Your article is published under the Creative
Commons Attribution license which allows
users to read, copy, distribute and make
derivative works, as long as the author of
the original work is cited. You may selfarchive this article on your own website, an
institutional repository or funder’s repository
and make it publicly available immediately.
1 23
Erkenn
DOI 10.1007/s10670-015-9793-3
ORIGINAL ARTICLE
Ontological Choices and the Value-Free Ideal
David Ludwig1
Received: 24 May 2015 / Accepted: 19 November 2015
Ó The Author(s) 2015. This article is published with open access at Springerlink.com
Abstract The aim of this article is to argue that ontological choices in scientific
practice undermine common formulations of the value-free ideal in science. First, I
argue that the truth values of scientific statements depend on ontological choices.
For example, statements about entities such as species, race, memory, intelligence,
depression, or obesity are true or false relative to the choice of a biological, psychological, or medical ontology. Second, I show that ontological choices often
depend on non-epistemic values. On the basis of these premises, I argue that it is
often neither possible nor desirable to evaluate scientific statements independently
of non-epistemic values. Finally, I suggest that considerations of ontological choices
do not only challenge the value-free ideal but also help to specify positive roles of
non-epistemic values in an often neglected area of scientific practice.
1 Introduction
Values are ubiquitous in scientific practice and play an important role in research
from the early planning stage to the dissemination of final results. Proponents of the
Value-Free Ideal (VFI) do not endorse the highly implausible claim that values
should be eliminated from science altogether but rather more specific claims about
the illegitimacy of non-epistemic values in the evaluation or justification of
scientific theories (e.g. Betz 2013; Hudson 2015; Lacey 2010; Schurz 2013; Sober
2007). Although there is not one authoritative formulation of VFI, the general idea
can be motivated by a distinction between value questions and fact questions. For
example, consider Max Weber’s classical defense of VFI in the social sciences
& David Ludwig
d.j.ludwig@vu.nl; davidundludwig@gmail.com
1
Vrije Universiteit Amsterdam, De Boelelaan 1105, 1081 HV Amsterdam, The Netherlands
123
D. Ludwig
(Weber 1917). Weber offers a detailed description of the ubiquity of values in
academia but still insists that we have to understand scientific inquiry in terms of a
distinction between two fundamentally different types of questions:
1.
2.
fact questions: what is the case in the world?
value questions: what should be the case in the world?
Even if scientific practice is obviously entangled with values in numerous ways,
it still remains plausible to separate fact questions and value questions in the sense
of (1) and (2). It seems to be only a small step from this distinction between fact
questions and value questions to the ideal of VFI. If science is concerned with facts
in the sense of (1) while values in the sense of (2) have no bearing on what is the
case in the world, then the ‘‘epistemic integrity of science’’ (Ruphy 2006) seems to
require that we avoid non-epistemic values in the evaluation and justification of
scientific theories.
Far from endorsing the highly implausible claim that values can be eliminated
from science altogether, proponents of VFI subscribe to the far more specific idea
that non-epistemic values should play no role in evaluating or justifying scientific
theories (cf. Elliott and McKaughan 2014; Büter 2015). There are at least four
common arguments that are directed against VFI in this more narrow sense. The
first argument is based on the general assumption that the underdetermination of
theories by evidence leaves room for non-epistemic values in theory choice (e.g.
Longino 2002; cf. Brown 2013). The second argument challenges VFI by pointing
out that scientific inferences often involve a high degree of uncertainty and by
suggesting that an evaluation of the ‘‘inductive risk’’ in theory choice requires nonepistemic values (e.g. Douglas 2009; John 2015). The third argument questions the
very distinction between epistemic or non-epistemic values. While philosophers of
science have often assumed that only the former are legitimate in theory choice
because they ‘‘promote the attainment of truth’’ (Steel 2010, p. 17), a breakdown of
the distinction would undermine VFI (Rooney 1992; Longino 1996; Machamer and
Douglas 1999). According to the fourth argument, scientists often have to employ
‘‘thick ethical concepts’’ in which factual and normative components are inextricably entangled (e.g. Putnam 2002). If it is not possible to separate these aspects,
then it is neither possible to evaluate or justify scientific theories independently of
values.
The aim of this article is to develop a different argument against VFI that is based
on a specification of the roles of non-epistemic values in scientific ontologies. In a
first step, I argue that the truth values of scientific statements depend on ontological
choices. For example, statements about entities such as species, race, memory,
intelligence, depression, or obesity are true or false relative to the choice of a
biological, psychological, or medical ontology (Sects. 2, 3). In a second step, I
argue that ontological choices often depend on non-epistemic values (Sect. 4).
These two steps imply that the truth values of scientific statements often depend on
non-epistemic values. In a third step, I try to show that this value-dependency of the
truth values of scientific statements undermines VFI (Sect. 5) as it is not possible to
avoid value-laden ontological commitments (Sect. 6). Finally (Sect. 7), I suggest
123
Ontological Choices and the Value-Free Ideal
that a consideration of non-epistemic values in scientific ontologies does not only
provide a negative argument against VFI but also makes more constructive
suggestions about legitimate roles of non-epistemic values in scientific theories.
2 Ontological choices
The aim of this section is to develop the idea that the truth values of scientific
statements depend on ontological choices. This general claim is best discussed on
the basis of more specific examples. Consider the following statements:
(a)
(b)
(c)
(d)
(e)
(f)
(g)
At least forty species of orchids became extinct in Indonesia during the
twentieth century.
There are two different tiger species in the San Diego Zoo.
The human short-term memory has a capacity of seven plus or minus two
units.
35.9 % of adults in the US are obese.
Grief is a natural response to death or loss and therefore not a depression.
Members of population a are on average more intelligent than members of
population b.
Andy Warhol had an I.Q. of 86.
Although I want to claim that the truth values of all statements (a)–(g) depend on
ontological choices, my case is most conveniently illustrated by the biological
examples (a) and (b). Consider the question whether there are two different tiger
species in the San Diego Zoo. At first, it seems that this question has a pretty
straightforward empirical answer that can be determined by a biologist in a valuefree manner: she simply has to go to the zoo and look whether there are two tiger
species.
Unfortunately, the situation is more complex because tigers raise taxonomic
issues that are not always answered in the same way by biologists. Let us assume
that there are Bengal Tigers (Panthera tigris tigris) and Sumatran Tigers (Panthera
tigris sumatrae) in the San Diego Zoo. Most biologists assume that there is only one
tiger species (Panthera tigris) and that the Bengal Tiger and the Sumatran Tiger
only qualify as subspecies. According the influential ‘‘biological species concept’’,
two populations belong to the same species if they are able to produce fertile
offspring of both sexes (Mayr 1969). Bengal Tigers and Sumatran Tigers are able to
produce fertile offspring of both sexes and a biologist who uses the biological
species concept will therefore conclude that (b) is false.
However, there are many alternatives to the biological species concept. For
example, the so-called ‘‘phylogenetic species concept’’ defines species as ‘‘the
smallest diagnosable cluster of individual organisms within which there is a parental
pattern of ancestry and descent’’ (Cracraft 1983, p. 170; LaPorte 2009, pp. 70–76).
Bengal and Sumatran Tigers are different clusters in this sense and proponents of
the phylogenetic species concept will therefore conclude that (b) is true. The truth
123
D. Ludwig
value of (b) therefore depends on the account of species that is used in developing a
biological ontology.
A similar point can be made with respect to orchids and the statement (a).
Orchids constitute a highly diverse biological family of flowering plants with more
than 20,000 species. The exact number of orchid species depends on the criteria for
species membership. As there is not one universally accepted set of criteria,
biologists will disagree on the exact number of orchid species. This disagreement
also affects more specific scientific statements. If two biologists use different
criteria for species membership of orchids, they may very well also end up
disagreeing on the truth value of (a).
The examples of tigers and orchids illustrate my claim that the truth values of
scientific statements depend on ontological choices. Different accounts of species
imply the existence of different biological kinds and in this sense different
biological ontologies. Furthermore, it makes sense to speak of ‘‘choice’’ in this
context because scientists have to decide which account they want to work with.
While the label ‘‘ontological choice’’ may be controversial in analytic metaphysics,
I therefore assume that it refers to a common phenomenon in scientific practice (cf.
Ludwig 2015d, chapter 4).
While biological kinds provide convenient examples for the claim that the truth
values of scientific statements depend on ontological choices, similar points can be
made with respect to other scientific statements including (c)–(g). (c) is one of the
most famous results of early cognitive psychology and has been formulated in
George Miller’s landmark article ‘‘The Magical Number Seven, Plus or Minus
Two’’ (1956). Although Miller’s hypothesis is based on empirical discoveries in
experimental psychology, one can still argue that the truth value of (c) depends on
ontological choices as (c) presupposes a certain account of short-term memory.
Consider debates about externalist accounts of memory and cognition in general
(e.g. Clark and Chalmers 1998). Externalists argue that cognitive processes such as
memory can be extended through external devices from paper notebooks to digital
media (Ludwig 2015d). Given such an externalist account, it does not make sense to
claim that the short-term memory is restricted to seven plus or minus two units.
Instead, the capacity of the short-term memory changes dynamically with the
cognitive extensions that are available in the environment. (c) therefore turns out to
be true only under the assumption of internalist ontology and should be considered
false under the assumption of externalism.
Given my discussion of (a)–(c), it is not hard to see in what sense the truth values
of (d)–(g) may also depend on ontological choices. Whether 35.9 % of adults in the
US are obese obviously depends on scientists’ accounts of obesity and whether grief
(and if so, what kinds of grief) qualifies as depression depends on psychiatrists’
accounts of depression (Intemann 2001; Zachar 2015). The same argument can be
made with respect to intelligence and I.Q. One crucial question is what cognitive
abilities should be considered relevant for intelligence. Furthermore, psychologists
disagree on the question whether we should assume the existence of a general
intelligence at all and proponents of multiple intelligences theories have proposed
large variety of more specialized intelligences such as ‘‘verbal intelligence’’,
‘‘mathematical intelligence’’, or ‘‘spatial intelligence’’ (Ludwig 2014b). Clearly the
123
Ontological Choices and the Value-Free Ideal
truth values of (f) and (g) depend on these issues of psychological ontologies. A
psychologist who works with a framework of multiple intelligences will reject both
statements on the basis of her rejection of a general intelligence. Furthermore, even
psychologists who agree on the assumption of a general intelligence may disagree
on the criteria for intelligence and therefore end up disagreeing on the truth values
of (f) and (g).
To sum up, scientists often have to choose between different accounts of entities
such as species, memory, obesity, depression, or intelligence. These different
accounts imply different scientific ontologies in the sense that they imply the
assumption of different entities. Finally, these different ontologies clearly affect the
truth values of statements such as (a)–(g).
3 Explanatory Interests and Natural Kinds
In the last section, I argued that scientists often have to choose between different
ontologies that imply different attitudes towards scientific statements such as (a)–
(g). This claim raises the obvious question of how scientists choose between
different ontologies. Philosophers of science commonly answer this question by
pointing to different explanatory interests that shape ontological choices in scientific
practice (e.g. Brigandt 2009; Leonelli 2012; Ruphy 2013; Ludwig 2014a; Danks
2015). Insofar as scientists are concerned with different explanatory projects, they
will often find different ontologies useful for their research.
Kitcher’s (1984) influential discussion of species pluralism provides a classical
illustration of this strategy by discussing the explanatory benefits and limits of
different species concepts: One commonly mentioned achievement of Mayr’s
biological species concept is that it allows to distinguish between two different
species of mosquitoes within the Anopheles complex and therefore provides an
explanation of the distribution of malaria in Europe (e.g. Mayr 1969). This
explanation would not be possible in the context of traditional morphological
species concepts due to the morphological similarities of mosquitoes within the
Anopheles complex. While this example illustrates why the biological species
concept is preferable in the context of some research projects, there are other areas
of biological research that suggest different species concepts. For example, the
biological species concept is clearly not an attractive choice for biologists who work
on asexual organisms.
The constitutive role of explanatory interests in ontological choices suggests that
the truth values of statements such as (a) and (b) depend on explanatory interests of
biologists: the truth values of (a) and (b) depend on the choice of a species concept.
The choice of a species concept depends on the explanatory interests of biologists.
As ‘‘dependency’’ in the present sense is clearly a transitive relation, the truth values
of (a) and (b) also depend on the explanatory interests of biologists.
Furthermore, one can extend this claim beyond biological ontologies to
statements such as (c)–(g). Recall that (c) will be true in the context of an
internalist ontology that restricts the short-term memory to biologically realized
processes but false in the context of an externalist ontology that accepts that external
123
D. Ludwig
media such as notebooks or cell phones can become part of someone’s short-term
memory. It seems reasonable to assume that the choice of an externalist or
internalist ontology also largely depends on explanatory interests of cognitive
scientists. An internalist account of memory as it has been used by Miller (1956) in
his formulation of (c) can be justified in the context of many traditional approaches
in cognitive psychology that are primarily concerned with the internal machinery of
cognition as it is realized by the human brain. Given this research interest in the
internal ‘‘brainbound’’ structure of cognition, there is little use for cognitive
processes that extend beyond the organism. Although the internalist account has
been a defining aspect of cognitive psychology in the second half of the twentieth
century, there are also research projects that suggest a different ontological
framework. This is most obvious in the case of research on problem solving
strategies that is not concerned with properties of internal mechanisms but with
human behavior in complex environments. Problem solving usually relies on
internal and external resources and humans often use both kinds of resources in
surprisingly similar ways. Externalist often motivate their ontological choices by
referring to research on problem solving such as Gray and Fu’s experiments (2004)
that illustrate how little the external-internal distinction matters in the context of
many cognitive tasks. It therefore seems reasonable to assume that the choice for a
specific account of memory also reflects the explanatory interests of scientists
(Pöyhönen 2014; Ludwig 2015c, chapter 4.2).
The case for ontological pluralism in philosophy of science suggests a rather
straightforward justification for my claim that the truth values of scientific
statements (a)–(g) depend on ontological choices. A statement such as (a) may be
true under the assumption of one ontology A but false under the assumption of
another ontology B. If we assume that the choice between the ontologies A and B
depends on explanatory interests, it seems that we should also accept that the truth
values of many scientific statements depends on these epistemically driven choices.
Still, one may worry that my endorsement of a pluralist account of scientific
ontologies has been too quick as it does not adequately consider that some
ontologies may be more successful than others in ‘‘carving nature at its joints’’.
Indeed, my claim would become trivial in the light of a philosophical position that
considers the boundaries of entities to be conventional constructions and rejects any
realism whatsoever. As recent controversies about natural kinds illustrate (e.g.
Khalidi 2013; Ludwig 2015b; Slater 2015), however, a general conventionalism
about scientific ontologies is highly controversial. Insofar my argument depends on
such a conventionalism, it will arguably be considered a non-starter by most
proponents of VFI.
While a conventionalist account of scientific ontologies would be indeed
sufficient for my claim that truth values of scientific statements depend on
ontological choices, conventionalism is by no means necessary. In fact, a pluralist
account of ontological choices is compatible with most of the currently popular
accounts of natural kinds in philosophy of science. This compatibility is most
obvious in the case of ‘‘epistemology only’’ (MacLeod 2010) formulations of
natural kinds. The idea that the naturalness of scientific kinds depends on their
epistemic relevance introduces at least some non-conventional elements (cf.
123
Ontological Choices and the Value-Free Ideal
Brigandt 2009). The choice between ontologies is not conventional in a similar
sense as the choice between signifiers such as ‘‘snow’’, ‘‘Schnee’’, and ‘‘neige’’ is
conventional as it will often turn out that one ontology is much more fruitful than
others in achieving our epistemic aims. Furthermore, the fruitfulness of an ontology
cannot be determined a priori but only on the basis of an open-ended empirical
inquiry. While epistemological accounts of natural kinds introduce non-conventional elements, they are still often compatible with my claim that truth values of
scientific statements depend on ontological choices. Indeed, if we would find that
one ontology is epistemologically superior across all contexts of application, we
may conclude that scientific practice does not involve ontological choices in a
substantial sense. However, examples such as different accounts of species illustrate
that scientists with different explanatory interests often find different ontologies
epistemically fruitful and therefore opt for different ontologies.
While purely epistemological accounts of natural kinds may be compatible with
my discussion of ontological choices, many philosophers have argued that any
convincing account of natural kinds will require at least some metaphysical depth.
Some kinds are epistemically more fruitful than others because they have certain
features. Many current accounts of natural kinds with metaphysical ambitions rely
on claims about property clustering—whether characterized as ‘‘stable property
clusters’’ in general (Slater 2015), causally unified property clusters (Khalidi 2013;
cf. Craver 2009), or more specifically homeostatic property clusters (Wilson et al.
2007).
However, even these more ambitious notions of natural kinds are arguably
compatible with my discussion of ontological choice. Again, consider my example
of orchid species (a) and two different ontologies A and B that imply different truth
values of (a). While it may be that only A or B provides an account of orchids that
satisfies characterizations of natural kinds as property clusters, it may also happen
that both ontologies pick out stable albeit somewhat different property clusters.
Indeed, decades of debates about different species ontologies have taught us that the
latter result is actually more realistic. The plurality of different species concepts is
so persistent precisely because there are different ways of picking out stable and
causally unified property clusters. For example, recall Kitcher’s case of a shift
towards the ecological species concept in the case of research on interactions
between asexual organisms in a coral reef. Both a focus on reproductive patterns
and on ecological niches will lead to the discovery of (partly overlapping) property
clusters but the relevance of discovered clusters depends on explanatory goals. A
metaphysical specification of natural kinds in terms of property clusters therefore
does not stand in the way of pluralist account of different ontological choices.
To sum up, my claim that the truth values of scientific statements depend on
ontological choices does not presuppose a general conventionalism about scientific
ontologies but is compatible with a variety of epistemological and metaphysical
characterizations of natural kinds. Of course, this does not mean that my proposal of
ontological choices is compatible with every possible account of natural kinds. Most
obviously, one could endorse a strong metaphysical realism that insists on exactly
one correct scientific ontology and one correct way of carving nature at its joints.
While such a monism retains a certain popularity in some areas of analytic
123
D. Ludwig
metaphysics (e.g. Sider 2011), it is usually limited to the ‘‘fundamental sciences’’
and not extended to the idea of one fundamental biological or psychological
ontology. The assumption that scientific practice in the life sciences comes with
ontological choices therefore only requires the (at the most) mildly controversial
assumption that we should not expect one biological or psychological ontology to be
preferable across all explanatory contexts.
4 Non-epistemic Values in Ontological Choices
I have argued that the truth values of many scientific statements depend on
ontological choices and that ontological choices depend on explanatory interests of
scientists. There is a tension between these claims and any variant of VFI that is
motivated by the allegedly clear distinction between fact questions and value
questions. Consider questions such as ‘‘How many species of orchids became
extinct in Indonesia during the twentieth century?’’ or ‘‘What is the capacity of the
human short-term memory?’’. While these questions are clearly fact questions in the
sense of Weber, the value-dependency of ontological choices suggests that they
cannot be understood as completely independent of value questions.
Proponents of VFI may attempt to solve this tension by emphasizing the
distinction between epistemic and non-epistemic values. Proponents of VFI are
happy to accept epistemic values in theory choices (e.g. McMullin 1982) and can
also accept epistemic values in ontological choices as long as social and other nonepistemic values remain excluded. A look at my examples (a)–(c) can further
strengthen the intuition that we can and should avoid non-epistemic values in
ontological choices. My examples of species and short-term memory suggest that
epistemic values such as explanatory power are crucial in ontological choices. A
plurality of equally legitimate ontologies has to be understood as a consequence of
explanatory power being dependent on explanatory interests: While the biological
species concept has proven to be extraordinarily helpful in some explanatory
contexts (e.g. the Malaria case), it evidently lacks explanatory power in other
contexts (e.g. asexual organisms). The same point can be made with respect to
memory research. Whether an internalist or externalist account of cognition has
more explanatory power crucially depends on explanatory interests. All of this
seems not to be a problem for proponents of VFI who are happy to admit that
epistemic values such as explanatory power are crucial for the evaluation of
scientific theories. As long as ontological choices are made solely on the
background of these epistemic values, my discussion of ontological choices does
not seem to pose a threat to VFI.
There are at least three problems with this defense of VFI. As I mentioned in the
introduction, some philosophers reject VFI by arguing that we cannot neatly
separate epistemic and non-epistemic values. While this objection may undermine
the idea of value-free ontological choices, it would also make my argument
dependent on a more general objection against VFI. I will therefore presuppose the
distinction between epistemic and non-epistemic values for the sake of argument.
But even if we grant proponents of VFI epistemic values, there remain two
123
Ontological Choices and the Value-Free Ideal
problems with the exclusion of non-epistemic values in ontological choices. First, I
will argue that explanatory interests are often shaped by non-epistemic factors and
cannot be reduced to epistemic values. Second, ontological choices are often not
only dependent on explanatory interests but also directly on non-epistemic factors.
Let us first consider the question whether explanatory interests can be understood
independently of non-epistemic values. A look at the vast and diverse literature on
‘‘the edges and boundaries of biological objects’’ (Haber and Odenbaugh 2009)
raises doubts that the explanatory interests that shape biological ontologies can be
understood as purely epistemic. In the case of some biological subdisciplines such
as conservation biology, it appears almost trivial that non-epistemic concerns shape
the explanatory interests of researchers. For example, the assumption of properties
such as the diversity, integrity, sustainability, or health of ecosystems (Callicott
et al. 1999; Wallace 2007) reflects explanatory interests of conservation biologists
but these explanatory interests reflect non-epistemic concerns regarding ecosystems.
Biologists in a society with different non-epistemic interests would have different
explanatory interests and would therefore end up with different ontologies.
The situation appears less clear in the case of other biological issues such as
different accounts of species. For example, the biological species concept is not
attractive in research on asexual organisms but it is at least not immediately clear
that an explanatory interest in asexual organisms should be understood as partly
non-epistemic. One possible response is to accept that explanatory interests in
biology are sometimes but not always shaped by non-epistemic factors. Therefore,
some biological ontologies reflect only epistemic interests while others depend on
non-epistemic factors.
However, one may also argue for the stronger claim that explanatory interests in
biology are always shaped by non-epistemic factors even if they are often not as
obvious as in the case of conservation biology. Explanatory interests have to be
evaluated on the background of socially shaped assessments of scientific significance (Kitcher 2011). For example, different species concepts come with different
explanatory strengths and weaknesses which need to be weighted by scientists (e.g.
Stanford 1995). While it is clear that biological taxonomies should make significant
distinctions, it is far from clear that there is a purely epistemic account of the
significance of a distinction. Consider, for example, the case of Mayr’s biological
species concept and its success in explaining the distribution of Malaria in Europe
by distinguishing between two species in the Anopheles complex. While the
example indicates an area where the biological species concept comes with
explanatory resources that are superior to (some of) its rivals, the importance of this
example also clearly depends on our non-epistemic interests. The ability to discern
different species within the Anopheles complex is an impressive example for the
explanatory benefits of the biological species concept because we care about human
diseases and therefore also about the distribution of Malaria.
Of course, the roles of non-epistemic values vary with their importance for an
ontological issue. In the case of species concepts, the relevance of non-epistemic
values may often appear obscure as it is at least not obvious that the choice of one
species concept over another has important social implications. The situation is
different if we consider the case of race instead of species. The rejection of human
123
D. Ludwig
races as legitimate biological entities is often associated with their limited
explanatory usefulness (Maglo 2011; Hochman 2013; Mncube 2015). For example,
Lewontin’s (1972) classical argument that most genetic variation is found in and not
between populations such as African or Asian was enormously influential in
questioning the explanatory value of races and rejecting their taxonomic legitimacy.
While a primarily epistemic case against race may work in biological taxonomy,
the situation is more complex in current debates about race in the biomedical
sciences. Racial realists in the biomedical sciences typically grant epistemic
limitations that have been stressed in the tradition of Lewontin but still insist on the
significance of race in socially relevant biomedical applications. Risch et al. (2002,
1), for example, argue for an ‘‘epidemiologic perspective on the issue of human
categorization in biomedical and genetic research that strongly supports the
continued use of self-identified race’’. Even if we accept Lewontin’s findings and
the limited explanatory use of ‘‘race’’ in taxonomy, Risch et al. insist that racial
distinctions capture relevant differences such as the prevalence of Mendelian
diseases that are often ‘‘found only in specific races (for example, cystic fibrosis and
hemochromatosis in Caucasians)’’ (2002, 9).
The issue of socially relevant explanations is not only prominent in defenses but
also in challenges of current biological accounts of race. First, critics of biological
accounts of race usually accept that races can be used as proxies for some
biomedically relevant genes but insist that they are unnecessarily coarse-grained and
unreliable proxies (e.g. Root 2003). Consider one of Risch et al.’s examples such as
hemochromatosis. Mutations that are causally responsible for hemochromatosis are
indeed more common in Caucasian populations such as Irish and Norwegian but the
use of ‘‘Caucasian’’ as a proxy for determining risk groups would create a large and
unnecessary amount of false positives (e.g. Italians or Italian-Americans). Maybe
even more worrying is that the use of race as a proxy can lead to false negatives in
cases such as sickle-cell anemia and children of Southern European ancestry
(Kaufman and Cooper 2010).
Second, philosophers who understand races as social and not as biological kinds
(e.g. Mills 1998; Haslanger 2012; Taylor 2013) could grant that biological accounts
of race come with some relevant explanatory resources but insist that they are
dwarfed by the explanatory resources of social accounts of race. Given that current
biological realists accept that the social realities of race cannot be explained in terms
of biological causes, a social constructionist can argue that it is more fruitful to think
of races as social instead of biological groups.
Of course, a proponent of biological realism could react to the second argument
by proposing a discipline-relative pluralism: ‘‘race’’ in the biomedical sciences
refers to biological populations and ‘‘race’’ in the social sciences refers to social
groups. Furthermore, a social constructionist could reject this proposal by pointing
out that socially caused differences between racialized groups are of crucial
importance not only in the social but also in the biomedical sciences. Lorusso and
Bacchini (2015; cf. Kaplan 2010), for example, have argued that ‘‘race-based’’
studies in the context of complex diseases are justified only if ‘‘self-identified race’’
is used as a proxy for social realities and not for genetic differences.
123
Ontological Choices and the Value-Free Ideal
The goal of my discussion is not to decide these complex issues but to point out
that explanatory interests are entangled with social concerns in the case of race.
Indeed, current controversies about the ontological status of race reflect questions
about the explanatory relevance of racial distinctions. A discussion of these
questions, however, does not provide a purely epistemic ground that would allow us
to decide the ontological status of races independently of non-epistemic concerns.
On the contrary, a substantial discussion of explanatory benefits and disadvantages
of different accounts of race requires careful consideration of our non-epistemic
aims. Recognizing this entanglement of epistemic and non-epistemic issues is of
crucial importance for understanding the structure of current debates about racial
ontologies. Risch et al. for example, combine their defense of a biological account
of race with the appeal to an ‘‘objective scientific perspective’’ (2003, p. 1) that is
supposed to validate their proposal and contrasts with the socially motivated claims
of eliminativists. Framing the debate by opposing ‘‘objective scientific’’ and
‘‘socially motivated’’ perspectives reflects a fundamental misunderstanding of the
debate as the usefulness of racial distinctions cannot be evaluated in isolation of our
non-epistemic aims.
So far I have argued that an exclusion of non-epistemic values from ontological
choices is not possible because explanatory interests of scientists are shaped by nonepistemic factors. However, there is a further problem: many ontological choices
not only depend on explanatory interests but also directly on non-epistemic values.
In order to illustrate this dependency on non-epistemic values, we do not even need
to move to new examples but can stick with the cases of species and race. In fact, it
is not difficult to find real-life examples of how choices between species ontologies
become directly entangled with non-epistemic concerns. For example, consider
phylogenetic accounts of species that follow cladistic approaches and often lead to
the recognition of far more species than traditional accounts of species in terms of
interbreeding or even morphology. The often staggering differences in the number
of recognized species do not only lead to epistemic but also straightforward nonepistemic implications of ontological choices. For example, a conservation biologist
who is concerned with the preservation of vulnerable species has to choose what
counts as a species and the choice will affect conservation practices. Zachos et al.
