Content uploaded by Teresa Marques
Author content
All content in this area was uploaded by Teresa Marques on Sep 07, 2020
Content may be subject to copyright.
Comp. by: ANEESKUMAR A Stage : Proof ChapterID: 0004768053 Date:18/2/20
Time:20:57:40 Filepath:d:/womat-filecopy/0004768053.3D
Dictionary : OUP_UKdictionary 260
13
C13 Amelioration vs Perversion
Teresa Marques
C13.S1 13.1 Introduction
C13.P1 Theories about the meaning of words for social categories and roles raise several
interconnected questions. One is metaphysical: What is the nature of the social
things we talk about? Another is conceptual: How do we think about the social
world, and how does our thinking relate to the nature of social reality? And how
should it relate? The latter question concerns so-called ameliorative projects. Such
projects aim to address other additional questions, for instance, what is the point
of having the concept in question? What concept (if any) would better serve our
social or political goals? And who are we?¹
C13.P2 In this paper, I argue that we should add one more question to this list, namely:
How can we assess the legitimacy of ameliorative projects? I will try to answer this
question by drawing from deeply problematic historical cases. Based on these
cases, I’ll introduce two notions of meaning perversion. This, I believe, will put us
in a better position to understand the limits of the moral and political legitimacy
of ameliorative projects. Permissible meaning revisions are those that are not
meaning perversions. This does not tell us which are revisionary projects we
should develop. But it gives us a way of circumscribing the meaning revisions
we should not pursue.
C13.P3 Suppose we are interested in understanding crucial concepts of the language of
justice and politics, e.g. in understanding the concepts expressed by words like
‘democracy’,‘freedom’,‘fair elections’,‘citizen’,‘the people’, etc. Sally Haslanger
has persuasively made the case that we can mean different things by ‘conceptual
analysis’, and that our reflection on the role of the language of politics can take
different approaches. Haslanger (2006) distinguished between manifest, operative,
and target concepts. A manifest concept is the concept one thinks guides one’s
categorizing, whereas the operative concept is the one that corresponds to actual
categorization patterns. The target concept is the concept that, all things
¹ These questions are illustrated in debates about race, for instance in Appiah (2006), Glasgow
(2003), Andreasen (2000, 2005), Kitcher (2007), Haslanger (2003, 2006), Machery et al. (2009), Diaz-
Leon (2015), Ludwig (2018).
OUP UNCORRECTED PROOF –FIRST PROOF, 18/2/2020, SPi
Teresa Marques, Amelioration v. perversion, In: Shifting Concepts: The Philosophy and Psychology of Conceptual Variability.
Edited by: Teresa Marques and Åsa Wikforss, Oxford University Press (2020). ©the several contributors.
DOI: 10.1093/oso/9780198803331.003.0014
Comp. by: ANEESKUMAR A Stage : Proof ChapterID: 0004768053 Date:18/2/20
Time:20:57:40 Filepath:d:/womat-filecopy/0004768053.3D
Dictionary : OUP_UKdictionary 261
considered, we should employ given our interests, the facts, etc. (Haslanger, 2006,
99); it is the concept that, at the end of the ameliorative project, we should be
using.
C13.P4 The question I address in this paper—how do we assess the moral and political
legitimacy of revisionary projects?—is thus essential to know if a project is ameli-
orative. Haslanger considers a close question, and says ‘whether or not an analysis
is an improvement on existing meanings depends on the purposes of the inquiry’
(Haslanger, 2012, n 1, p. 367). But I take it that it would be too naïve to believe that
an analysis is an amelioration just in case the ‘engineers’of the revision have good
purposes. Some features of what Maynard and Benesch (2016) call Dangerous
Speech, which I introduce below, show that people can be convinced of the
goodness of a purpose while bringing out very bad outcomes. Hence, whether a
revision is an improvement does not depend exclusively on how convinced
theorists or activists are of the presumed goodness of their ends.
C13.P5 I take it, also, that there is some tension between the concept we should be
using, all things considered and the concept we should be using, given our aims,
interests, or the facts. The intersection of what we should do, all things considered,
and what we should do, given our aims, leaves open the possibility that the answer
to the legitimacy question may be epistemically unavailable, or that we can be
wilfully ignorant of the right answer. It would thus be desirable to have some
constraints or guidelines to assess a revisionary projects’legitimacy.
C13.P6 Now, when theorists consider conceptual engineering or conceptual ethics, they
are often concerned with the concepts a word should express, given the specific
aims of the theory. Theoretical concerns are normally removed from the political
domain. For instance, Tarski (1943) considered that ‘truth’as used in natural
languages is an incoherent notion, and that it gives rise to paradox. He argued that
‘truth’should be defined in a way that would allow it to play its foundational role
in a semantic theory, while insulating the theory from the contradictions and
paradoxes that the natural language use gives rise to. For the purposes of the
theory, a revision of the meaning of ‘true’is desirable if the intended meaning
fulfils the relevant theoretical aims.
C13.P7 Theoretical goals can also be pursued in social and political domains. Yet, there
are differences between the theoretical domains that do not touch on socially
relevant notions, and those that transpire into language use ‘in the wild’. Whereas
the notion of truth in Tarski’s theory is not meant to replace the use of ‘truth’in
informal discourse, meaning revisions about social kinds or social relations often
have both theoretic and social aims. For example, redefining what ‘marriage’
means in actual legislations has arguably changed the extension of ‘marriage’in
the real world.²
²I’m grateful to Pablo Rychter and Esa Díaz-León for discussion about this point.
OUP UNCORRECTED PROOF –FIRST PROOF, 18/2/2020, SPi
261
Comp. by: ANEESKUMAR A Stage : Proof ChapterID: 0004768053 Date:18/2/20
Time:20:57:40 Filepath:d:/womat-filecopy/0004768053.3D
Dictionary : OUP_UKdictionary 262
C13.P8 This paper differs from other recent articles that try to pin down the normative
constraints of conceptual engineering. For instance, Simion (2017) is optimistic
about the prospects of conceptual ameliorations, but considers only epistemic
constraints: a concept should be ameliorated, she claims, only insofar as this does
not entail epistemic loss. Podosky (2018) argues contra Simion that epistemic loss
is not an adequate criterion, and that conceptual revision should be allowed
whenever it has the “capacity to causally influence the world”. Moreover, he
suggests, normative questions should address feasibility constraints. He says that
the claim that a concept should be engineered is right only if
C13.P9 ...its desired causal influence on social reality is feasible; and the claim that a
concept should be engineered is wrong if its desired causal influence on social
reality is infeasible. Importantly, the notion of feasibility mentioned here is
epistemic: It is about the feasibility of representational accuracy.
(Podosky 2018, 13)
C13.P10 In other words, the conceptual engineering is feasible only if it is possible for a
concept to come to accurately represent (social) reality. For instance, the revision
of ‘marriage’and the progressive change of legislation in liberal democracies
around the world is feasible in Podosky’s sense, and this answers the question of
whether ‘marriage’should be ameliorated. This is an answer to one reading of the
normative question.
C13.P11 This is not the reading of the ‘should’in ‘when should a concept be amelior-
ated?’that I have in mind.³ I have in mind moral and political legitimacy.
A conceptual engineering project could be feasible in Podosky’s sense, and still
be morally and politically illegitimate. In other words, we should ask when are
revisionist projects not ameliorative but perversive. I will argue that there is no
obvious answer as to what ‘the value of a meaning revision’can be beyond serving
specific theoretical purposes, and that it is hard to balance the value of pursuing
theoretical aims against the possible harms of using the words ‘in the wild’,soto
speak. The conviction that an analysis is pursued with good intentions does not
guarantee the goodness of the outcome. And even when revisionary projects
achieve their aims, the aims themselves may not be desirable. On the contrary,
people should be cautious of deep feelings that their purposes are desirable or
good. There may be cases where there is no clear answer to the question of what
people should do. Given this, it would be desirable to establish guidelines for
³ Lawford-Smith (2013) argues that the appeal to political feasibility is often used for ruling out
political theories that can’t be implemented, but that it should rather be used as a tool for ranking
alternative theories ‘along one of the dimensions relevant to making decisions about what to actually
do’(Lawford-Smith, 2013, 245). Furthermore, feasibility is independent of both desirability and risk.
Hence, although understanding feasibility is an essential step in assessing political projects, it does not
assess the desirability of the projects’ends or risks.
OUP UNCORRECTED PROOF –FIRST PROOF, 18/2/2020, SPi
262
Comp. by: ANEESKUMAR A Stage : Proof ChapterID: 0004768053 Date:18/2/20
Time:20:57:40 Filepath:d:/womat-filecopy/0004768053.3D
Dictionary : OUP_UKdictionary 263
assessing the moral legitimacy of revisionist projects. It would be useful to know,
for instance, which projects should not be pursued. To this end, I’ll consider two
alternative definitions of meaning perversions.
C13.P12 Section 2 introduces some historical examples of meaning perversions, and
points to cognitive and affective biases that are conducive to condoning harmful
practices and behaviour, while protecting agents’moral self-righteous convictions.
Section 3 explores the limits of our capacity to establish the moral and political
legitimacy of revisionary projects, and takes race as a case study to illustrate the
point. Section 4 characterizes meaning perversions. Ultimately, meaning perver-
sions are revisionary projects that are politically or morally illegitimate. The
section offers two different non-equivalent and non-extensional definitions of
meaning perversion: one characterized in terms of harmful (perlocutionary)
effects, and one characterized in constitutive terms. The possibility of meaning
perversions in either sense establish the limits of permissible meaning revisions.
C13.S2 13.2 Meaning Revisions in the Wild
C13.S3 13.2.1 Lessons from the Past and the Present
C13.P13 In LTI,The Language of the Third Reich, Victor Klemperer offers a chilling
description of the corrupting power of language. Klemperer witnessed how,
under the Third Reich, ‘huge number of concepts and feelings’were corrupted
and perverted. This was the case of words like ‘heroic’,‘heroism’,or‘fanatic’. In his
diaries, he registered how the use of these words shifted in a way that indicated
that, for instance, one could not be a hero unless one were a fanatic. At the same
time, fanaticism was no longer regarded as a negative trait. Klemperer reported
that even after the Second World War ended,
C13.P14 ...young people in all their innocence, and despite a sincere effort to eliminate
the errors in their neglected education, cling to Nazi thought processes ...as
soon as this concept (heroism) was touched upon, everything became blurred in
the fog of Nazism ...and then replaced it with ‘fanatical’
(Klemperer, 1957/2000, p. 14).
C13.P15 He continues by drawing an illustrative analogy where Nazi language ‘comman-
deers for the party that which was previously common property and steps words,
groups of words, and sentence structures, in poison.’
C13.P16 Masha Gessen has also written on the corrupting power of discourse. In a 2017
article, she warns us of the damage of Donald Trump’s discourse to social and
political reality. She compares the way Trump talks with the way autocrats in the
former Soviet Union, and present-day rulers in Russia, talk:
OUP UNCORRECTED PROOF –FIRST PROOF, 18/2/2020, SPi
263
Comp. by: ANEESKUMAR A Stage : Proof ChapterID: 0004768053 Date:18/2/20
Time:20:57:40 Filepath:d:/womat-filecopy/0004768053.3D
Dictionary : OUP_UKdictionary 264
C13.P17 A Russian poet named Sergei Gandlevsky once said that in the late Soviet period
he became obsessed with hardware-store nomenclature. He loved the word
secateurs, for example. Garden shears, that is. Secateurs is a great word. It has
a shape. It has weight. It has a function. It is not ambiguous. It is also not a
hammer, a rake, or a plow. It is not even scissors. In a world where words were
constantly used to mean their opposite, being able to call secateurs ‘secateurs’—
and nothing else—was freedom.
C13.P18 ‘Freedom,’on the other hand, was, as you know, slavery. That’s Orwell’s 1984.
And it is also the USSR, a country that had ‘laws,’a‘constitution,’and even
‘elections,’also known as the ‘free expression of citizen will.’The elections, which
were mandatory, involved showing up at the so-called polling place, receiving a
pre-filled ballot—each office had one name matched to it—and depositing it in
the ballot box, out in the open. Again, this was called the ‘free expression of
citizen will.’There was nothing free about it, it did not constitute expression, it
had no relationship to citizenship or will because it granted the subject no
agency. Calling this ritual either an ‘election’or the ‘free expression of citizen
will’had a dual effect: it eviscerated the words ‘election,’‘free,’‘expression,’
‘citizen,’and ‘will,’and it also left the thing itself undescribed. When something
cannot be described, it does not become a fact of shared reality. Hundreds of
millions of Soviet citizens had an experience of the thing that could not be
described, but I would argue that they did not share that experience, because
they had no language for doing so. At the same time, an experience that could be
accurately described as, say, an ‘election,’or ‘free,’had been pre-emptively
discredited because those words had been used to denote something entirely
different (Gessen, 2017).
C13.P19 So, where Klemperer talks of ‘that which was previously common property’,
Gessen talks of a lost ‘shared reality’. This is, I think, one of the harmful effects
of meaning perversions, but here I will not elaborate further on what that amounts
to, or how it opens the way for autocracy.
C13.P20 I call these uses of ‘freedom’,‘free expression of citizen will’,‘election’,‘heroic’,
‘fanatical’,meaning perversions. I consider two non-synonymous and (possibly)
non-co-extensional senses of meaning perversion in Section 13.4: a causal conse-
quentialist sense, and a constitutive sense. The next subsection illustrates the extent
of the consequences of our cognitive failures and biases in some extreme cases.
C13.S4 13.2.3 The Road to Hell is Paved with Good Intentions
C13.P21 The end aim of meaning revisions is sometimes self-interested. Some politicians’
use words like ‘law and order’or ‘justice’to manipulate the justice system in order
protect themselves or consolidate their power, not to promote the rule of law in
OUP UNCORRECTED PROOF –FIRST PROOF, 18/2/2020, SPi
264
Comp. by: ANEESKUMAR A Stage : Proof ChapterID: 0004768053 Date:18/2/20
Time:20:57:40 Filepath:d:/womat-filecopy/0004768053.3D
Dictionary : OUP_UKdictionary 265
the service of justice and fairness. Other politicians’use words like ‘democracy’
and ‘free elections’to perpetuate their hold on power, not to guarantee govern-
ments representative of citizens’rights and of the rule of law. Yet, meaning
revisions often do have genuine social or political justice aims, or at least, aims
that their proponents believe will bring about a better world. This section dis-
cusses some historical cases where extremely harmful consequences were seen as
morally and politically permissible, or even required. People’s inability to foresee
those consequences were often due to common cognitive limitations. How does
this relate to conceptual engineering projects? Revisionists may genuinely believe
in the goodness of their purposes and methods. But there can be gaps between the
best course of action for a certain end, and people’s epistemic capacity to assess
those means and end. And the fact that certain ends are desired does not make
them desirable (all things considered), or desirable independently of the means set
to achieve them.
C13.P22 In recent work, Maynard and Benesch (2016) discuss the conditions under
which so-called Dangerous Speech (DS) can occur. They characterize DS as
speech acts that are capable of encouraging approval of violence by an audience.
This provides an extreme example of discourse that, from the perspective of the
speaker and the target audience, is perceived as morally legitimate. As Maynard
and Benesch characterize DS, it is a product of both its context and its content,
which feed into and overlap with each other. The specific features of the contexts
of DS are the (influential) speaker, the audience in its the socio-historical envir-
onment, and the availability of means of dissemination (Maynard & Benesch,
2016, p.77). If a community relies mainly on one source of information, the
message spread by that source has a greater impact. But an influential or authori-
tative speaker addressing a volatile audience through mass means of information
is not producing DS if the content is not inflammatory or hateful. DS often occurs
in social and historical contexts that increase the likelihood that the audience
accepts that violence against certain people is morally permissible. It can be seen
as rightful punishment for presumed past crimes, or as means to prevent pre-
sumed existential threats. Socio-historical contexts of DS may include longstand-
ing grievances, resentment, the memory of historic injustices (real or imagined), a
weak or dysfunctional justice system, competition for resources, or land disputes.
Influential speakers may manipulate and exacerbate the resentment against mem-
bers of another group for political gains.
C13.P23 Maynard and Benesch (2016) then offer a characterization of six possible
features of the content of Ds. These features need not all be present, and need
not be present in the same way in all cases. The first is dehumanization through
language,⁴which includes forms of discourse that can do direct harm by the
⁴See also Tirell (2012) or Jeshion (2016).
OUP UNCORRECTED PROOF –FIRST PROOF, 18/2/2020, SPi
265
Comp. by: ANEESKUMAR A Stage : Proof ChapterID: 0004768053 Date:18/2/20
Time:20:57:40 Filepath:d:/womat-filecopy/0004768053.3D
Dictionary : OUP_UKdictionary 266
offense, denigration, or derogation of members of a target group as undeserving of
the duties that are owed to them qua persons. The second feature of the content of
dangerous speech is guilt attribution: members of a group are said to be guilty (as
members of the group) of past crimes, e.g. rape or murder, stealing, responsible for
current difficulties, destruction of the economy, occupation, or oppression, etc.,
and their guilt is offered as the moral justification for feelings of resentment and
retributive action. The third feature content is threat construction: an in-group
accuses an out-group of posing an existential threat, which contributes to morally
justify acting in self-defence. The fourth feature is the destruction of alternatives
and the representation of a course of action as a historical necessity, or of
alternative courses of action as impractical or inefficacious. For instance, Figes
(2002) describes how many citizens of the Soviet Union thought that the violence
of the Stalinist era was the only possible and necessary path to Communism,
reporting someone that said: ‘I had my doubts about the Five Year Plan ...but
I justified it by the conviction that we were building something great ...a new
society that could not have been built by voluntary means’(Figes, 2002, p. 111).
