Stevan Harnad - Computation is symbol manipulation

It extends the Church-Turing Thesis, which asserts that any effectively calculable function can be computed by a Turing machine. A system where actions are determined by the current state, possible transitions, and symbols read. Whatever physical systems can do, computers can do.

Manipulating symbol tokens based solely on their shapes.

Proposes cognition as a form of computation.

The symbols and operations within a computational system have meaningful and contextually appropriate interpretations that correspond to real-world entities and concepts. Important, because it requires understanding of semantics and not just syntax for a machine to replicate cognition.

Nontrivial systems possess inherent structure and resist arbitrary changes, unlike trivial systems.

Allows replacing symbols without altering computation, abstracting it from physical details.

States that if someone or something behaves indistinguishably from others, it’s rational to attribute similar capabilities.

Involves connecting symbols directly to what they represent without external interpretation

Weak CTTP suggests physical systems are equivalent to Turing Machines, but simulations don’t equate to genuine understanding.

Strong CTTP overlooks implementation independence, crucial for distinguishing computation from other natural phenomena.