Humour and Neural Networks
p at dorrell.demon.co.uk
Sat Nov 11 03:36:18 EST 1995
I have formed a theory of Humour that explains it in terms of neural network
If a neural network meets an input that is different to those it has
previously been trained on, then its reponse will depend on some way to
the learned responses it has to inputs it has been trained on.
In practice this might be revealed as a sort of "reasoning by similarity",
and this describes the way that the human brain tends to think.
But in some cases reasoning by similarity might produce a completely wrong
answer. No doubt the neural network would eventually correct this after
further training. But if the response is completely wrong, there may be
some other part of the brain that can recognise this, and immediately
apply a correction, i.e. suppression of the incorrect response. I call
this the "ridiculousness detector".
More details in my web page -
I have been wondering if there exist individuals who suffer a neurological
deficiency that involves a lack of sense of humour. Further symptoms might
include a readiness to accept bizarre belief systems (religion ?). In fact the
patient may have a catalog of bizarre symptoms including lack of sense of
humour, and it may be the lack of humour that is actually the primary
cause of the condition.
Philip Dorrell Email: p at dorrell.demon.co.uk
Articles on consciousness, dreams, music, evolution, anthropic principle at
More information about the Neur-sci