C13.P24 The fifth important feature is what Maynard and Benesch (2016) call virtuetalk,
through which the audience is motivated by ‘deep and unreflected feelings that
something feels “good”or “bad”, in particular feelings that induce positive moral
self-appraisal’,a‘satisfactory mental image of themselves ...often shaped by
notions of ideal group-identities, that produces considerable self-esteem’
(Maynard & Benesch, 2016, p. 84). It also contributes to the moral justification
of actions against an out-group. The final feature of the content of dangerous
speech is future-bias, i.e. the biased belief in a promise of future goods. Future-bias
is presumed to outweigh the short-term difficulties the audience may have to
endure, or the moral costs of the violence against others:
C13.P25 But the anticipated benefits can also be extravagant and utopian—promises that
a positive transformation of society will be brought about through a temporary
violent transition, or that national unity and prosperity for a long-mistreated
people can be obtained. In light of the expectation that Soviet violence would
protect the revolution and usher in Communist utopia, Lenin assured his
followers that in the future ‘the cruelty of our lives, imposed by circumstance,
will be understood and pardoned. Everything will be understood, everything’
(Maynard and Benesch, 2016, pp. 85–6).
C13.P26 DS thus arises when a perfect storm of cognitive biases and contextual condi-
tions come together. The great motivational strength of DS—its nearly irresistible
pull—comes from its reliance on moral reasoning and motivation. This motiv-
ational force of DS arguably depends on the reactive attitudes involved. The
notion of reactive attitudes was introduced by P. F. Strawson (1962/2008), and
it includes attitudes like gratitude, resentment, and contempt. These are attitudes
OUP UNCORRECTED PROOF –FIRST PROOF, 18/2/2020, SPi
266
Comp. by: ANEESKUMAR A Stage : Proof ChapterID: 0004768053 Date:18/2/20
Time:20:57:40 Filepath:d:/womat-filecopy/0004768053.3D
Dictionary : OUP_UKdictionary 267
that we have in reaction to another person’s action towards us. They are essential
both in interpersonal relations and in our moral lives.⁵Yet, our cognitive failures
and biases—in ascribing guilt to others, in ruling out alternative courses of action,
in seeing ourselves as virtuous, as just knowing that we are right, or towards some
utopic future ideal, show that reactive attitudes and the actions they motivate can
be easily misguided and morally unjustified.
C13.P27 Meaning perversions are not a minor issue. It should not surprise anyone to
learn that meaning perversions played a central role in cases of DS that motivated
historically inhumane acts of mass violence. Through meaning perversions, as
Victor Klemperer put it, ‘language does not simply write and think for me, it also
increasingly dictates my feelings and governs my entire spiritual being the more
unquestioningly and unconsciously I abandon myself to it’. Meaning perversions
played a part in making it possible for Soviet citizens to tolerate and condone the
actions that brought about the hunger of millions of people, for instance the
Holodomor in Ukraine (a fact many still deny or minimize).⁶
C13.P28 How can we circumscribe the limits of morally and politically legitimate
meaning revisions, in particular when politics or justice are at stake?
C13.S5 13.3 The Limits of Ameliorative Projects
C13.P29 This section advocates for defensive pessimism⁷in the assessment of the legitim-
acy of meaning revisions, i.e. in the assessment of the value and desirability of a
meaning revision to a certain end. Revisionary projects should be approached not
just with defensive pessimism, but also with honest humility. Indeed, theorists
should not only be aware that any possible meaning revision may, and likely will,
have harmful unintended consequences. Theorists should also be aware of the
cognitive biases that often prevent them from seeing the harm their plans and
actions bring about. Insofar as it is possible, such harm should be minimized.
Below, the debate about the amelioration of the concept of race is introduced as a
case study that illustrates the difficulty of circumscribing the politically relevant
revisions that, ultimately, do no harm.
C13.P30 One way of deploying defensive pessimism, i.e. of taking proactive behaviour to
counteract possible negative outcomes, is to adopt ‘reflective equilibrium’as a
method for finding coherence among diverse considered judgments, bringing
them into relations of mutual support and explanation.⁸The process should be
⁵The notion has been deployed in recent moral and political philosophy, for instance by Manne
(2017), Björnsson and Hess (2017), Goldman (2014), or Couto (2016).
⁶See Snyder (2017a, 2017b).
⁷The notion of defensive pessimism was introduced in Norem (2001). It is a cognitive strategy that
helps people to take proactive behaviour to counteract possible negative outcomes.
⁸After Rawls (1971).
OUP UNCORRECTED PROOF –FIRST PROOF, 18/2/2020, SPi
267
Comp. by: ANEESKUMAR A Stage : Proof ChapterID: 0004768053 Date:18/2/20
Time:20:57:40 Filepath:d:/womat-filecopy/0004768053.3D
Dictionary : OUP_UKdictionary 268
understood as allowing for the revision of convictions, given the facts and the
foreseeable outcomes, a revision that honest humility can facilitate. But what are
the morally or politically criteria that should guide our assessment of the facts and
the foreseeable outcomes of concrete revisionary projects?
C13.P31 In a 2018 lecture, Haslanger argued against ideal theory and in favour of critical
non-ideal theory to decide how to answer questions in social ontology—what
possible and actual scenarios are relevant to consider. This decision, she argued,
depends on what we want to know and for what purposes. Haslanger claimed that
to address concrete justice issues—e.g. ‘How should we revise the educational
system in Boston to be more fair?—it is not helpful to start from idealized
examples and ideal theory:
C13.P32 We can learn a lot about justice from considering concrete social circumstances.
I would argue that such ‘bottom up’(v. ‘top-down’) theorizing is the best way to
learn about justice, for our background presuppositions are tested against the
messy reality we are trying to address (Haslanger, 2018, p. 2).
C13.P33 I don’t have a definite view here, but I am sceptical of doing away entirely with
ideal theory, and relying only on ‘bottom up’theorizing. After all, without gaining
distance from immediate practical concerns (interests and purposes) people may
never move away from the deep feelings that something feels right, while wilfully
ignoring the harm that they can cause by acting only on such feelings.⁹In the case
of the rightness of a revision of meaning, the mere reliance on bottom up
theorizing, and on feelings that something feels rights, cannot suffice to guarantee
the moral or political legitimacy of a revisionary project.
C13.P34 Perhaps ‘the value of a meaning revision’can refer to its positive overall
consequences, all things considered. The distinction between an amelioration
and a perversion would then depend on the balance of the good versus the
harmful consequences of a revisionist project. But how can a theorist setting out
to advance a new categorization know that good consequences will outweigh the
harmful ones?
C13.P35 The opening section distinguished between meaning revisions that are pursued
to meet specific theoretical aims, and those that are intended to have real social
consequences. This is a distinction that has been made in discussions about the
usefulness of the concept of race. For instance, Kitcher (2007) defended a prag-
matic biological view about race, considering that there are non-denumerably
many ways to sort people into biological categories. These possible divisions, he
claimed, depend on our cognitive capacities and our purposes. Other authors like
Andreasen (2004) argued for a cladistic conception of human races (see also
⁹In any case, following a reflective equilibrium approach can be pursued in a nonideal way. See, for
instance, Stemplowska and Swift (2012).
OUP UNCORRECTED PROOF –FIRST PROOF, 18/2/2020, SPi
268
Comp. by: ANEESKUMAR A Stage : Proof ChapterID: 0004768053 Date:18/2/20
Time:20:57:41 Filepath:d:/womat-filecopy/0004768053.3D
Dictionary : OUP_UKdictionary 269
Andreasen, 1998, 2000, 2005). One could perhaps ameliorate the concept of race
to meet the purposes of a theory of human biology, and define races as ‘ancestor-
descendant sequences of breeding populations that share a common origin’
(Andreasen, 2004, p. 425).
