luciano.floridi@philosophy.ox.ac.

uk
www.philosophyofinformation.net

Luciano Floridi
Research Chair in Philosophy of
Information, UNESCO Chair in
Information and Computer Ethics
Research Group in Philosophy of
Information, University of
Hertfordshire
Information Ethics Group, OUCL &
Philosophy, University of Oxford

The Beginning: Plato

The Makers Knowledge Tradition
Hobbes on Constructionism
Constructionism: Four Consequences
Conceptual Engineering and Computer Science
Six Constructionist Principles
The Turing Test Revisited
Turing Test and Constructionism
Conclusion: back to Plato

Who really knows an artefact?

(A) The imitator, who paints the lyre?
(B) The maker, who builds the lyre?
(C) The user, who plays the lyre?

Imitator

The user possesses better

knowledge than the maker.
Republic
601a-e;
Cratylus
390b;
Euthydemus 289b; Phaedrus 274e.

User

artefact

Maker

Socrates: Then the sort of knowledge we

require [..] is that in which there happens
to be a union of making and knowing
how to use the thing made. [...]
So we ought, it seems, to aim at
something far other than being lyremakers.
Plato, Euthydemus 289b

Problems
1.Friendly fire: meant against imitators, ends up being
against makers.
2.Separation user/maker and lower regard for techne.
3.How can God know better despite being the
maker/engineer?
4.How can conceptual/semantic artefacts be better known
by their users?

If genuine knowledge is knowledge of the intrinsic

nature of the object known (knowledge of the
ontology of the known), we as epistemic agents
can only know what we make.

Constructionism:
knowledge is acquired
through the construction
of semantic artefacts
(information modelling).

Epistemic research and

information modelling: two
sides of the same coin.
It is not just seeing but
handling that makes a
difference.

Hobbes (1656), Six Lessons to the Professors of Mathematics

Geometry
therefore
is
demonstrable,
[Because]
the
science
of every
subject
is
Of
arts some
are demonstrable,
others
for the lines
figures
from which
derived
fromanda and
precognition
of we
the
indemonstrable;
demonstrable
are
reason construction
are drawn and
and
described
by
causes,
generation,
of
those
of construction
the subject
ourselves;
civil
is
the
same;is and
consequently
the
whereof
inand
the
powerphilosophy
of where
the artist
demonstrable,
the
causes
are known,
thereweis make
placedoes
for
himself,
who,
in because
his demonstration,
commonwealth
ourselves.
demonstration,
but
notthe
where
the causes
no
more but deduce
consequences
are
tobecause
seekoperation.
for.of natural bodies we know
of
his
own
But
not the construction, but seek it from
the effects, there lies no demonstration
of what the causes be we seek for, but
only of what they may be.

Bacon and knowledge as power we can

improve our knowledge by improving the knowledge
of the techniques by which we investigate reality.
Vico and the verum ipsum factum or verum
et factum convertuntur (what is true and what is
made are interchangeable) comprehension of
the world will fail, so better focus on those
sciences whose subject is created by man.
Kant and the noumenon the ultimate
nature of reality in itself remains unknowable.
Friedrich Dessauer technology
establishes a positive contact with noumena.

White boxes: systems about

which we know everything,
because we built them.

Black boxes: systems of which

we ignore the internal structure,
rules and composition.

Things in themselves like black

boxes: we can never know their
intrinsic nature because we didnt
make them. Constructionism can learn
some lessons from Computer Science.

1. Poietic Knowledge: we can

know only what we make.
2. Constructability: working
hypotheses are investigated by
(theoretical or practical) simulations.
3. Controllability: the simulation
has to be controllable.
4. Confirmation: confirmations or refutations of the
hypothesis concern the simulation, not the simulated.
5. Non-descriptivism: Reality and Knowledge are in
a resource-to-product relation, not original-to-copy or
source-to-reproduction relation.
6. Economy: the fewer conceptual resources we use,
the better it is. (Ockham revisited).

Constructability

Controllability

Given a theory, we implement

and test it in a system.
The resulting system is
totally
controllable,
i.e.
modifiable, compositional,
teleological
and
predictable.

Neither machines nor programs are black boxes; they are

artefacts that have been designed, both hardware and
software, and we can open them up and look inside.
A. Newell and H. A. Simon (1976), Computer Science as
Empirical Enquiry: Symbols and Search.

Modifiable

we can change its internal structures

and rules at any time.

Compositional

we have control of any single part of the

system.

Teleological

Predictable

The system has been built with an intent

(makers goals) and it acts following an
end, though not necessarily intentional
(systems goals).
We know the rules of the system, so we
can know its behaviour and use it to
predict the behaviour of another system
or of the natural system that our
simulation is trying to model.

Confirmation

Prevents from generalizing the

working hypotheses, as if the
simulation were THE real cause (or
internal structure) of the simulated.

Non-descriptivism

Reality is a resource for our

knowledge, not a source of it.
Constructionism provides effective
methods to work with available
affordances and constraints.

Economy

Careful management of resources

Green policy. Constructionism seeks
the minimal ontological commitment.

Imitation Game aka Turing Test

According to Turing (1950):
in about fifty years' time it will
be possible to programme
computers [...] to make them
play the imitation game so well
that an average interrogator
will not have more than 70
percent chance of making the
right identification after five
minutes of questioning.

Turing Test:
respects the minimalist criterion
uses the levels of abstraction
is constructionist.
Minimalism. Turing refuses to provide an answer to the
question can a machine think?. Not well-defined problem
because of such vague concepts as machine and thinking.
Imitation Game = better management of resources.
Levels of Abstraction. Turing Test is a Level of Abstraction.
The rules of the game define the conditions of observability.
By changing the rules of the game one changes the LoA so
the answer will change too.

Turing refuses to define what

intelligence really is.

Makers Knowledge

He makes an hypothesis, and

devises a system to assess it.

Constructability

The system is fully controllable.

Controllability

The fact that a machine passes

or fails the test implies only
that the machine can, or can
not, be considered intelligent
at that Level of Abstraction.

Confirmation

Turing does not consider those

ways requiring a large amount
of conceptual resources.

Minimalism

Turing Test based on Questions and Answers.

Limit: agents not required to ask questions.
To know is to be able to wonder and build semantic
artefacts that can satisfactorily and successfully
address such wonders.

In Platos Cratylus (390c)

someone who knows is
defined as: the man who
knows how to ask and
answer questions.

luciano.floridi@philosophy.ox.ac.uk
www.philosophyofinformation.net
ACKNOWLEDGEMENTS
Many thanks to Jan van
Leeuwen, the NIAS-Lorentz
Center and everybody else who
made the meeting possible.
COPYRIGHT DISCLAIMER
Texts, marks, logos, names, graphics, images,
photographs, illustrations, artwork, audio clips, video
clips, and software copyrighted by their respective
owners are used on these slides for non-commercial,
educational and personal purposes only. Use of any
copyrighted material is not authorized without the
written consent of the copyright holder. Every effort
has been made to respect the copyrights of other
parties. If you believe that your copyright has been
misused, please direct your correspondence to:
l.floridi@herts.ac.uk stating your position and I shall
endeavour to correct any misuse as early as possible.

Luciano Floridi
Research Chair in Philosophy of
Information, UNESCO Chair in
Information and Computer Ethics
Research Group in Philosophy of
Information, University of
Hertfordshire
Information Ethics Group, OUCL &
Philosophy, University of Oxford