The
word ‘real’ is derived from the Latin res, which means things both in
the concrete and abstract senses. Thus, ‘reality’ refers to the totality of all
real things, and ‘realism’ is a philosophical doctrine about the reality of
some of its aspects. But, as philosophy is divided into several subdisciplines,
the doctrines of realism are likewise divided into a number of varieties. For
our purposes, without attempting to be exhaustive, it is appropriate to divide
philosophy—and thereby the problems of realism—into six areas: ontology,
semantics, epistemology, axiology, methodology, and ethics.
Ontology
studies the nature of reality, especially problems concerning existence.
Semantics is interested in the relation between language and reality.
Epistemology investigates the possibility, sources, nature, and scope of human
knowledge. The question of the aims of enquiry is one of the subjects of
axiology. Methodology studies the best, or most effective, means of attaining
knowledge. Finally, ethics is concerned with the standards of evaluating human
actions and alternative possible states of the world.
Given
these brief characterizations, it now seems easy to distinguish six different
problems of realism:
|
It
may also appear that these questions can be answered independently of each
other: each of them has typical positive and negative answers whose supporters
can be identified as ‘realists’ of the appropriate type, and their opponents as
‘anti-realists’.
The
situation is more complicated, however. The relationships between these six
disciplines are a fundamental point of departure that divides philosophical
schools. Such disagreement about the philosophical method is also an important
source in the debates between realism and anti-realism.
Plato's
theory of forms was a bold attempt to solve together the problems of ontology,
semantics, epistemology, axiology, and ethics. The traditional view, formulated
by Aristotle, takes ontology to be the ‘first philosophy’ and, hence, primary
to epistemology.1 However, many philosophers have followed
Immanuel Kant in rejecting such an approach as ‘metaphysical’: the first task
of philosophy is to study the possibility and conditions of knowledge by
uncovering the innate structures of the human mind (see Section 4.3). Kant's
followers have changed the mental structures to languages and conceptual
frameworks. Analytic philosophy in the twentieth century has studied questions
of existence through the ‘ontological commitments’ of conceptual systems and
theories (Quine 1969). The pragmatist tradition has developed variations of
Charles Peirce's proposal to ‘define’ reality and truth by the ultimate
consensus of the scientific community. These approaches place epistemology and
methodology before semantics, and semantics before ontology.
Similarly,
many pragmatists have denied the fact–value distinction (cf. Putnam 1992),
whereas the independence of OR and VR is often defended by ‘Hume's guillotine’
(i.e. is does not logically imply ought).
Michael
Devitt's Realism and Truth (1991) starts with a clear formulation of
five ‘maxims’. They imply that the ontological issue of realism (OR) should be
settled before any epistemic (ER) or semantic (SR) issue, where the latter two
should be sharply distinguished. If these maxims are taken to be premisses in
arguments for realism, pragmatists may complain that the question has been
begged against them—they are not ‘confusing’ ontology, semantics, and
epistemology, but rather ‘blurring’ such distinctions on purpose. Moreover,
other kinds of scientific realists, like Raimo Tuomela (1985), who accept the scientia
mensura principle (‘science is the measure of all things’), would claim
that all questions about existence are a posteriori, and therefore they will be
decided last, only after science has reached its completion.
Devitt's
main thesis is that truth should be separated from ontological realism (Devitt 1991:
p. x): ‘No doctrine of truth is constitutive of realism: there is no entailment
from the one doctrine to the other’ (ibid. 5). This thesis is correct at least
in the sense that an ontological realist may accept an anti-realist notion of
truth, and semantical or representational realism alone does not tell which
particular statements are true and which are false (ibid. 42). On the other
hand, Devitt himself defines the realist correspondence theory of truth so that
it presupposes (i.e. entails) the existence of a mind-independent reality
(ibid. 29). Similarly, many formulations of epistemological realism would not
make sense without some minimal assumptions of ontological realism.
Michael
Dummett (1982) argues that we should approach OR via SR and ER. Let S be a
‘disputed’ class of statements (e.g. about unobservable entities or about the
past). Then ‘realism’ about S is the thesis that the statements in S satisfy
the principle of bivalence, i.e. they are objectively true or false in
virtue of a mind-independent reality. Anti-realists about S instead reject
bivalence (cf. Luntley 1988), and suggest a theory where truth and meaning
depend on our actual epistemic opportunities and capabilities.
It
is fair to expect that the philosophy of language will help to clarify the
issues of realism and anti-realism. Ontological questions are certainly
intertwined with semantical and epistemological ones. But it is very doubtful
whether the study of language and meaning as such could resolve any
metaphysical issues. For example, there may be statements h such that h does
not have a determinate truth value. Is this an argument against realism? I
think the answer is no, since h may simply be a vague statement (e.g. ‘It is
raining now’), while the reality which h speaks about (e.g. weather) is
completely mind-independent. In other words, ontological realism need not
entail the principle of bivalence.
To
keep the various problems of realism conceptually separate—as far as we can.
But this does not imply that we could—or should—‘settle’ the ontological issue
‘before any epistemic or semantic issue’, or that we should eventually avoid
‘hybrid’ doctrines (Devitt 1991: 47). Realism is a philosophical world view, a
‘large-scale philosophical package’ in Richard Boyd's (1990) sense, and its
successful defence requires that we try to find the most plausible combinations
of ontological, semantical, epistemological, axiological, methodological, and
ethical positions. In other words, it does not seem promising to expect that OR
could be convincingly solved without considering SR, ER, AR, MR, and VR—and the
interconnections between these theses—as well.
