Re: Babbage/Menabrea: Analytical Engine

From: Marinho Francis-Oladipo (
Date: Sat Feb 24 2001 - 13:26:31 GMT

> Was there a clear idea at the time of what was meant by "able to
> perform any function or calculation"?...

I think that there was in a sense, because there was distinct talk of
analytic calculations as well as basic operations. This was the main
difference between the Difference Engine and the Analytical Engine. I guess
the question is did their understanding of the phrase in question encompass
all that has been identified to be computation. If it didn't then I would
say that there was not a strictly clear idea.
I'm not quite clear on whether this is the intuitive or formal sense of the
Church/Turing thesis. The thesis talks about an "effective" method of
achieving results of a calculation. Turing and Church from what I
understand both arrived at the same solution from different angles with
Turing saying that the informal method can be obtained using the formal
method. Church appears to me to be saying that that without the formal
method, the informal method cannot be obtained via this "effective" method.
Have I missed the point?

> And was Babbage's Analytical Engine an approximation to a Universal Turing
> Machine?

The universal Turing Machine is one that can function without having to
have its functionality directly changed for a new type of calculation. The
algorithm just needs to be written to its tape. The Analytical Engine works
similarly as the operator needs to feed in the new algorithm to be used for
new calculations. I'm not sure whether the Analytical Engine is an
approximation but the following comments from Menabrea could help on the
way to a decision. I draw from them that the Analytical Engine probably is
an approximation.

> ...and similarly the cards merely command the engine to perform these same
> operations; but in order that the mechanisms may be able to act to any
> purpose...problem must in every particular case be introduced.

> What was the motivation for eliminating the need for human intervention...

Human intervention was needed to be eliminated as it would improve accuracy
of calculations and increase efficiency thus it was mainly for practical
reasons, I would say. With humans, the time taken by different people on
performing tasks would be a big hang-up if they were left directly involved
in the calculation process.

Analytic problems unlike numerical ones are those that do not necessarily
have a formula that can be directly applied with all known values. It
contains incomplete data. Computation is described as symbol manipulation
and so even though values are unknown, symbols can be substituded for the
symbols that would normally be the values and a solution should still be
obtainable without actually knowing what the symbols mean as long as a rule
is being followed.

> What does "reflection"... "executes concepts of intelligence" mean?

Reflection is what we feel we do when we do things like making choices or
decisions. There is a feeling of doing more than just rule following
because there is sometimes no way of telling why a particular choice was
made, it is just put down to an instinct or feeling.
I think that Conceptions of intelligence imply doing what intelligent
systems can do and perhaps also knowing how intelligent systems do them. A
concept is an idea, knowledge about something. I think a concept is more of
a visualisation than an actualisation.

Francis-Oladipo Marinho < >

This archive was generated by hypermail 2.1.4 : Tue Sep 24 2002 - 18:37:16 BST