Naturoids: On the Nature of the Artificial
by Massimo Negrotti.
World Scientific Publishing, River Edge, NJ, U.S.A., 2002.
150 pp., ISBN: 981-02-4932-2.
Reviewed by Rob Harle
recluse@lis.net.au
Much has been written about the "artificial". Books
and journal articles on artificial intelligence, Alife and artificial
body parts are numerous and quite often extensively detailed. What
many of them lack however, is a coherent, well articulated theory
of just what it is that constitutes "artificiality". This
slim volume, Naturoids: On the Nature of the Artificial goes
a long way towards correcting this rather odd situation.
Negrotti has succeeded in outlining a basic theory of artificiality
which he methodically and systematically expands throughout the book.
Researchers, especially in the field of artificial intelligence, often
bemoan the intractability of the practical problems associated with
this work. Generally, artificial intelligence progress has been disappointing
in relation to its original expectations. This book shows why this
has occurred. The main reason has been an unrealistic, confused and
naive understanding of the fundamental attributes of what constitutes
artificiality, together with a serious underestimation of the complexity
of natural systems.
This book is arranged in two sections. Part One introduces and develops
the basic theory and Part Two, "The Reality of the Artificial"
discusses practical applications with numerous examples of the artificial.
Part Two serves as a good introduction for the student or lay reader
on a whole range of artificiality from medical prostheses to artificial
zoos. As such this section does not really cut any new ground, other
than applying Negrottis theory to the various practical applications
of the artificial.
Before outlining Negrottis theory I must say this book is graphically
uninspiring, has no index and has more than an acceptable level of
typographical and grammatical errors, I am not sure if this is due
to translation Negrotti is affiliated with the University of
Urbino, Italy or poor proof reading. However, this in no way
detracts from the importance this theory will play in the future of
the artificial, especially artificial intelligence.
Negrotti starts by clearly describing the difference between the technology
of the artificial and conventional technology. The "artificial",
aims to reproduce something existing in nature whereas conventional
technology creates something new, for example we talk of artificial
flowers but not an artificial telephone (p. 6).
He then outlines, "...the three main concepts of the theory of
the artificial the observation level, the exemplar and the
essential performance..." (p. 33). To this can be added the "principle
of inheritance", that is, qualities and quantities that cannot
be predicted a priori. For example the side effects of drugs.
Of these principles the concept of "the observation level"
is extremely important to understand in the process of any artificial
object or system design. Whilst this observation level concept is
similar to the quantum mechanics notion of the observer changing that
which is observed, it goes further in suggesting that only one level
of observation can be realised at a time. "The choice of an observation
level allows us to capture a true reality, but only the one which
is compatible with that level (p. 21). For example, if we wish to
create an artificial flower, at which level of the numerous possible
levels of observation do we want it to be like the natural flower?
If we choose appearance, rather than say fragrance, then it does not
matter what is hidden behind its form and colour, it could be hollow.
Negrotti argues that we cannot satisfy all possible levels of observation,
in fact, cannot know all levels and it is this problem which makes
designing complex artificial systems, such as a bionic person extremely
difficult, perhaps impossible.
The book is sprinkled with interesting and relevant historical accounts
of how humans have attempted to create the artificial, from poor old
Icarus wax wings to Kurzweilian style super robots ones
that make humans obsolete (p. 71). There are also quite a few references
to different art forms over the centuries which fall into Negrottis
fourfold classification system of the artificial.
I think this book is an important read for all professionals who are
involved in designing and creating artificial objects and systems,
at the very least it will save much wasted effort in attempting the
impossible.