(2013, p. 1) provide illuminating examples of this as they argue that phylogenetic
accounts can lead to an inflation of ‘‘new species [that] creates an unnecessary
burden on the conservation of biodiversity.’’
While the case of phylogenetic species concepts illustrates that overly fine-grained
ontologies can be problematic for conservation biology, conservation concerns can also
lead in the other direction as the case of the Alabama Sturgeon illustrates (Scharpf 2000;
cf. Winther and Kaplan 2013, footnote 15). The Alabama sturgeon (Scaphirhynchus
suttkusi) is a critically endangered fish that exists only in a small portion of the lower
Alabama River. Recognizing the Alabama sturgeon as a distinct species would place it
on the Endangered Species List and lead to conservation measures that were strongly
opposed by an industrial interest group along the Alabama River. The consequence of
this opposition was a long political, scientific, and legal struggle over the status of the
Alabama sturgeon until a lawsuit in 2000 mandated protection of the fish.
123
D. Ludwig
The case of the Alabama sturgeon is a helpful reminder that biological taxa often
become agents in complex networks that include diverse entities such as the US Fish
and Wildlife Service, industrial lobby groups, legal documents and laws,
environmental activists, conservation policies and budgets, and so on. The
introduction or elimination of taxa as well reconsiderations of their boundaries
can have considerable effects on these networks (cf. Mol 1999; Latour 1999). It is
therefore a misunderstanding to think that the choice between different species
ontologies will exclusively reflect explanatory issues.
While direct influences of non-epistemic values may be somewhat unexpected in
the case of species, debates about the ontological status of race are more commonly
interpreted as normative (e.g. Mallon 2006; Gannett 2010; Winther and Kaplan
2013; Ludwig 2015a). I have already argued that controversies about racial
ontologies reflect non-epistemic concerns that shape explanatory interests regarding
human diversity. However, there are also more direct influences as illustrated by
debates about the effects of racial distinctions in public understanding of human
diversity (e.g. Morning 2011; Donovan 2014; Phelan et al. 2013). The worry that
biological accounts of race create a burden for science communication and science
education can be motivated by the observation that they require constant distinctions
between correlation and causation. On the one hand, membership to biologically
defined races will be correlated with social differences in education, criminal justice
system, health care, and so on. On the other hand, the biological properties will not
be causally relevant and can therefore easily lead to a conflation of correlation and
causation.
This worry about public misunderstanding of human diversity can be substantiated by recent research on the use of biology textbooks in high school education.
Donovan (2014) compared the impact of using racialized and non-racialized
textbooks on reasoning about human diversity. In his study, racialized textbooks
introduced diseases such as sickle cell anemia or cystic fibrosis in racial terms (e.g.
‘‘particularly common among people of African descent’’) while non-racialized
textbooks referred more generally to geographic variability (e.g. ‘‘carriers of sickle
cell anemia are more resistant to malaria, a common and serious disease in many
parts of the world’’) (Donovan 2014, p. 470). While the racialized passages did not
make any assumptions about cognitive or behavioral differences, Donovan’s data
suggests that ‘‘subtle references to race in the modern biology curriculum can lead
students to agree more strongly that races differ in complex human traits (e.g.
academic ability and artistic ability) because of genetics’’ (Donovan 2015, p. 1; cf.
Condit et al. 2004; Phelan et al. 2013 for complementary studies on the effects of
media reports).
Concerns about public (mis)understanding of race in science communication and
science education can provide another non-epistemic reason to reject biological
accounts of race. In this case, the non-epistemic concern does not shape explanatory
interests but rather directly affects an ontological choice. As a matter of empirical
description, it should be uncontroversial that ontological choices often involve a
complex mix of epistemic and non-epistemic considerations and can also be shaped
by direct social concerns such as public (mis)understanding of scientific research.
However, my discussion of ontological choices in the last sections suggests the
123
Ontological Choices and the Value-Free Ideal
stronger claim that there is no good reason to exclude these non-epistemic concerns
from debates about racial ontologies. Given the assumption that there is not one
metaphysically correct way of distinguishing between human populations, we will
have to choose between different ways of ‘‘carving up’’ human diversity. And given
that this is a genuine choice, it is unclear why we should consider non-epistemic
values secondary or even exclude them from decision-making processes. Epistemic
values are often considered to be of prior relevance in theory choice because they
are truth-indicating in the sense that they ‘‘promote the attainment of truth’’ (Steel
2010, p. 17). If there is not only one correct ontology, however, this line of
argument clearly fails prioritize epistemic values in ontological choices.
5 The Argument Against VFI
The considerations of the last sections suggest a quite simple argument against VFI:
Premise 1) The truth values of scientific statements depend on ontological
choices. Premise 2) Ontological choices depend on non-epistemic values. Conclusion 1) The truth values of scientific statements depend on non-epistemic values.
Premise 3) If the truth values of scientific statements depend on non-epistemic
values, then we should reject VFI. Conclusion 2) We should reject VFI.
In Sects. 2 and 3, I have argued for the first premise of the argument. I have
admitted that this premise presupposes the rejection of a strong metaphysical
realism that implies a monist account of biological and psychological ontologies. A
philosopher who is convinced that there is exactly one fundamental account of
entities such as species, race, memory, intelligence, obesity, or depression will
reject this premise and insist that scientific statements such (a)–(g) have truth values
that do not depend ontological choices but rather on fundamental ontological facts. I
have argued that even the endorsement of a substantial metaphysical account of
natural kinds is compatible with a moderate pluralism that requires ontological
choices in the life sciences. Proponents of VFI who do not want to presuppose such
a monism have to reject either the second or the third premise. However, it seems
that the second premise is well justified. First, ontological choices are shaped by
explanatory interests that are often inextricably entangled with non-epistemic
values. Second, the examples of species in conservation biology and of racial
ontologies show that ontological choices are also often directly shaped by nonepistemic concerns.
However, one may still reject the relevance of all of this for debates about VFI by
rejecting the third premise of the argument. Maybe truth values of scientific
statements such as (a)–(g) depend on non-epistemic values but this valuedependency does not threaten a moderate version of VFI. Maybe we can somehow
disentangle the value-laden component of ontological choice from the value-free
component of empirical evidence and limit VFI to the latter.
I take this objection seriously and I will discuss it in the next section. However,
first I want to point out the prima facie incompatibility of the claim that the truth
values of scientific statements depend on non-epistemic values and most common
formulations of VFI. Consider Weber’s strict distinction between fact questions
123
D. Ludwig
(what is the case in the world?) and value questions (what should be the case in the
world?) that provide the basis of his account of value-free science. While it is
certainly often helpful to distinguish between fact questions and value questions, the
claim that both types of questions are entirely independent from each other is highly
dubious in the case of my examples (a)–(g). Certainly, (a)–(g) are factual statements
in Weber’s sense but given their dependency on value-laden ontological choices,
they are clearly not completely independent from non-epistemic values.
