User blog comment:Pellucidar12/Peng-Pellucid-Armstrong Number/@comment-29915175-20170102011204/@comment-27513631-20170102234128

Not really; you'd have to define what you mean by 'n symbols'. 'n total applications of inference rules' is more valid, and is completely well defined if you use a given collection of inference rules as in e.g. https://www1.maths.leeds.ac.uk/~rathjen/ICMend.pdf p5-p6.