Computer Science is a Natural Science
Posted on
by
Chris Warburton
Many people don’t think of Computer Science as a natural science, in the same sense as Physics, Chemistry, etc. I think this is flat out wrong, and don’t understand why people might think this; for example, my undergraduate degree was Physics with CS, and people often told me that’s a “strange combination”.
Some examples of CS being a natural science:
- Strong/physical Church-Turing thesis: treats the Church-Turing thesis (one of the founding ideas of CS) as a physical law, which places strict limits on the behaviour of any physical system, namely that it must be computable by a Turing machine. Some call this “unproven”, but no physical law is “proven”; relativity, gravitation, thermodynamics, etc. are all experimentally tested and theoretically sound; so is the physical Church-Turing thesis.
- Information theory is a core part of CS, and is also a fundamental part of nature, with deep connections to probability (the outcome of a 50/50 event is precisely 1 bit), thermodynamics (entropy; Maxwell’s Demon; etc.), causality (the speed limit in relativity only applies to information transfer), quantum mechanics (non-classical probability using complex numbers; entanglement == mutual information; observation == information transfer), black hole information paradox, holographic principle, etc.
- Algorithmic information theory is a very nice way to quantify and compare scientific theories, by treating them as programs which output observations.
- Computational complexity theory places further restrictions on the behaviour of physical systems. Similar to the physical Church-Turing thesis, we can regard “no system can perform NP-hard calculations using polynomial resources” as a theoretically sound, empirically tested law of physics.
- CS principles provide straightforward, rigorous, objective explanations to longstanding problems in science, like free will (computational irreducibility), wavefunction collapse (information transfer), mind/body duality (hardware/software), and so on. Unfortunately, many don’ see CS as a natural science, and hence don’t encounter its perspectives/explanations, and instead repeat the same old woo (quantum consciousness, multiverses, anthropic arguments, etc.)