C13.P36 Yet, eliminativists like Appiah (Wilkins, Appiah, & Gutmann, 1998) argued
that there are no human races not just because ordinary racial categories do not
track any actual natural differences between people. The harms caused by the false
belief in a unique biological base for racial categories is an additional reason to
abandon all talk of human races. In European Portuguese, for instance, the term
‘raça’is conatively loaded. Most people on the left avoid using the word. In fact,
the end of the fascist dictatorship in Portugal in the 1970s led also to the
independence of former Portuguese colonies in Africa (the revolution to end the
dictatorship was carried out by the military who wanted to put an end to the war
for independence of those countries, which had been going on for approximately
thirteen years). In the new constitution drafted by a mostly socialist and social-
democratic parliament, all uses of ‘raça’were removed. A similar process led to the
removal of mentions of ‘Rasse’from German legislation:
C13.P37 Recent debates about Weissein (‘whiteness’) in Germany provide further evi-
dence that recognition of social realities of white supremacy does not presuppose
an account of race in terms of these realities. For example, Amjahid’s (2017) book
Unter Weissen (‘Among Whites’) dissects German practices of white supremacy
but explicitly avoids realist appeals to Rasse by pointing out the ‘historical
burden’(2017, p. 49) of the concept. In this sense, the German concept of race
may indeed be more adequately understood in analogy to other failed concepts
such as witch. While alleged witches were forced in very real social positions,
claims about the reality of races in Germany seem just as misleading as claims
about the reality of witches (Ludwig, 2018, p. 8).
C13.P38 In reply to similar concerns, Kitcher says that the usefulness of racial categories
will depend on the theoretical purposes that are best served by having those
categories.¹⁰But, as he puts it:
C13.P39 [E]ven if the concept of race plays a role in some lines of biological inquiry, the
values of those lines of inquiry, and of pursuing them through retention of the
concept of human race, would have to be sufficiently great to outweigh the
potential damage caused by deploying this concept in the other contexts in
which it plays so prominent a role, namely in our social discussions
(Kitcher, 2007, p. 302).
¹⁰See also Díaz-León’s appeal to Kitcher’s criterion, Chapter 10 in this volume.
OUP UNCORRECTED PROOF –FIRST PROOF, 18/2/2020, SPi
269
Comp. by: ANEESKUMAR A Stage : Proof ChapterID: 0004768053 Date:18/2/20
Time:20:57:41 Filepath:d:/womat-filecopy/0004768053.3D
Dictionary : OUP_UKdictionary 270
C13.P40 Kitcher’s suggestion—that we can ameliorate the concept of race when the value
of pursuing biological lines of inquiry is sufficiently great to outweigh the potential
damage caused by deploying the concept in social contexts—seems a bit cavalier.
The theorist must assume that the value of his inquiry justifies not only categor-
izing people into clades, i.e. ancestor-descendant sequences of breeding popula-
tions, which of course is useful for biology. The theorist is, on assumption, also
using the word ‘race’to refer to a clade knowing that millions of people who have
suffered because of their categorization under different ‘races’, and that talking of
biological races will continue to feed the social discrimination perpetrated on this
basis. Redefining the meaning of race terms for the important goals of a theory
does not do away, or minimize, the negative impact the use of ‘race’has in
society.¹¹
C13.P41 What can we learn from this discussion? Can we say that the Germans and the
Portuguese have ameliorated their race concepts by revising the terms used as
empty, and hence removing them from legislation? That can’t be right. Learning
that a term has no referent is not the same as revising its meaning. ‘Unicorn’does
not refer to anything because there are no unicorns, not because people amelior-
ated the meaning of the word to give it a null-extension. Pointing out that there
are species of mammals that may have given origin to the concept of unicorn—
horses and narwhals—and that the study of these organisms is useful in biology,
does not entail in any way that biology would benefit from using the term
‘unicorn’to refer to either species. Equally, ‘witch’does not refer to anything
because there are no witches. Pointing out that there have been women who were
considered inconvenient and who may have been at the origin of the concept of
witch—maybe those unmarried, maybe those with a mind of their own—and that
our knowledge of their lives is useful historical knowledge, and perhaps a useful
sociological warning sign of misogyny, does not entail in any way that social
theories would benefit from using the term ‘witch’to refer to actual women.
C13.P42 A meaning revision that gives an empty term a new extension can hardly be
ameliorative when giving the term any extension raises the probability of causing
¹¹ I don’t address social constructionist non-eliminativism were, except briefly in a paragraph below.
First, because race is not the main focus of this chapter. It is enough here to focus on what most people
take races to be—biological kinds (see Machery & Faucher, Chapter 9 this volume, on the folk concept
of race). Second, it is unclear that the most appropriate way of addressing social inequality is on the
basis of a social constructionist non-eliminativist notion of race. Races are not obviously social
constructions, unlike property, presidents, judges, professors, stock exchanges, etc. The claim that
race is socially constructed is a theoretical view that social theorists can adopt to explain both the fact
that it is not biologically real and that there are social relations (in several countries) that seem to be
based on racial or ethnic group membership. But the point of treating race as a social construction is
not just theoretical, it is practical—the theorist offers the ameliorated concept to replace the folk
concept. But how do we know that deploying the theoretical notion of race in the real world would
contribute to fairness? For all we know, it may be that reifying social identities is an obstacle to the
creation of a more equalitarian common ground., It may be that the social constructionist proposal
faces the same challenges as the biological one: that redefining the meaning of race terms for the
important goals of the theory does not erase the negative social impact of the use of ‘race’.
OUP UNCORRECTED PROOF –FIRST PROOF, 18/2/2020, SPi
270
Comp. by: ANEESKUMAR A Stage : Proof ChapterID: 0004768053 Date:18/2/20
Time:20:57:41 Filepath:d:/womat-filecopy/0004768053.3D
Dictionary : OUP_UKdictionary 271
very real harm. In the case of ‘race’, science can continue to research human
reproduction and migration without using race-talk, as social sciences can without
using ‘witch’. Race-talk has had, and continues to have, hugely harmful conse-
quences. In the German or Portuguese contexts, the usefulness of pursuing lines of
inquiry about human clades can be served with the technical term, ‘clades’, and
not by the loaded terms ‘raça’or ‘Rasse’. And as Ludwig (2018) suggests, issues
related to justice can be addressed instead by talking of racialized practices and
conceptions.¹²
C13.P43 Now, even if it were true that there is only one way of classifying humans into
different clades, it would still be the case that those divisions should not warrant
racially based discrimination. For any biologically real characteristic, as for
instance biological sex is, we should refrain from discriminating anyone on its
basis. I take discrimination to differ from mere differential treatment.¹³ This is a
different issue than the race problem discussed above. In the race case, there’s
good reasons to deny that there are races, because the underlying biological reality
does not provide a coherent way of matching biological categories to pre-
theoretical folk race conceptions, and race words are normatively loaded in highly
pernicious ways. But we do track other biological categories fairly well, as is the
case with sex categories, or conditions like albinism. Female people, or albino
people, have suffered discrimination and harm for being what they are. That harm
does not justify doing away with categorizing sexually male and sexually female
organisms, any more than the harm that albino organisms suffer (they are prime
targets for predators, for a start) does not justify doing away with the distinction
between albinos and animals with (some) melatonin. Doing away with the
categorization does not do away with the harm. When those organisms are
human, recognizing people for what they are may indeed be the required differ-
ential treatment that is necessary to address the injuries they are more prone to
suffer. For instance, recognizing the differences between female and male symp-
toms of cardiovascular disease is necessary for properly diagnosing and treating
females under forty who may die from heart attacks or strokes (Stamp, 2018).
C13.P44 In recent work, Haslanger (2019) offers a more sophisticated answer to the
question of how to assess the adequacy of ameliorative projects. She draws from
work by Knobe et al. (2013) and Knobe and Prasada (2011) on dual character
concepts, i.e. concepts that have both a descriptive and a normative dimension,
where dual character concepts
C13.P45 ...are represented by (a) a set of concrete features and (b) a set of abstract values
that the concrete features are seen as realizing. These two representations are
¹² This is also, in fact, the strategy for which Robin Andreasen argues (Chapter 11 in this volume),
departing from her earlier views.
¹³ See Lippert-Rasmussen (2013) for a discussion on ways to define discrimination.
OUP UNCORRECTED PROOF –FIRST PROOF, 18/2/2020, SPi
271
Comp. by: ANEESKUMAR A Stage : Proof ChapterID: 0004768053 Date:18/2/20
Time:20:57:41 Filepath:d:/womat-filecopy/0004768053.3D
Dictionary : OUP_UKdictionary 272
intrinsically related, but they are nonetheless distinct, and they can sometimes
yield opposing verdicts about whether a particular object counts as a category
member or not (Knobe & Prasada, 2011, p. 2,965).
C13.P46 The dual character view offers a promising prospect for understanding the
complexity of meaning change and will help with defining one of the senses of
meaning perversions in the next section.
C13.P47 This section showed that it is unclear what the value of a meaning revision can
be beyond serving theoretical purposes. The value of pursuing theoretical aims
should be balanced against the possible harm caused by using the word in the wild.
