Information Technology Reference
In-Depth Information
placed in a square. Obviously, the electronic version of a 'bit' has many practical
advantages; it is compact, fast and computationally viable. Other objects that fit are
space, time, colour (being coloured), force and mass (as per Newton).
Although Wittgenstein's Tractatus satisfied formal languages, such as predicate
calculus, he found that it did not completely explain natural language. The referential
paradigm of meaning cannot cope with the human use of a language. For example,
you can take the notion of a 'game' and explain this by pointing to examples of
games. But what defines a game? Is there any definition that will both contain the
idea of all games and exclude all that are not games? The answer seems to be no. Yet
this negative case cannot happen within the bounds of the Tractatus.
This indeterminacy of meaning in natural language seems to be a flaw in his initial
great work. Recognising this he then went on to explore these flaws in his second
great work 'the Philosophical Investigations' (Wittgenstein
1953
). Here he expands
the principles of meaning to include the flexible way language is used. This involves
what he called 'a language game' where the rules of the 'game' can be changed. The
challenge we make is
“can computer science make the same leap?”
Can we extend
the computer interactions to also include a programming game?
I propose that because of this essential flaw identified by Wittgenstein, computers
are unlikely to have the possibility of natural communication with people unless
we apply a different approach to program design. I will come to this conclusion by
considering the two major works on the philosophy of language by Wittgenstein.
I will show that such a lack of natural communication with machines is related to
same reason that Wittgenstein made the paradigm shift away from his first work—
the Tractatus. I will explain why the Tractatus clearly aligns with formal computer
modelling and that his second work, the Philosophical Investigations, best fits the
more flexible requirements of human communication. How his second work might
be used to make computer communication more human becomes another question
to be answered.
8.2
Inferring Internal Experience
As we have discussed, classical linguistic philosophy suggests that language un-
derstanding arrives from denotational (referential) semantics. If we examine what
people talk about, we find that many of the conversations are descriptions of our own
internal lives. Since nobody can have direct access to another's internal experiences,
then the only way in which such experiences can be understood is indirectly through
inference. We can infer each other's experience because we share the state of being a
person, in the same culture, using a common language and in the context of similar
external events (such as a musical performance; see Chap. 6 and Billinge and Addis
2003
). It is hence possible through conversation to build an internal model of another
person's view of the world. The only requirements for this model is to be able to
make predictions from conversations about:
