What I think Corpus Collosum is getting at - or what makes more sense - is this:

The pattern of organisation of an intelligent system is self-similar across a range of scales.

Admittedly, this is less concise than 'fractal' - but possibly more precise. The pattern of organisation of living systems is autopoetic - self making - and some claim that cognition is living. That is, living systems are cognitive sytems and vice versa.

Instead of throwing buzzwords around, let me make this more specific - living systems are networks within networks and so are intelligent systems. Inasmuch as networks are similar to other networks, life is fractal (in this limited way). The higher levels of organisation mentioned above are also heirarchical networks (see process physics). However, are they 'similar' to each other in any rigorous way? Probably not; after all, at different levels of organisation, different behaviours are present in the 'nodes' (humans, neurons, countries). This would tend to suggest that they are organised into different types of network.

It's mildly amusing to me that people have voted this down. Presumably they think that I am suffering from Modern Physics Abuse Syndrome. Ironically, as a biologist, I study much more complex physics than any (or most) of you high-and-mighty physics students out there. Its not for nothing that physicists are trying to bring the rigour that worked so sucessfully for (relatively) simple systems -planets, inclined planes, charges in a vacuum or whatever -to the complex and complicated sytems of biology. In case any of you physicists are interested in the beginnings of an explanation for this stuff, see:

Philip Ball The Self-Made Tapestry
Peter Coveny and Rodger Highfield Frontiers of Complexity

THEN you can downvote me! :)