You are on page 1of 2

A new way of classifying ontological or metaphysical positions can be based on

Karl Poppers terminology of three worlds (see Popper, 1972, 1974). World 1
consists of physical things, events, and processes in space and time, including
lawlike relations between such entities. This is the domain of inorganic and
organic nature, studied by physics and biology. World 2 includes subjective
mental states and events (e.g., beliefs, emotions, and volitions) in individual
human minds. This is the domain of human psyche, studied by psychology and
cognitive science. World 3 contains the products of human social action, such as
cultural objects, social institutions, and abstract entities like concepts,
propositions, arguments, theories, problems, and numbers. This domain is
studied by cultural and social sciences, logic and mathematics. Three monistic
metaphysical doctrines can now be identied (cf. Niiniluoto, 1999). Materialism in
its radical eliminative form claims that everything real belongs to World 1.
Reductive materialism states that reality is reducible to World 1 entities and their
complexes. Eliminative and reductive materialism are forms Constructive
Realism in Mathematics | 343 of physicalism. Emergent materialism takes World
1 as primary, but admits that suciently complex material systems may have
emergent non-physical properties. Subjective idealism makes parallel claims
about World 2. Its eliminative and reductive forms constitute the doctrine of
spiritualism, but emergent idealism is also a possible view. Objective idealism in
its classical versions has taken some non-material and non-subjective ideas
(such as Platos forms, thoughts of supernatural gods, and Hegels objective
spirit) as the ultimate source of all being, but more mundane variations could
replace them by some abstract World 3 entities. Idealist views (e.g.,
phenomenalism, social constructivism) are ontologically antirealist, as they treat
the material reality in World 1 as mind-dependent or humanmade. Besides such
monistic views, dualist ontologies may accept World 1 and World 2 as two
independent (but possibly causally interacting) domains of reality. This is the
Cartesian tradition initiated by Descartes. Another kind of dualism could accept
World 1 and World 3. Trialist doctrines accept the reality of all three worlds. When
Popper introduced his theory of the third world in a paper with a teasing title On
the Theory of the Objective Mind, Mario Bunge was shocked that in 1967 Popper
had a sudden conversion to objective idealism (Bunge, 1981, p. 138). However,
Poppers position can be interpreted as a form of emergent materialism (see
Niiniluoto, 1999): in his evolutionary account he sees Worlds 2 and 3 as historical
products of World 1. Neither of them could exist without the material World 1,
but they have achieved a relatively independent status by being able to in-uence
material entities by a causal feedback mechanism. Popper mentions Bernard
Bolzanos sentences in themselves (Stze an sich) and Freges thoughts
(Gedanke) as anticipations of his World 3. Frege stated in 1918 that thoughts
are neither things in the external world nor ideas, so that a third realm has to
be recognized (Frege, 1984). So for Popper propositions as the objective
contents of thought are paradigmatic examples of World 3 entities. Freges
position about propositions is Platonist, however, and Popper wants to dissociate
his man-made and changing World 3 from Platos divine and unchanging third
world. World 3 entities emerge as by-products of human social action, especially
the creation of symbolic languages. Poppers theory of World 3 is thus a kind of
poor mans Platonism: as no higher being or god has created abstract objects
for us, man must himself make them (see Niiniluoto, 1992). Popper does not
refer to Carnaps Aufbau, which in 1928 gave a three-fold classication of objects
into physical, psychological, and cultural (see Carnap, 1967, cf. Niiniluoto, 2006).
Cultural objects include concrete artifacts (e.g., books, paintings, coins) which
have a material kernel in World 1, but also non-physical relational properties
(e.g., meaning, value, price) through their relations to human practices. Cultural
objects include also abstract artifacts (e.g., poems, sym- 344 | Ilkka Niiniluoto
phonies) which can be copied, recorded, and reproduced by their embodiments
that belong to Worlds 1 and 2. Using Carnaps terms, a symphony can be
documented in World 1 by notes on a score, acoustic waves, recording tapes, or
compact discs, and it can be manifested in World 2 as the composers idea or the
listeners experience. A virtue of the theory of World 3 is that it allows us to say
that there is one and only one Beethovens Eroica Symphony: it would be quite
arbitrary to identify this masterpiece with any of its multiple documentations and
interpretations in Worlds 1 and 2. We can also see that abstract cultural objects
in World 3 exist in time but not in space: they are created at a certain moment of
time, and they could be destroyed if all of their documentations and memories
disappeared from Worlds 1 and 2. Popper defends the reality of World 2 and 3
entities by their causal powers. For example, mental states inuence our bodily
behavior by downward causation, and coins function as money via existing
social institutions. It is less evident that purely abstract entities could have a
causal impact on us. For such entities one can apply Charles Peirces medieval
de-nition of reality: real things have characters that are independent of what
anybody may think them to be (see Peirce 1931- 35, 5.405). Poppers own
example is prime numbers: the largest prime number known in 2013 is
2578851611. Probably no one had ever thought about this number or written
down its 17,425,170 digits, but still it had (even before it was found by The Great
Internet Mersenne Prime Search on January 25, 2013) objectively the property
of being prime. In the same spirit, Haim Gaifman (2012) characterizes the realist
ontology of mathematics in terms of factually meaningful questions which have
determinate answers, independent of our knowing it or our abilities to -nd it.
Peirces criterion is not unproblematic in sophisticated branches of mathematics
like set theory. The de-nition of measurable cardinals was given Stanislaw Ulam
in 1930, and a great number of theorems about them have been proved. For
example, Ulam and Tarski showed that measurable cardinals are strongly
inaccessible. Further, if is measurable, then is the th weakly inaccessible
cardinal. Solovay proved in 1967 that ZF together with the axiom of determinacy
AD implies the measurability of the cardinal 1. On the other hand, it is known
that the existence of measurable cardinals is independent of the axioms of ZFC.
Scott proved in 1961 that the axiom of constructibility V=L implies the non-
existence of measurable cardinals, but it is still unknown whether their existence
is compatible with the axioms of ZFC. So in a sense, after 70 years of intensive
studies, we still dont know whether these entities exist or not (see Kanamori,
2003). One possible conclusion for a realist, compatible with Peirces criterion, is
that mathematicians are able to rigorously investigate alternative universes of
set theory Hamkins (2012) defends a Platonist version of this set-theoretic
multiverse. Constructive Realism in Mathematics | 345 Peirce himself used his
criterion to distinguish reality and ction, and by the same principle we can
distinguish real public entities in World 3 and merely ctional objects (e.g., Santa
Claus and Anna Karenina). Thus, World 3 di-ers from Meinongs jungle of all
possible conceivable objects (see Routley, 1980; cf. Niiniluoto, 2011)