Thomasina Coverley on the subject of AI

Thomasina Coverley on the subject of AI

 I am presently reading the notes of one “A.A.L.”,
whom I have deemed to call Al for the sake of expediency and convenience,
as published in this past month’s edition of Scientific Memoirs.
The notes make reference to Babbage’s wonderful Analytical Engine,
and they ascribe to it a much grander scope than seems to have occurred
even to the namesake of the machine.
 Our mysterious Al proposes a departure from the
mundane mathematical drudgery for which the engine was designed.
They conclude, as had I, that the mechanisms by which
mathematical symbols can be combined and interpreted could be applied
just as soundly to, for instance, the symbols comprising music.
 They write, and I agree, that such a machine could surely formulate
the structure of a song, whose intricacy and extent were limited only by the
capacity and complexity of the machine itself, limitations which Al
identifies as being arbitrary, meaning there are no limitations at all if we
allow ourselves to speak theoretically, the only worthwhile way of speaking.
 As much as I admire the writings of Al, I must express that I feel
they stopped short of the true possibilities of the engine. As enamored as I
am with the idea of symphonies composed by the hand of machines,
I feel that the simple transition from mathematics to music is simply too
unambitious.
 If we are to take this machine, as Al posits, for a means by which
symbols of any kind can be encoded and operated upon, why limit oneself to
mathematical notation, or even to its close cousin musical score?
 As I read through these published and printed notes, how could I fail
to notice that the very language used to describe the engine takes the form of
symbols as well?
 Though the patterns of meaning encoded therein are of high
complexity, we have established already that this presents no barrier
to the purposes of speculation. Perhaps it would be possible to treat each
letter as a unique icon, and the combinations thereof as one might algebraic
expressions. Perhaps it would be more practical to treat the word as our
unitary symbol, increasing the number of unique icons but decreasing the
complexity of their interactions.
 The details of the implementation are in no way my concern;
my claim is simply that, following the description of the analytical engine as
presented by Al, it should be possible to tailor such a machine to parse
and operate upon the written word. By these means a machine could be
rendered capable of understanding and perhaps even responding to typeset
sentences and queries.
 Rather than being limited to purely scientific calculations,
or to the strictly artistic medium of musical arrangements,
such a machine would be free to run the gamut of expression just as any
human could, exploring all manner of written forms from poetry to scientific
conjecture.
 Of course the computational complexity that would be needed for
such a machine falls well beyond the realm of realism, but it is the nature of
advanced thinkers such as Al and myself not to be troubled by such things.