a ground-breaking new theory

Not really, but I thought I should share my lunacy with our faithful readers lest they think sanity a cohort of the academy. Be warned. This is what a poor upbringing in Latinity and too little sleep produces.

It is my contention that the gerund and the gerundive are wholly unrelated.

I've been wondering for a while about the old Indo-European -r/-n stems and the Latin gerund. We know that the infinitive is used in place of the nominative and the accusative of the gerund. We also know that the infinitive and the gerund are both old verbal nouns. Even if they did not initially comprise a de-verbal -r/-n declension (e.g. Gk. u(dwr, u(datos wherein the -a- is a vocalized nu), it may be helpful to treat them so. A simple example:

N. amare, 'loving'
G. amandi, 'of loving'
D. amando, 'to/for loving'
Ac. amare, 'loving'
Ab. amando, 'by loving'

The one wrench thrown in the works here is the accusative with a preposition, which would befor example 'ad amandum.' This has doubtless arisen through comparison with the gerundive, which I think was originally the middle participle equivalent to Greek -menos. Sihler treats of this possibility briefly, though the evidence for a shift from *-mnos to -ndus is at best unclear.

But I find it helpful to think of it all this way because the gerund and gerundive are not given the attention due in early education and, much like the supine, are a source of endless difficulty for people like me who learned Latin from underprepared grad students through a high school textbook.

If we treat the Latin gerund in this way, its primary difference from the Greek infinitive is in that it expresses case through inflection while the Greek does so through the article.

If you think that's crazy, wait until I post my thoughts about the supine.

UPDATE: I found a document on my computer filled with things like this, and here's the section on -nd-. The Greek and the macrons require a good unicode font, so enter the 21st century already:

nd infix (as in the so-called gerundive—but not in the gerund.)

This is best treated as an old marker of the middle voice. The gerundive would then properly be a present middle participle [cf. verbal adjectives in (u)ndus, which are participial and often have reflexive force].`

A&G collect together with this the suffixes minus and mnus, which are found in some old participial nouns like fēmina and alumnus. This was prescient on the part of A&G as the most attractive formal explanation for the gerundive suffix is a syncopated middle participle ending of the ομενο type familiar from Greek. Why mn should result in nd at some stage of Latin development is presently beyond us, but some brave adherent of the comparative method may find a sliver of evidence in the following pair:

Gk. τέμνω, L. tondeō , “to cut”

In much the same way that the supine is compared with and taught alongside the perfect passive participle on the grounds of a formal coincidence , the so-called gerundive is named after the gerund on the basis of its ending alone.

The mysterious gerund, like the supine (and to a lesser extent the infinitive), is simply an old type of verbal noun gone defective and used only in special constructions.