Similar problems arise in the context of contemporary formulations of VFI that
are not explicitly based on the distinction between fact questions and value
questions but assume that the evaluation (e.g. Lacey 2010; Sober 2007) or
justification (e.g. Betz 2013) of scientific theories or statements should not involve
non-epistemic values. Again, the tension between my discussion and these
suggestions seems almost trivial: If the truth values of statements such as (a)–
(g) are dependent on ontological choices and if ontological choices are dependent
on non-epistemic values, then it is not possible to evaluate or to justify (a)–
(g) independently of non-epistemic values.
6 Avoiding Ontological Commitments?
The basic idea of my argument against VFI is that scientific statements presuppose
ontological choices that are often shaped by non-epistemic values. One possible
response to this challenge is to suggest that scientists can somehow avoid
ontological commitments and restrict themselves to a discussion of empirical
evidence that remains independent of any non-epistemic concerns. For example,
consider the case of tigers in the San Diego Zoo. I have argued that there are two
species of tigers according to the phylogenetic species concept but only one
according to the biological species concept. One may argue that a scientist could
limit herself to making this statement without having to endorse either of these
accounts.
There is certainly some truth to this objection. Scientists do not always have to
choose sides in ontological conflicts and can often formulate their results in terms of
conditionals that leave an ontological issue open. However, this suggestion will not
work if it is supposed to exclude all ontological commitments and I will argue in
this section that scientists often have to make ontological choices in order to
conduct empirical research.
To illustrate why general avoidance of ontological commitments is not viable in
scientific practice, let us consider one last case study. Research on domestic
violence has considerably broadened its scope in the past 20 years through
intersectional approaches that do not only consider gender but rather entangled
social structures such as class, gender, race, and sexual orientation. This
reorientation reflects the now widely shared assumption that earlier research on
domestic violence was biased in not investigating a ‘‘universal family reality but the
conditions of white middle-class heterosexual families’’ (Bograd 2005, p. 27).
As Kourany (2010) points out in Philosophy of Science After Feminism,
intersectional studies on domestic violence provide rich examples for the
123
Ontological Choices and the Value-Free Ideal
importance of values in scientific research. Kourany focuses on Carolyn West’s
research on the experiences of domestic violence of African American women in the
United States. On the one hand, this project requires careful attention to structures
that shape domestic violence in this specific context and have been ignored by
accounts that focus on white middle-class households. On the other hand, West’s
project is also committed to analyzing unique social positions such as ‘‘African
American couples who are young, undereducated, impoverished, unemployed,
urban dwellers’’ (West 2004, 1487) without perpetuating the stereotype ‘‘that black
Americans are biologically or culturally more prone to violence than other ethnic
groups’’ (West 2004, p. 1489).
West’s project provides an excellent case study for discussions of non-epistemic
values in scientific practice. First, considering the experiences of African American
women clearly shapes explanatory interests in the sense that it raises questions about
social structures that were not considered in earlier research. Second, it also
provides clear examples of direct influences of social considerations beyond
explanatory interests such as the aim to avoid racial stereotypes. These considerations affect not only the collection of data but also answers to the question what
counts as domestic violence in the first place. For example, Taft et al. (2009, p. 50)
argue that the limited focus of earlier research has led to ‘‘conceptualizations [that]
are predominantly grounded in the experience and worldview of Caucasian
women.’’
Insofar as the boundaries of domestic violence are in question, there are many
variables to consider. Some of them reflect general issues with boundaries of
violence such as the questions whether violence should be considered illegitimate
and/or physical by definition (Dempsey 2005). Other variables are more specific to
the domestic context and have been questioned in the light of intersectional
approaches. For example, accounts that presuppose gender asymmetry undermine
research on domestic violence in same-sex couples. In the case of race, West (2008)
stresses the need for an account that includes phenomena that have been masked by
racial stereotyping as ‘‘normal’’ experiences of black women.
Returning to the question whether scientists can avoid ontological commitments,
it is certainly plausible that researchers do not always have to be committed to
exactly one account of domestic violence. In fact, a consideration of different
accounts of domestic violence may be beneficial and could elucidate relevant
patterns of how rates of domestic violence differ depending on the chosen account.
However, even a consideration of different accounts of domestic violence will
require the choice of a relevant subset. There are far too many variables that can
also be weighed differently to consider all logically possible accounts of domestic
violence. First, one may therefore argue that an avoidance of choices is not even
logically possible. There are not only countless logically possible ways of defining
‘‘domestic violence’’ but every attempt to provide a complete list would arguably
also lead to a regress problem. Definitions of ‘‘domestic violence’’ will refer to other
entities such as intimate partners or physical abuse that can also be interpreted in
different ways. And if we would also try to avoid one specific account of intimate
partners or physical abuse, we would end up in a definitional regress by again
having to refer to new entities.
123
D. Ludwig
Logical problems aside, it should be uncontroversial that avoiding specifications
of ‘‘domestic violence’’ is not viable in scientific practice. What matters for
researchers is not the logical space of all possible accounts of domestic violence but
rather the choice of an (epistemically and socially) adequate account. For example,
researchers who want to study the prevalence of domestic violence in a specific
context will know what to look for only after having specified what counts as
domestic violence. One of the charges against earlier accounts of domestic violence
is precisely that they led to the collection of data that ignored the experiences of
African American women and other marginalized groups. In contrast, a broader
notion of domestic violence that reflects intersectional structures implies different
criteria and therefore the need for different data. One can therefore not even design a
meaningful study of domestic violence without making at least some choices
regarding its boundaries.
The case of domestic violence provides a plausible illustration of the necessity of
ontological choices in scientific practice and also connects the arguments of this
article with research in feminist philosophy of science on diverse case studies from
the social sciences such as divorce (Anderson 2004), employment rates (Anderson
1995), inflation (Dupré 2007), and rape (Dupré 2007). However, the same
arguments can also be developed with examples from the natural instead of the
social sciences. For example, attempts to avoid ontological choices in the case of
species lead to analogous problems. First, one can reject the very idea of a list of all
logically possible accounts of species. Second, avoiding any specifications of
species would clearly not be viable in scientific practice as biologists need to specify
criteria in order to know what data to collect. No matter whether we look at the
natural or the social sciences, the general lesson of this section therefore remains the
same: In order to get to the empirical issues, scientists need to specify the
boundaries of entities and ontological choices are therefore ubiquities in scientific
practice.
7 Conclusion
In this article, I have argued that the truth values of scientific statements depend on
ontological choices (Sects. 2, 3) and that ontological choices often depend on nonepistemic values (Sect. 4). Furthermore, I have argued that this value-dependency of
scientific ontologies undermines VFI (Sect. 5) as it is neither desirable nor possible
to avoid ontological commitments (Sect. 6).
