IEEE Technology and Society Magazine - March 2018 - 42

Abstract Entity

is-a

Physical Entity

is-a
Yeast Strain

is-a

Model of Yeast
Metabolism

is-a

Part-of
Part-of

Model of Robot
Architecture

Nutrient

Gene

Incubator

Denotes

Yeast Strain

is-a is-a

is-a
Nutrient

is-a

Biochemical
Entity

Manipulates

Gene

Robotic
Equipment

Incubator
is-a

is-a
Plate

Denotes

Plate

Manipulates
Manipulates

Manipulates

Computer
Computer
Software

Computer
Hardware

Figure 2. The Robot Scientist Universe (a fragment).

associated with different ontologies: correspondence theories with realism, and pragmatism, verification, and coherence with idealism, anti-realism, or relativism [14].
A robot scientist's physical effectors (laboratory
robots) can test the truth or falsehood of an abstract scientific proposition by specific physical experiments: an
Abstract entity of type Proposition is assigned a truth
value by a Physical entity that participates in a specified Process. This is achieved through the designed isomorphism between an abstract Denotation rule and a
physical Denotation process (Figure 2). This operational approach to truth does not discriminate between correspondence, pragmatism, verification, or coherence
theories of truth. For a human scientist these different
approaches may possibly inspire different ways of doing
science, but given the current state of development of
robot scientists it is unclear to us whether there is any
operational difference between these approaches.
One of the most debated questions in the philosophy
of science is that between realism and anti-realism.
Realism is "the viewpoint that accords to the objects of
knowledge an existence that is independent of whether
anyone is perceiving or thinking about them" [14]. The
alternative position regards the existence of the real world

as a metaphysical question that cannot be answered, and
regards scientific theories as instruments of prediction
[15], [16]. As physical devices, robot scientists necessarily
adopt a realist position as defined above. However, their
approach to determining the truth of propositions is
also consistent with that of anti-realism. Therefore, with
robot scientists there would seem to be no difference
in regarding scientific theories as descriptions of reality or as tools for prediction. This approach is related to
quietism [17].
The realism/anti-realism debate is closely connected
to another area of interest in the philosophy of science
that is important in the design of robot scientists: the
relationship between observed and theoretical entities.
This subject has long been a matter of debate in the philosophy of science, with some philosophers claiming
that the distinction is not real and/or important [16]. We
argue that the distinction clarifies the robot scientist's
reasoning, and that what are observed and theoretical
entities is relative to defined instrumentation.
A common view in the philosophy of science is that
hypothesis formation necessarily requires human creativity [10]. This view has long been challenged by AI,
[e.g., 19]. Most work within scientific discovery has

IEEE Technology and Society Magazine

∕

MARCH 2018

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - March 2018