The conviction that an analysis is pursued to attain valuable theoretical or political
aims does not guarantee the goodness of the outcome. On the contrary, theorists
should be cautious of ‘deep feelings’about the desirability of their projects and
aims. Section 13.4 offers a way to circumscribe those meaning revisions that
should not be pursued. It will take advantage of the notion of dual character
concepts, and of the difference between harmful illocutionary effects and harmful
perlocutionary effects.
C13.S6 13.4 The Legitimacy of Ameliorative Projects
C13.P48 This section introduces some theoretical resources that, together with the notion
of dual character concepts, can help us understand language’s motivational power,
and the illocutionary and perlocutionary effects that contribute to structure social
relations. These resources are drawn not only from philosophy of language, but
also from philosophy of action and of emotions. It makes use of these resources to
distinguish between two senses of meaning perversion. Some of the notions
introduced play a crucial role in the explanation of the expressive normative
dimensions of meaning. The notion of dual-character concepts can be reinter-
preted as multidimensional aspects of meaning, and that efforts to change one
dimension of meaning may not be accompanied by changes in the other
dimension.
C13.S7 13.4.1 The Illocutionary Structure of Contexts
C13.P49 In recent joint work, García-Carpintero and I propose an expressive presupposi-
tional account of derogatory language. Our view can be extended beyond deroga-
tory language to other kinds of discourse that encode normative and evaluative
expressive presuppositions. On our view, expressive presuppositions are not just
propositions to be added to a common ground as shared beliefs. Expressive
language, e.g. that involving slurs, includes a cognitive and a conative dimension.
OUP UNCORRECTED PROOF –FIRST PROOF, 18/2/2020, SPi
272
Comp. by: ANEESKUMAR A Stage : Proof ChapterID: 0004768053 Date:18/2/20
Time:20:57:41 Filepath:d:/womat-filecopy/0004768053.3D
Dictionary : OUP_UKdictionary 273
Expressives make requirements on a shared conative record, governed by sui
generis norms specific to affective attitudes and their public manifestations
(Marques & García-Carpintero, 2019).¹⁴
C13.P50 If emotions are sui generis normative states (Mulligan, 1998, D’Arms &
Jacobson, 2000, Deonna & Teroni, 2015), and the speech acts that express them
are defined by distinctive norms, then in order to incorporate the presuppositional
view of pejoratives we should add further illocutionary structure to the context set.
The intentional objects of the emotional states provide this additional structure.
For instance, a pejorative or a slur presents its target as an adequate recipient of
mistreatment, as worthy of contempt. The ‘formal object’of the emotion is the
normative condition that allegedly justifies the emotional attitude towards it. But
this normative condition—the property of being contemptible—is not part of the
represented content, and hence what speakers accept in accepting a use of a
pejorative is not the belief that a target is worthy of contempt.
C13.P51 The nature of the expressive conative dimension of language use requires
conversational contexts to have illocutionary structure,¹⁵including the different
classes of contents to which speakers are committed in different modes: in the
ways they are committed to their beliefs, to their intentions, to their affective
attitudes, and to the questions guiding their inquiries. In felicitous contexts, these
different commitments are mutually shared, and license presuppositions. As
Stalnaker’s (1978) account of assertion emphasizes, an accepted assertion comes
to be presupposed afterwards, allowing for the satisfaction of presuppositional
requirements later on in the discourse. Similarly, accepted questions under dis-
cussion (QUDs), directives, and expressives can come to be taken for granted,
constraining the legitimate moves that can later be made.
C13.P52 Our account fills in some of the details of the proposal made by Langton (2012):
C13.P53 I want to propose, in an exploratory spirit, the idea that the phenomenon of
accommodation might extend beyond belief—beyond conversational score, and
common ground, as originally conceived—to include accommodation of other
attitudes, including desire and hatred. My remarks here will inevitably be pro-
grammatic. But to convey the general idea: just as a hearer’s belief can spring into
being, after the speaker presupposes that belief, so too a hearer’s desire can spring
into being, after the speaker presupposes the hearer’s desire; and so too a hearer’s
hatred can spring into being, after the speaker presupposes that hatred.
Stalnaker’s common ground can perhaps be extended to include not just com-
mon beliefs, and other belief-like attitudes, but common desires, and common
¹⁴See also Cepollaro and Stojanovic (2016) for a related hybrid presuppositional view.
¹⁵Charlow (2016) and Portner (2016) suggest that directives have a content to be added (when
successful) to a collection of propositions that represent the mutually known active projects of the
interlocutors, a ‘To Do List’or ‘Plan Set’, and Roberts (2012) argues that contexts are further structured
by the questions under discussion.
OUP UNCORRECTED PROOF –FIRST PROOF, 18/2/2020, SPi
273
Comp. by: ANEESKUMAR A Stage : Proof ChapterID: 0004768053 Date:18/2/20
Time:20:57:41 Filepath:d:/womat-filecopy/0004768053.3D
Dictionary : OUP_UKdictionary 274
feelings, as well. Speakers invite hearers not only to join in a shared belief world,
but also a shared desire world, and a shared hate world
(Langton, 2012, p. 140).
C13.P54 The context update made with the literal use of an expressive is an update not only
on the shared cognitive common ground—the set of propositions that come to be
commonly accepted under a belief mode—but also on shared ‘motivational
set’¹⁶—the set of intentions, evaluative dispositions, or desires that are shared
under intentional and sentimental modes.
C13.P55 A shared motivational set can be part of common ground, which would include
those motivational attitudes that are in fact common, and taken to be such. But we
can also represent a ‘motivational score’as part of a conversational score, i.e. those
attitudes whose permissible expression is part of the conversational score. The
distinction is useful. People often accept to follow norms that they don’t agree
with or that they disavow, and accept to act in accordance with widely shared and
permissible evaluative attitudes that they do not actually have.
C13.P56 People may have doubts about the prospects of sharing emotions under a
sentimental mode, even if the prospect of sharing plans does not raise the same
doubts.¹⁷But those like Salmela and Michiru (2016) have outlined an account of
collective emotions that links the intentional structure of joint actions and the
underlying cognitive and affective mechanisms. Collective emotions, they argue,
can function as both motivating and justifying reasons for jointly intentional
actions, in some cases even without prior joint intentions of the participants.¹⁸
C13.P57 Before embarking on meaning or conceptual revision projects, theorists should
not only incorporate the developments on dual character concepts, as Haslanger
proposes. They should also consider the hybrid nature of evaluative and norma-
tive language, and seek a better understanding of the role of emotions in motiv-
ating and persuading through discourse. If words’meanings can serve to make not
only assertive illocutionary acts, but also to make expressive speech acts, the things
(situations, events, or people) described can be taken for granted as realizing
certain norms or values through the emotional or affective attitudes expressed.
The next subsection shows how harmful revisionary projects will often result from
this hybrid dimension of language.
¹⁶The suggestion adapts Bernard Williams’s (1979) notion of a ‘motivational set’, i.e. the set of
dispositional attitudes, plans, intentions, emotions, etc., that identify the reasons agents have for acting,
and are an integral part of their practical reasoning.
¹⁷On shared plans, see for instance Kutz (2000).
¹⁸Salmela and von Scheve (2017) deploy research on collective emotions to explain right-wing
radical populism, illustrating the usefulness of a notion of a common motivational set to explain the
functioning of Dangerous Speech. They argue that there are two psychological mechanisms underlying
the rise of right-wing populism: ressentiment and emotional distancing.
OUP UNCORRECTED PROOF –FIRST PROOF, 18/2/2020, SPi
274
Comp. by: ANEESKUMAR A Stage : Proof ChapterID: 0004768053 Date:18/2/20
Time:20:57:41 Filepath:d:/womat-filecopy/0004768053.3D
Dictionary : OUP_UKdictionary 275
C13.S8 13.4.2 Harmful Perlocutionary Effects or Constitutive
Norm Erosions?
C13.P58 The main question this chapter addresses is how can we assess the legitimacy of
ameliorative purposes and projects. The literature on conceptual engineering has
not devoted sufficient attention to the possibility of meaning perversions.
However, conceptual or meaning revisions can be either ameliorative or perver-
sive. Ameliorations are improvements on existing meanings. Perversions, in
contrast and by analogy, are corruptions of existing meanings.
C13.P59 This is still fairly imprecise, but it can be understood in two ways. First,
perversions can be understood in causal effects. These could include many of
the effects that Klemperer and Gessen mention: an impoverished experience, a
destitute language, the loss of a shared reality, the loss of individual autonomy to a
language that ‘thinks for us, and dictates our feelings’, as well as actual discrim-
ination or oppression. The political dangers of these effects should not be neg-
lected or minimized. These are often the effects autocrats intend, since they
diminish a population’s capacity to resist the autocrat’s control over social and
political reality:
C13.P60 Causal meaning perversions are attempts to hijack language (e.g. of justice, politics,
social roles, or moral or epistemic virtues) in a way that makes people worse off. In
other words, meaning perversions are revisionary or engineering projects that
cause harm.
