic commitments.
The choice between an unacceptably balkanizing interest-group politics and a more or less
harmonious republican and liberal democratic politics of the common good is one confronted
only in a society free of structural group-based oppression.  Only there will a commitment to
basic institutional arrangements and norms function as a common point if departure for all social
groups in democratic deliberation.  In a society not free of oppression, the choice is a false
dilemma.

 C: Do we really have or want social groups such as Blacks, gays, Jews, etc.?

 It is sometimes said that the sort of politics behind hate crimes laws and similar initiatives
presupposes and creates incentives to coerce a degree internal group unity neither possible nor
desirable.  So, critics will object to hate crimes laws on the grounds that Blacks (or gays, or Jews,
etc.) are too varied a group for such laws to be justified and that such laws create objectionable
incentives for Blacks (or gays, or Jews, etc.) to police their own membership for internal
conformity to certain criteria of group membership.

 This objection arises out of two mistakes.  The first is to forget that within the context of
hate crimes laws and their justification references to Blacks, gays, Jews and other groups are
references to structural groups, not cultural or identity groups or associationist groups.  Thus, it is
entirely beside the point that group members neither share important ends nor self-identify
through group membership.  The second mistake is to suppose that structural groups are properly
defined through some set of essential membership criteria, a set of necessary and sufficient
conditions the proper application of which determines whether any particular individual is in or
out of the group.  But structural groups cannot be so defined.  They are constituted and must be
understood in terms of how they are related to other structural groups within and through the
basic social structure that determines the ability of individuals, and thus groups of individuals, to
access basic social resources.  Individuals may and often will differ in the degrees to which they
belong to various structural groups, depending on how they are positioned within the network of
institutional social arrangements that facilitate or frustrate access to basic social resources.  This
does not preclude speaking intelligently of groups, however, so long as there are (as there in fact
are) concentrated nodal points of numerous individuals situated similarly to a very high degree
within and by basic institutional norms and arrangements.

 To the extent that the argument from oppression presupposes Blacks, gays, Jews and the
like as structural social groups, it neither presupposes nor encourages a false or high degree of
internal group unity.  Indeed, at least with respect to the politics behind hate crimes laws, the
term identity politics is misleading.  It is simply not in the nature of structural groups to make
claims as such about their identity or the identity of their members.  It is in their nature, instead,
to make claims about the justice of how they, and thus their members, are positioned within and
through the basic social structure.

 To be fair, it must be noted that political initiatives aimed at improving the social
positioning of oppressed structural groups are unlikely to succeed in a democracy without
substantial support and activism from group members, and that as a strategic and rhetorical
matter, cultivating a group-based identity (above and beyond ordinary “consciousness raising”)
among group members may prove desirable.   But this is a contingent and avoidable feature of
such political initiatives.  The less the resistance within the democratic polity, the less there is to
be gained by such rhetoric and political strategy.

V.  Conclusion

 Critics of hate crimes laws are correct about two things.  First, the standard arguments
given as justification for these laws are either unsound or weak.  Second, these laws and the
standard justifications given for them at least appear to arise out of and affirm a balkanizing,
identity/interest-group politics of resentment.  But the critics of hate crimes laws are incorrect in
concluding that there is no compelling moral case for such laws.  The argument from oppression
is sufficiently compelling to constitute a prima facie justification for such laws.  And hate crimes
laws justified by appeal to and suitably revised in light of it cannot be saddled with the charge
that they arise out of and affirm a balkanizing, identity/interest-group politics of resentment.

 Nevertheless, hate crimes laws and the politics from which they arise are undeniably
divisive.  But that is to be expected and by itself is unobjectionable; democratic initiatives aimed
at responding to and remedying structural group-based oppression are almost always divisive.
The question with all such initiatives is whether they are so divisive that their cost in terms of
social unity outweighs whatever moral and political gains they promise.   That is a complex
question I have neither asked nor answered in this paper.

 But two points bear emphasizing, as a final thought, here.  First, among the moral and
political gains promised by hate crimes laws and the politics which surround them is a broader,
deeper and more accurate understanding of social and political life.  And that is a gain perhaps
worth pursuing, even at the (one must hope temporary) cost of an increase in social tension and
division between groups.  Second, no proponent of hate crimes laws imagines them as a silver
bullet capable of working significant social change by themselves.  They are just one part of a
broader set of social, political and legal initiatives aimed at responding to and remedying
structural group-based oppression.  Ultimately, their merits and demerits ought to be assessed
within the context of that larger set of initiatives of which they are a part.

Endnotes