While I have formulated a general objection against VFI, the article also suggests
a more positive picture of the roles of non-epistemic values. Focusing on these
positive elements can be motivated by the case for ‘‘a new direction for science and
values’’ (Hicks 2014) that moves beyond general controversies about VFI to more
nuanced discussions of the appropriate roles of non-epistemic values (cf. Anderson
2004; Longino 2004 for related points). First, the ideal of a value-free science has
become widely rejected by philosophers of science and one may worry that general
objections against VFI are therefore only going to confirm what many philosophers
already agree on. Second, a general rejection of VFI may be helpful in reminding us
123
Ontological Choices and the Value-Free Ideal
of the ‘‘socially engaged’’ dimension of philosophy of science (Cartieri and
Potochnik 2014) but it does not provide a positive characterization of the roles that
non-epistemic values should play in scientific practice. Given that criticism of VFI
should not lead to acceptance of every social value in every context of scientific
practice (cf. Wilholt 2009; Kitcher 2011; Biddle and Leuschner 2015), it becomes of
crucial importance to specify legitimate roles of non-epistemic values beyond the
generic statement that they cannot be completely excluded.
Reflecting the need for a more nuanced picture, one can also read this article as
an attempt to specify an area of scientific practice in which non-epistemic values
play a productive and legitimate role in the formulation of scientific theories. Given
that recent debates about science and values tend to focus on issues of uncertainty
and induction (e.g. Winsberg 2012; Biddle 2013; Morrison 2014; Intemann 2015;
Leuschner 2015 cf. Brown 2013), a discussion of ontological choices broadens the
perspective of debates about social values in scientific practice. The case studies of
this article illustrate positive roles of social values in scientific ontologies from
species in conservation biology and race in biomedical research to intelligence in
psychology and domestic violence in the social sciences. While these case studies
lead to a positive perspective on the roles of non-epistemic values in scientific
ontologies, they clearly do not provide a fully developed account. Such an account
would not only require examples of non-epistemic values in scientific ontologies but
also a discussion of how we should integrate them with epistemic considerations
and the empirical evidence. Addressing the structure of such an integration requires
that we move beyond general arguments about VFI and engage with ontological
choices in their unique disciplinary, historical, and social settings.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original
author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were
made.
References
Anderson, E. (1995). Knowledge, human interests, and objectivity in feminist epistemology. Philosophical Topics, 23(2), 27–58.
Anderson, E. (2004). Uses of value judgments in science: A general argument, with lessons from a case
study of feminist research on divorce. Hypatia, 19(1), 1–24.
Betz, G. (2013). In defence of the value free ideal. European Journal for Philosophy of Science, 3(2),
207–220.
Biddle, J. (2013). State of the field: Transient underdetermination. Studies in the History and Philosophy
of Science Part A, 44, 124–133.
Biddle, J., & Leuschner, A. (2015). Climate Skepticism and the manufacture of doubt. European Journal
for Philosophy of Science, 5(3), 261–278.
Bograd, M. (2005). Strengthening domestic violence theories. In N. Sokoloff & C. Pratt (Eds.), Domestic
violence at the margins (pp. 15–38). New Brunswick: Rutgers University Press.
Brigandt, I. (2009). Natural kinds in evolution and systematics. Acta Biotheoretica, 57(1–2), 77–97.
Brown, M. (2013). Values in science beyond underdetermination and inductive risk. Philosophy of
Science, 80(5), 829–839.
Büter, A. (2015). The irreducibility of value-freedom to theory assessment. Studies in History and
Philosophy of Science Part A, 49, 18–26.
123
D. Ludwig
Callicott, J. B., Crowder, L. B., & Mumford, K. (1999). Current normative concepts in conservation.
Conservation Biology, 13, 22–35.
Cartieri, F., & Potochnik, A. (2014). Toward philosophy of science’s social engagement. Erkenntnis,
79(5), 901–916.
Clark, A., & Chalmers, D. (1998). The extended mind. Analysis, 58(1), 7–19.
Condit, C., et al. (2004). Exploration of the impact of messages about genes and race on lay attitudes.
Clinical Genetics, 66(5), 402–408.
Cracraft, J. (1983). Species concepts and speciation analysis. Current Ornithology, 1, 159–187.
Craver, C. F. (2009). Mechanisms and natural kinds. Philosophical Psychology, 22(5), 575–594.
Danks, D. (2015). Goal-dependence in (Scientific) Ontology. Synthese, 1–16.
Dempsey, M. (2005). What counts as domestic violence? William and Mary Journal of Women and the
Law, 12, 301–333.
Donovan, B. M. (2014). Playing with fire? The impact of the hidden curriculum in school genetics on
essentialist conceptions of race. Journal of Research in Science Teaching, 51(4), 462–496.
Donovan, B. M. (2015). Putting humanity back into the teaching of human biology. Studies in History
and Philosophy of Biological and Biomedical Sciences, 52, 65–75.
Douglas, H. (2009). Science, policy, and the value-free ideal. Pittsburgh: University of Pittsburgh Press.
Dupré, J. (2007). Fact and value. In H. Kincaid, J. Dupré, & A. Wylie (Eds.), Value-free science: Ideals
and illusions? (pp. 27–41). Oxford: Oxford University Press.
Elliott, K. C., & McKaughan, D. J. (2014). Nonepistemic values and the multiple goals of science.
Philosophy of Science, 81(1), 1–21.
Gannett, L. (2010). Questions asked and unasked: how by worrying less about the ‘really real’
philosophers of science might better contribute to debates about genetics and race. Synthese, 177(3),
363–385.
Gray, W. D., & Fu, W. T. (2004). Soft constraints in interactive behavior. Cognitive Science, 28,
359–382.
Haber, H., & Odenbaugh, J. (2009). The edges and boundaries of biological objects. Biological Theory,
4(3), 219–222.
Haslanger, S. (2012). Resisting reality: Social construction and social critique. Oxford: Oxford
University Press.
Hicks, D. J. (2014). A new direction for science and values. Synthese, 191(14), 3271–3295.
Hochman, A. (2013). Against the new racial naturalism. The Journal of Philosophy, 110(6), 331–351.
Hudson, R. (2015). Why we should not reject the value-free ideal in science. Perspectives on Science
(online first).
Intemann, K. (2001). Science and values: Are moral judgments always irrelevant to the justification of
scientific claims? Philosophy of Science, 68(3), S506–S518.
Intemann, K. (2015). Distinguishing between legitimate and illegitimate values in climate modeling.
European Journal for Philosophy of Science, 5, 1–16.
John, S. (2015). Inductive risk and the contexts of communication. Synthese, 192(1), 79–96.
Kaplan, J. M. (2010). When socially determined categories make biological realities. The Monist, 93(2),
281–297.
Kaufman, J. S., & Cooper, R. S. (2010). Use of racial and ethnic identity in medical evaluations and
treatments. In I. Whitmarsh & D. S. Jones (Eds.), What’s the use of race (pp. 187–206). Cambridge:
MIT Press.
Khalidi, M. A. (2013). Natural categories and human kinds. Cambridge: Cambridge University Press.
Kitcher, P. (1984). Species. Philosophy of Science, 51, 308–333.
Kitcher, P. (2011). Science in a democratic society. New York: Prometheus.
Kourany, J. (2010). Philosophy of science after feminism. Oxford: Oxford University Press.
Lacey, H. (2010). The constitutive values of science. Principia: An International Journal of
Epistemology, 1(1), 3–40.
LaPorte, J. (2009). Natural kinds and conceptual change. Cambridge: Cambridge University Press.