C13.P61 In a very literal sense, if a meaning revision produces harmful consequences it is
not an amelioration. We may be concerned, however, that the causal notion of
perversion cannot guide us in distinguishing between the merely feasible projects
(in Podosky’s sense, introduced at the start of this chapter) and the morally
legitimate projects.
C13.P62 We need a more perspicuous way of distinguishing ameliorations from perver-
sions, one that goes beyond a focus on possible harmful effects. The reason is,
fundamentally, epistemic. Shelly Kagan spells it out:
C13.P63 Perhaps the most common objection to consequentialism is this: it is impossible
to know the future. This means that you will never be absolutely certain as to
what all the consequences of your act will be. An act that looks like it will lead to
the best results overall may turn out badly, since things often don’t turn out the
way you think they will. Something extremely unlikely may happen, and an act
that was overwhelmingly likely to lead to good results might—for reasons beyond
your control—produce disaster. Or there may be long-term bad effects from your
act, side-effects that were unforeseen and indeed unforeseeable. In fact, lacking a
OUP UNCORRECTED PROOF –FIRST PROOF, 18/2/2020, SPi
275
Comp. by: ANEESKUMAR A Stage : Proof ChapterID: 0004768053 Date:18/2/20
Time:20:57:41 Filepath:d:/womat-filecopy/0004768053.3D
Dictionary : OUP_UKdictionary 276
crystal ball, how could you possibly tell what all the effects of your act will be? So
how can we tell which act will lead to the best results overall—counting all the
results? This seems to mean that consequentialism will be unusable as a moral
guide to action (Kagan, 1998, p. 64).
C13.P64 Kagan’s argument raises a problem for conceptual engineering projects gener-
ally.¹⁹It is a problem for efforts to discriminate ameliorations from perversions,
and presses us to find a more useful way of identifying the engineering concepts
we should be engaged in, morally and politically.
C13.P65 A second sense of meaning perversions, I suggest, is constitutive. The cases that
Klemperer and Gessen focus on, together with the theoretical resources intro-
duced in the previous subsection, and a modification of Jason Stanley’sdefinition
of undermining propaganda (Stanley, 2015) provide the elements for a different
better definition.
C13.P66 Stanley defines undermining propaganda thus:
C13.P67 Undermining Propaganda: A contribution to public discourse that is presented
as an embodiment of certain ideals yet is of a kind that tends to erode those very
ideals (Stanley, 2015, pp. 52–3).
C13.P68 Now, I don’t use Stanley’sdefinition of undermining propaganda for two reasons.
First, propaganda as he defines it will include uses of code words or dog whistles
(Mendelberg, 2001, Khoo, 2017, Saul, 2018), or of racial fig leaves (Saul, 2017), and
I think that meaning perversions function in different ways than these other
phenomena. Second, I think that meaning perversions can occur outside the
remit of propaganda, for instance, in a relationship, where one of the parties can
pervert the meaning of ‘good friend’. Hence, not all propaganda is a meaning
perversion, and not all meaning perversions are propaganda.
C13.P69 In my proposed definition, the idea of an undermining of norms and ideals is
still essential. Normally, a correct use of word that expressively presupposes a
norm or a value reinforces that value or norm, which is taken for granted as
common ground. For instance, thanking a child for being polite helps to reinforce
polite interactions and rules of etiquette. Meaning perversions are uses of words
that also presuppose shared norms or values, but pervert the norm enforcement
process because they involve the application of terms or phrases to things that
don’t realize the presupposed values:
C13.P70 Constitutive meaning perversions A speaker Sperverts the meaning of a word w
just in case S’s use of wis presented as an enforcement or application of norms or
¹⁹This argument is made more precise in Lenman (2000).
OUP UNCORRECTED PROOF –FIRST PROOF, 18/2/2020, SPi
276
Comp. by: ANEESKUMAR A Stage : Proof ChapterID: 0004768053 Date:18/2/20
Time:20:57:41 Filepath:d:/womat-filecopy/0004768053.3D
Dictionary : OUP_UKdictionary 277
values that wexpressively presupposes, which erodes those very same norms or
values by being misapplied to an unsuitable referent.
C13.P71 A use of a word is a perversion when it is false that the word applies to what the
speaker intends to refer to, and the use of the word nonetheless has the illocut-
ionary expressive effect it constitutively has—it expresses a conative state to the
effect that what the speaker is referring to realizes a certain value or norm.
C13.P72 Anyone who accepts the utterance making this expressive presupposition will
take for granted that what is referred realizes that value. In other words, she will
take for granted the permissibility of the relevant evaluative attitudes, and will
possibly also share them. Those attitudes are often themselves motivating and
justifying reasons for jointly intentional action. The use is perversive in that what
is referred does not actually realize the presupposed value. And hence the motiv-
ated actions are not the appropriate actions to take towards the presumed referent.
C13.P73 What goes on with meaning perversions—like Putin’s‘the dictatorship of the
rule of law’, the Nazi ‘fanatic hero’, Medvedev’s‘managed democracy’, Soviet
elections as ‘free expressions of citizen will’—is that they take advantage of the
hybrid contents these phrases encode. There are, on the one hand, legitimate
referents of words like ‘democracy’,‘free’,‘elections’. The referents would be
picked by the set of concrete features that correspond to the first dimension of
dual-character concepts. But there are, on the other hand, the expressive norma-
tive or evaluative presupposition that these phrases express. This would corres-
pond to the set of abstract values those concrete features would realize.
C13.P74 Thus, a sincere literal use of ‘democracy’made by a competent speaker denotes
any form of political organization or government that displays some minimal
features (allowing for more and less fitting cases—from full to flawed democra-
cies). To accept that a state is democratic licenses certain presuppositions, not just
about features of its form of government, but also about how it realizes certain
desirable normative values. To take for granted that a form of government is
democratic is to take for granted that it is desirable as good. Mutatis mutandis, the
same can be said for ‘freedom’or ‘elections’.
C13.P75 Besides these illocutionary effects, there are additional probable harmful per-
locutionary effects of the use of the word w. The constitutive sense of meaning
perversions and their perlocutionary effects are related. One sense in which we can
pinpoint the nature of meaning perversions is that they have the harmful per-
locutionary effects they have precisely because they contribute to undermine the
presupposed norms or values. But the two senses are not equivalent and may not
be co-extensional. Some revisions that have harmful consequences will not be
constitutively perversive.
C13.P76 Take ‘free elections’: they are good things, they are essential for democracies, a
recognition of the citizens’sovereignty through their representation in the insti-
tutions of their countries. The expression has a positive expressive connotation.
OUP UNCORRECTED PROOF –FIRST PROOF, 18/2/2020, SPi
277
Comp. by: ANEESKUMAR A Stage : Proof ChapterID: 0004768053 Date:18/2/20
Time:20:57:41 Filepath:d:/womat-filecopy/0004768053.3D
Dictionary : OUP_UKdictionary 278
That positive aspect of the meaning of ‘free election’can be taken for granted, at
least for a good amount of time, to manipulate people. By describing the ritual
practice of the so-called elections in the Soviet Union, the Soviet regime was
perverting the meaning of ‘free elections’. The ritual that was called a ‘free election’
did not display any of the concrete features that democracies must minimally
exemplify. And hence, that ritual did not actually realize the positive normative
value of free elections.
C13.P77 The normative presupposition expressed by ‘free election’, which is taken for
granted as part of the motivational conversational set, pragmatically contradicts
the actual application of the phrase to something that does not meet the minimal
constraints for being a free election. By doing this, the Soviet regime was eroding
the positive value of ‘free election’, while normalizing the new undemocratic
practice. People lived under a pragmatic contradiction between the official nor-
mative ideal—democracies are valuable and participating in them is desirable—
and the reality they were forced to inhabit. And people who are deprived of the
means to appropriately describe the situation they live in are people that are
deprived of the means to appropriately address it, and are more easily controlled
by authoritarian regimes, as both Klemperer and Gessen testify. These are some of
the harmful perlocutionary effects of meaning perversions.
