I had not heard of this. Is this the reason why I see some diagrams where gerunds - clearly used as nouns - are diagrammed like participles on pedestals rather than on stepped pedestals?
It's possible. I'm not sure. I don't have much experience with Reed-Kellogg diagrams, since we've never done them in school, and when I learned English more in depth at university we skipped over traditional grammar altogether, since linguists don't use it (much).
There's a problem with the term "clearly used as nouns", though. In "Running marathons is fun," would you say "running" is used as a noun? The subject is the entire phrase, not just the participle/gerund thingy (I'm not calling it either, because classification is at issue here). Traditional grammar would call it a nominal phrase, because of its subject function. I'm not sure whether traditional grammarians think that a nominal phrase has to contain a noun in the prominent position ("marathons" is a noun, but it's not the word that defines the phrase).
Because of problems like this, linguists tend to use different theories. I'm sure you've seen those tree diagrams that always branch off into smaller and smaller units. Those are phrase structure diagrams, where you basically split clauses and phrases until you arrive at single words. The definition of word classes don't involve reference to their function, but they're defined by how they behave in a sentence (and what you can do with them). So "running marathons" in a variant of structural grammar is a participle phrase, no matter where it occurs in a sentence. That takes care of cases like "my loud singing" vs. "my singing loudly", because subjects or objects no longer require nouns, but can take phrases (or clauses). [The distinction between phrases and clauses is also handled differently than it is in traditional grammar.]
As I said, I have little experience with traditional grammar, but if a nominal phrase doesn't have to include a noun, then it's possible to have noun-less subject or object, with the phrase itself (rather than the individual words within it) taking over the function. That would also mean that a noun-phrase (a phrase defined over its noun) is only a sub-type of the broader type "nominal phrase". I could see traditional grammarians argue like that and thus treat the gerund as a participle. Whether this actually happens, or whether seeing a gerund diagrammed as a participle is just a "mistake" I don't know, since I've never gone that deeply into traditional grammar. I'm more a structuralist myself.