Minds and Computers : An Introduction to the Philosophy of Artificial Intelligence

a fortiori
even more strongly, by the same reasoning. ‘Brunswick is
in Melbourne, therefore it is in Victoria; a fortiori, it is in Australia.’
activation function
one of the two functions implemented by
each processing node of an artificial neural network which deter-
mines, given a level of activation, whether or not the node
should fire.
adicity
the number of inputs which a function takes.
a
fferent incoming (signal conduction).
algorithm
another name for an e
ffective procedure – a systematic
method which requires no understanding for its execution and
which can be completed in finite time.
allophone
a phonetic variant of a phoneme whose context-
dependent production is determined by phonetic realisation rules.
amphiboly
a property of certain syntactic structures such that they
admit of multiple semantic interpretations. ‘I saw the man on the
hill with the telescope.’
anomalous monism
the view that certain physical states have irre-
ducibly non physical properties.
antecedent
the left-hand side of a conditional.
aphasia
a language deficit. Common forms include Broca’s aphasia,
Wernicke’s aphasia and conduction aphasia.
artificial intelligence, strong
the contentious view that it is possible
to develop artefacts which have minds in the sense that we have
minds.
artificial intelligence, weak
the uncontentious view that it is possible
to create artefacts which are able to implement certain functions
which are held to be (weakly) constitutive of intelligence. Often
used to sell white goods.
artificial neural network
a structure of weighted connections
between simple processing units – each of which implements a
transfer function and an activation function – the propagation of
211

---

activation through which constitutes parallel distributed processing.
Artificial neural networks are modelled on ‘brain style’ information
processing.