C13.P78 Meaning perversions contrast with code words. Code words induce the accept-
ance into the motivational common ground of evaluative dispositions, plans, and
norms, that conflict with pre-accepted dispositions/plans/norms that are part of a
shared conversational score. But the expression of those values or norms is not
encoded (not even as a presupposition) in the meaning of a code word.²⁰This is
exemplified in uses of code words lacking a negative racial connotation, for
instance ‘we are doing a census of the people living in the inner-city to determine
the investment in new and much-needed pre-schools and centres for primary
medical attention’. Moreover, even those uses that do implicate a racial connota-
tion denote their proper referent in the world, e.g. ‘inner-city’refers to actual
urban areas. That’s why plausible deniability is possible—one can always point out
that what one said is factually true, and since the racial connotation is not part of
encoded meaning, there is no contradiction in that denial.
C13.P79 In contrast, the normative and evaluative connotation of a meaning perversion
is encoded in the word’s meaning, and thus it is automatically taken for granted as
part of the conversational record. Hence, speech that perverts meaning is easily
accepted in a context, since it aligns with what is already accepted as part of the
shared motivational set. This explains Klemperer’s (1957/2000) description: ‘as
soon as this concept (heroism) was touched upon, everything became blurred in
the fog of Nazism’. However, meaning perversions are not used to denote their
²⁰For discussion, see Khoo (2017).
OUP UNCORRECTED PROOF –FIRST PROOF, 18/2/2020, SPi
278
Comp. by: ANEESKUMAR A Stage : Proof ChapterID: 0004768053 Date:18/2/20
Time:20:57:41 Filepath:d:/womat-filecopy/0004768053.3D
Dictionary : OUP_UKdictionary 279
proper referents in the world. As a result, the evaluative connotation conflicts with
the use the word to talk about an improper referent that doesn’tfit the value it is
presented as realizing.
C13.P80 As a result of these differences, the argumentative strategies of speakers who use
code words and those who use meaning perversions differ. Code words allow
plausible deniability: ‘I wasn’t saying anything about race! I was talking about
criminality in certain urban areas’. Meaning perversions, in contrast, allow for
rhetorical norm-enforcement questions: ‘How can you be against our freedom?’,
‘How can you oppose democracy?’,‘How can you believe what an enemy of the
people says?’Interlocutors are, naturally, expected to reject that they are against
freedom, or that they oppose democracy, or that they believe their enemies. They
are also now pressed to accept that an improper referent has the concrete features
that realize the values that they are already take for granted.
C13.P81 It may be hard to know in advance if a given revision will have harmful effects,
although we can try to fend them off for instance by combining, as I suggested,
reflective equilibrium methods guided by defensive pessimism and honest humil-
ity. This means essentially that we should expect bad consequences to occur (and
try to foresee them to the best of our ability) and be prepared to revise our
expectations and plans if necessary. But it is not difficult to know in advance if a
given revision is harmful if it is a meaning revision in the second, constitutive,
sense. We can know, by understanding the mechanisms at play, that referring to
an improper referent and accepting that it fits the values or norms, it very plainly
cannot fit.
C13.P82 When a revision is a perversion in the constitutive sense, it is especially difficult
to resist: to give a reply to propagandistic calls for respecting the will of the people,
or for holding free elections, or for taking back control. Yet, in the mouth of many
demagogues, these are meaning perversions: uses of ‘the people’that exclude most
of the people, of ‘free elections’that are neither free nor an exercise in autonomous
individual choice, of ‘take back control’that give away control.
C13.P83 How do we resist meaning perversions, or spot them? Any direct criticism
invites replies like ‘how can you be against my freedom?’,or‘how can you be
against the people?’An interlocutor is left speechless, since in normal circum-
stances a normal reply would be obviously ‘No! I’m not against the people, and
I’m not against free elections’. Those rhetorical questions are effective to advance a
meaning perversion because they seem to reinforce shared norms, while eroding
them. Theorists who are interested in advancing a revisionist project can easily be
unaware that they we are putting forward a perversion instead of an amelioration.
If, as Maynard and Benesch (2016) put it, people are often moved by ‘deep and
unreflected feelings that something feels “good”or “bad”’, which induce ‘positive
moral self-appraisal’, theorists may resist taking the extra step required to disen-
tangle their conviction in the intended good results and the appraisal that the
OUP UNCORRECTED PROOF –FIRST PROOF, 18/2/2020, SPi
279
Comp. by: ANEESKUMAR A Stage : Proof ChapterID: 0004768053 Date:18/2/20
Time:20:57:41 Filepath:d:/womat-filecopy/0004768053.3D
Dictionary : OUP_UKdictionary 280
misapplication will erode the very same values or normative principles we believe
we are promoting.²¹
C13.P84 This is a small step forward in delimitating the scope of the answer to the
question how can we assess the legitimacy of ameliorative projects. We can
complement the discussion about why it is problematic to engineer race-talk,
for instance. Recall that race-talk has had, and continues to have, hugely harmful
consequences. Ordinary race concepts are presumed to track a set of concrete
biological features that identify a natural class, which turns out to be empty. The
values that are supposed to be realized by those concrete features present people as
less, or more, deserving of consideration, social standing, or respect as persons by
being presumed to exhibit certain biological features. We can try to ameliorate
what ‘race’refers—for instance, to argue that that it refers to ancestor-descendant
sequences of breeding populations that share a common origin. But the normative
presuppositions that come with race-talk are not amenable to amelioration by fiat,
particularly when there have been billions killed or enslaved on account of the
negative values they were presumed to realize. The decision to do away with talk of
Rasse or raça acknowledge this heavy burden. Moreover, there are alternative
ways to address resilient racism, and to pursue useful lines of inquiry in biology,
anthropology, or sociology.
C13.S9 13.5 Closing Remarks
C13.P85 I have argued for the importance of adding the question How can we assess the
legitimacy of ameliorative projects? to the list of questions that conceptual engin-
eering and conceptual ethics seek to address. I have also argued that our focus on
the normative constraints of conceptual engineering should emphasize the moral
and political legitimacy of the projects pursued. I was guided, in this regard, by
troubling historical lessons. By trying to characterize meaning perversions, I hope
to have offered a way of demarcating what a legitimate conceptual amelioration
cannot be. The two senses of meaning perversion offered demarcate the set of
morally permissible meaning or conceptual engineering projects. Revisionary
projects that are permissible are the set of meaning revisions that are not perver-
sions. That means that they are meaning revisions that (i) don’t have harmful
consequences, and that (ii) do not misapply a word to something unfitting the
values presupposed by the use of the word.
²¹ Bicchieri (2005, 2017) work models how shared social norms that are followed in a society rely on
people’s normative and empirical expectations with respect to what others will do and feel, and the
conditions under which normative change occurs.
OUP UNCORRECTED PROOF –FIRST PROOF, 18/2/2020, SPi
280
Comp. by: ANEESKUMAR A Stage : Proof ChapterID: 0004768053 Date:18/2/20
Time:20:57:41 Filepath:d:/womat-filecopy/0004768053.3D
Dictionary : OUP_UKdictionary 281
C13.S10 References
Andreasen, R. O. (1998). A new perspective on the race debate. British Journal for the
Philosophy of Science,49(2), 199–225.
Andreasen, R. O. (2000). Race: Biological reality or social construct? Philosophy of
Science,67(3), 666.
Andreasen, R. O. (2004). The cladistic race concept: A defense. Biology and Philosophy,
19(3), 425–42.
Andreasen, R. O. (2005). The meaning of ‘race’.Journal of Philosophy,102(2),94–106.
Andreasen, R. O. (2020). Conceptual fragmentation and the use of ‘race’in scientific
theorizing. In T. Marques & Å. Wikforss (Eds.), Shifting concepts (pp. XX–XX).
Oxford: Oxford University Press.
Appiah, K. A. (2006). How to decide if races exist. Proceedings of the Aristotelian
Society,106(3), 363–80.
Bicchieri, C. (2005). The grammar of society: The nature and dynamics of social norms.
Cambridge: Cambridge University Press.
Bicchieri, C. (2017). Norms in the wild: How to diagnose, measure, and change social
norms. Oxford: Oxford University Press.
Björnsson, G. & Hess, K. (2017). Corporate crocodile tears? On the reactive attitudes of
corporate agents. Philosophy and Phenomenological Research,94(2), 273–98.
Cepollaro, B. & Stojanovic, I. (2016). Hybrid evaluatives: In defense of a presupposi-
tional account. Grazer Philosophische Studien,93, 458–88.
Charlow, N. (2016). Decision theory: Yes! truth conditions: No! In N. Charlow &
M. Chrisman (Eds.), Deontic modality. Oxford: Oxford University Press.
Couto, A. (2016). Reactive attitudes, forgiveness, and the second-person standpoint.
Theory and Moral Practice,19, 1309–23.
D’Arms, J., & Jacobson, D. (2000). The moralistic fallacy: On the ‘appropriateness’of
emotions. Philosophy and Phenomenological Research,61(1),65–90.
Deonna, J., & Teroni, F. (2015). Emotions as attitudes. Dialectica,69, 293–311.
