I prefer to use the terms as they were originally defined by the authors who introduced them, and not with the modified meanings that become fashionable after the authors of some paper written some decades later decide randomly to change the definitions of the old terms. This is also true for the terms "garbage collector" and "reference counts", which were both introduced in 1960 to denote 2 clearly distinct methods of memory management, and the main difference between them does not consist in whether some kind of reference counts exist somewhere, or not.

The paper "A unified theory of garbage collection", which has started the fashion of considering reference counting as a kind of garbage collection, only shows correctly that both tracing and reference counting are complementary implementation techniques for a garbage collector.

It does not mention anywhere the essential practical difference between the traditional standalone memory management with reference counts and a garbage collector, which stays the same regardless whether the garbage collector also happens to use reference counts for some purposes, which is the difference between predictable and unpredictable times when memory reclamation is done.

Many of the modern authors of academic papers are a poor model of using computer terminology (or for the terminology in other domains), because very frequently it is obvious that they have not read the old works where such terms were introduced for the first time, even when such works are cited in the bibliography.

Unlike them, I have done an extensive research to find when and where various computing terms have been used for the first time, and I strive to use most terms with their original meanings, not with corrupted meanings, even in the cases when the latter have become more popular lately.

So as fellow digital archaeologist I would find interesting to find Lisp, BASIC, CLU, SIMULA, or Cedar papers, where the authors mention reference counting not being a form of garbage collection algorithms.

Cedar was already combining reference counting with a cycle collector, as one of the very first systems programming languages with automatic resource management.