Latour, B. (1999). Politiques de la nature. Paris: La Découverte.
Leonelli, S. (2012). Classificatory theory in data-intensive science. International Studies in the
Philosophy of Science, 26(1), 47–65.
Leuschner, A. (2015). Uncertainties, plurality, and robustness in climate research and modeling: On the
reliability of climate prognoses. Journal for General Philosophy of Science, 46(2), 367–381.
Lewontin, R. (1972). The apportionment of human diversity. Evolutionary Biology, 6, 381–398.
123
Ontological Choices and the Value-Free Ideal
Longino, H. E. (1996). Cognitive and non-cognitive values in science. In J. Nelson (Ed.), Feminism,
science, and the philosophy of science (pp. 39–58). Dordrecht: Springer.
Longino, H. E. (2002). The fate of knowledge. Princeton: Princeton University Press.
Longino, H. E. (2004). How values can be good for science. In P. Machamer & G. Wolters (Eds.),
Science, values, and objectivity (pp. 127–142). Pittsburgh: University of Pittsburgh Press.
Lorusso, L., & Bacchini, F. (2015). A reconsideration of the role of self-identified races in epidemiology
and biomedical research. Studies in History and Philosophy of Biological and Biomedical Sciences,
52, 56–64.
Ludwig, D. (2014a). Hysteria, race, and phlogiston. A model of ontological elimination in the human
sciences. Studies in History and Philosophy of Biological and Biomedical Sciences, 45, 68–77.
Ludwig, D. (2014b). Disagreement in scientific ontologies. Journal for General Philosophy of Science,
45(1), 119–131.
Ludwig, D. (2015a). Against the new metaphysics of race. Philosophy of Science, 82(2), 244–265.
Ludwig, D. (2015b). Indigenous and scientific kinds. The British Journal for the Philosophy of Science
(ahead of print).
Ludwig, D. (2015c). A Pluralist theory of the mind. Dordrecht: Springer.
Ludwig, D. (2015d). Extended cognition and the explosion of knowledge. Philosophical Psychology,
28(3), 355–368.
Machamer, P., & Douglas, H. (1999). Cognitive and social values. Science and Education, 8(1), 45–54.
Macleod, M. (2010). The epistemology-only approach to natural kind. In F. Stadler (Ed.), The present
situation in the philosophy of science (pp. 189–194). Dordrecht: Springer.
Maglo, K. (2011). The case against biological realism about race. Perspectives on Science, 19(4),
361–390.
Mallon, R. (2006). ‘Race’: Normative, not metaphysical or semantic. Ethics, 116, 525–551.
Mayr, E. (1969). Principles of systematic zoology. New York: McGraw-Hill.
McMullin, E. (1982). Values in Science. PSA: Proceedings of the biennial meeting of the philosophy of
science association, pp. 3–28.
Miller, G. (1956). The magical number seven, Plus or Minus Two. Psychological Review, 63(2), 81–97.
Mills, C. (1998). Blackness visible. Ithaca: Cornell University Press.
Mncube, Z. (2015). Are human races cladistic subspecies? South African Journal of Philosophy, 34(2),
163–174.
Mol, A. (1999). Ontological politics. A word and some questions. The Sociological Review, 47(S1),
74–89.
Morning, A. (2011). The nature of race. Berkeley: University of California Press.
Morrison, M. (2014). Values and uncertainty in simulation models. Erkenntnis, 79(5), 939–959.
Phelan, J. C., Link, B. G., & Feldman, N. M. (2013). The genomic revolution and beliefs about essential
racial differences. American Sociological Review, 78(2), 167–191.
Pöyhönen, S. (2014). Explanatory power of extended cognition. Philosophical Psychology, 27(5),
735–759.
Putnam, H. (2002). The collapse of the fact/value dichotomy: And other essays. Cambridge: Harvard
University Press.
Risch, N., et al. (2002). Categorization of humans in biomedical research. Genome Biology, 3(7), 1–12.
Rooney, P. (1992). On values in science. PSA: Proceedings of the biennial meeting of the philosophy of
science association, pp. 13–22.
Root, M. (2003). The use of race in medicine as a proxy for genetic differences. Philosophy of Science,
70(5), 1173–1183.
Ruphy, S. (2006). Empiricism all the way down. Perspectives on Science, 14(2), 189–214.
Ruphy, S. (2013). Pluralisme scientifique, enjeux épistémiques et métaphysiques. Paris: Hermann.
Scharpf, C. (2000). Politics, science, and the fate of the Alabama sturgeon. American Currents, 26(3),
6–14.
Schurz, Gerhard. (2013). Philosophy of science: A unified approach. New York: Routledge.
Sider, T. (2011). Writing the book of the world. Oxford: Oxford University Press.
Slater, M. (2015). Natural kindness. The British Journal for the Philosophy of Science, 66(2), 375–411.
Sober, E. (2007). Evidence and value-freedom. In H. Kincaid, J. Dupré, & A. Wylie (Eds.), Value-free
science (pp. 109–119). Oxford: Oxford University Press.
Stanford, P. K. (1995). For pluralism and against realism about species. Philosophy of Science, 62(1),
70–91.
123
D. Ludwig
Steel, D. (2010). Epistemic values and the argument from inductive risk. Philosophy of Science, 77(1),
14–34.
Taft, C., et al. (2009). Intimate partner violence against African American women. Aggression and
Violent Behavior, 14(1), 50–58.
Taylor, P. C. (2013). Race: A philosophical introduction. Malden: Polity.
Wallace, K. J. (2007). Classification of ecosystem services. Biological Conservation, 139(3), 235–246.
Weber, M. (1917). Der Sinn der ‘Wertfreiheit’ der soziologischen und ökonomischen Wissenschaften. In
M. Weber (Ed.), Gesammelte Aufsätze Zur Wissenschaftslehre 7 (pp. 489–540). Tübingen: Mohr.
West, C. M. (2004). Black women and intimate partner violence. Journal of Interpersonal Violence,
19(12), 1487–1493.
West, C. M. (2008). Mammy, Jezebel, Sapphire, and their homegirls. In J. Chrisler, C. Golden, & P.
Rozee (Eds.), Lectures on the psychology of women (pp. 286–299). New York: McGraw Hill.
Wilholt, T. (2009). Bias and values in scientific research. Studies in History and Philosophy of Science
Part A, 40(1), 92–101.
Wilson, R. A., Barker, M. J., & Brigandt, I. (2007). When traditional essentialism fails. Philosophical
Topics, 35, 189–215.
Winsberg, E. (2012). Values and uncertainties in the predictions of global climate models. Kennedy
Institute of Ethics Journal, 22(2), 111–137.
Winther, R. G., & Kaplan, J. M. (2013). Ontologies and politics of biogenomic ‘Race’. Theoria, 60(136),
54–80.
Zachar, P. (2015). Grief, depression, and the DSM-5: a review and reflections upon the debate. Revista
Latinoamericana de Psicopatologia Fundamental, 18(3), 540–550.
Zachos, F., et al. (2013). Species inflation and taxonomic artefacts. Mammalian Biology-Zeitschrift für
Säugetierkunde, 78(1), 1–6.
123