Diaz-Leon, E. (2015). What is social construction? European Journal of Philosophy,23
(4), 1137–52.
Figes, O. (2002). The whisperers: Private life in Stalin’s Russia. London: Penguin Books.
Gessen, M. (2017). The autocrat’s language. The New York Review of Books. 13 May
2017.
Glasgow, J. (2003). On the new biology of race. Journal of Philosophy,100(9), 456–74.
Goldman, D. (2014). Modification of the reactive attitudes. Pacific Philosophical
Quarterly,95(1),1–22.
Haslanger, S. (2003). Social construction: The ‘debunking’project. In F. Schmitt (Ed.),
Socializing metaphysics (pp. 301–25). Oxford: Rowman and Littlefield.
OUP UNCORRECTED PROOF –FIRST PROOF, 18/2/2020, SPi
281
Comp. by: ANEESKUMAR A Stage : Proof ChapterID: 0004768053 Date:18/2/20
Time:20:57:41 Filepath:d:/womat-filecopy/0004768053.3D
Dictionary : OUP_UKdictionary 282
Haslanger, S. (2006). What good are our intuitions? Philosophical analysis and social
kinds. Aristotelian Society Supplementary Volume,80(1),89–118.
Haslanger, S. (2012). Resisting reality: Social construction and social critique. Oxford:
Oxford University Press.
Haslanger, S. (2018). Nonideal, critical, ameliorative, realist social ontology: Say what?
Keynote lecture at Social Ontology 2018: The 11th Biennial Collective Intentionality
Conference,22–25 August 2018, Boston, MA.
Haslanger, S. (2020). Going on, not in the same way. In D. Plunket, A. Burgess, &
H. Cappelen (Eds.), Conceptual ethics and conceptual engineering (pp. XX–XX).
Oxford: Oxford University Press.
Jeshion, R. (2016). Slur creation, bigotry formation: The power of expressivism.
Phenomenology and Mind,11, 130.
Kagan, S. (1998). Normative ethics. Boulder, CO: Westview Press.
Khoo, J. (2017). Code words in political discourse. Philosophical Topics,45(2),33–64.
Kitcher, P. (2007). Does ‘race’have a future? Philosophy and Public Affairs,35(4),
293–317.
Klemperer, V. (1957/2000). The language of the Third Reich—LTI Lingua Tertii
Imperii, A philologist’s notebook. London: Continuum.
Knobe, J., & Prasada, S. (2011). Dual character concepts. In Proceedings of the 33rd
Annual Conference of the Cognitive Science Society (pp. 2965–70). Boston, MA:
Cognitive Science Society.
Knobe, J., Prasada, S., & Newman, G. (2013). Dual character concepts and the
normative dimension of conceptual representation. Cognition,127(2), 242–57.
Kutz, C. (2000). Acting together. Philosophy and Phenomenological Research,61(1),
1–31.
Langton, R. (2012). Beyond belief: Pragmatics in hate speech and pornography. In
M. K. McGowan and I. Maitra (Eds.), Speech and harm: Controversies over free
speech (pp. 72–93). Oxford: Oxford University Press.
Lawford-Smith, H. (2013). Understanding political feasibility. Journal of Political
Philosophy,21(3), 243–59.
Lenman, J. (2000). Consequentialism and cluelessness. Philosophy and Public Affairs,
29(4), 342–70.
Lippert-Rasmussen, K. (2013). Born free and equal? A philosophical inquiry into the
nature of discrimination. Oxford: Oxford University Press.
Ludwig, D. (2018). How race travels: Relating local and global ontologies of race.
Philosophical Studies,176(10), 2729–50.
Machery, E., Mallon, R., Nichols, S., & Stich, S. (2009). Against arguments from
reference. Philosophy and Phenomenological Research,79(2), 332–56.
Manne, K. (2017). Down girl: The logic of misogyny. Oxford: Oxford University Press.
OUP UNCORRECTED PROOF –FIRST PROOF, 18/2/2020, SPi
282
Comp. by: ANEESKUMAR A Stage : Proof ChapterID: 0004768053 Date:18/2/20
Time:20:57:41 Filepath:d:/womat-filecopy/0004768053.3D
Dictionary : OUP_UKdictionary 283
Marques, T. and M. García-Carpintero (2019). Really expressive presuppositions and
how to block them. Grazer Philosophischen Studien. Forthcoming.
Maynard, J. and S. Benesch (2016). Dangerous speech and dangerous ideology: An
integrated model for monitoring and prevention. Genocide Studies and Prevention:
An International Journal,9(3),70–95.
Mendelberg, T. (2001). The race card: Campaign strategy, implicit messages, and the
norm of equality. Princeton, NJ: Princeton University Press.
Mulligan, K. (1998). From appropriate emotions to values. Monist,81(1), 161–88.
Norem, J. (2001). Defensive pessimism, optimism, and pessimism. In E. Chang (Ed.),
Optimism & pessimism: Implication for theory, research, and practice (pp. 77–100).
Washington DC: American Psychological Association.
Podosky, P. C. (2018). Ideology and normativity: Constraints on conceptual engineer-
ing. Inquiry: An Interdisciplinary Journal of Philosophy [online], 1–15.
Portner, P. (2016). Imperatives. In M. Aloni and P. Dekke (Eds.), Cambridge handbook
of semantics (pp. 593–628). Cambridge: Cambridge University Press.
Rawls, J. (1971). A theory of justice. Cambridge, MA: Harvard University Press.
Salmela, M., & Michiru, N. (2016). Collective emotions and joint action. Journal of
Social Ontology,2(1),33–57.
Roberts, C. (2012) Information structure in discourse: Towards an integrated formal
theory of pragmatics. Semantics & Pragmatics,49(5),1–69. Originally published in
J. Yoon & A. Kathol (Eds.) (1998), OSU Working Papers in Linguistics, vol. 49.
Salmela, M., & von Scheve, C. (2017). Emotional roots of right-wing political popu-
lism. Social Science Information,56(4), 567–95.
Saul, J. (2017). Racial figleaves, the shifting boundaries of the permissible, and the rise
of Donald Trump. Philosophical Topics,45(2), n.p.
Saul, J. (2018). Dogwhistles, political manipulation and philosophy of language. In
D. Fogal, D. W. Harris, & M. Moss (Eds.), New work on speech acts (pp. 360–83).
Oxford: Oxford University Press.
Simion, M. (2017). The should in conceptual engineering. Inquiry: An Interdisciplinary
Journal of Philosophy,61(8), 914–28.
Snyder, T. (2017a). The deliberate starvation of millions in Ukraine: Review of Anne
Applebaum’sGreat Famine—Stalin’s War on Ukraine.The Washington Post 3
November 2017.
Snyder, T. (2017b). On tyranny: 20 lessons for the 20th century. New York: Duggan
Books.
Stalnaker, R. (1978). Assertion. Syntax and semantics (New York Academic Press),9,
315–32.
Stamp, N. (2018). Women with heart diseases are dismissed and it’s killing them. The
Guardian. https://www.theguardian.com/commentisfree/2018/jun/14/women-
with-heart-diseases-are-dismissed-and-its-killing-them.
OUP UNCORRECTED PROOF –FIRST PROOF, 18/2/2020, SPi
283
Comp. by: ANEESKUMAR A Stage : Proof ChapterID: 0004768053 Date:18/2/20
Time:20:57:41 Filepath:d:/womat-filecopy/0004768053.3D
Dictionary : OUP_UKdictionary 284
Stanley, J. (2015). How propaganda works. Princeton, NJ: Princeton University Press.
Stemplowska, Z., & Swift, A. (2012). Ideal and non-ideal theory. In D. Estlund (Ed.),
The Oxford handbook of political philosophy (pp. 373–89). Oxford: Oxford
University Press.
Strawson, P. (1962/2008). Freedom and resentment. Proceedings of the British
Academy,48,1–25. Reprinted in Freedom and resentment and other essays, 2nd
edn (pp. 1–28). New York: Routledge.
Tarski, A. (1943). The semantic conception of truth: And the foundations of semantics.
Philosophy and Phenomenological Research,4(3), 341–76.
Tirrell, L. (2012). Genocidal language games. In I. Maitra & M. K. McGowan (Eds.),
Speech and harm: Controversies over free speech (pp. 174–221). Oxford: Oxford
University Press.
Wilkins, D. B., Appiah, K. A., & Gutmann, A. (1998). Color conscious: The political
morality of race. Princeton, NJ: Princeton University Press.
Williams, B. (1979). Internal and external reasons. In R. Harrison (Ed.), Rational
action (pp. 101–13). Cambridge: Cambridge University Press.
OUP UNCORRECTED PROOF –FIRST PROOF, 18/2/2020, SPi
284