Before today, I thought that there was no neuter substantive1 noun with a nominative singular in a and a genitive singular in ae. However, I have encountered references to a possible exception: some sources indicate that the noun pascha/Pascha, referring to Passover or Easter, could inflect according to the first declension and at the same time be neuter in gender. I'd like to see if I can verify this surprising fact.
How the word pascha entered Latin
This is supposed to be explained by the background of the word: it comes from a masculine Aramaic word via Greek πάσχα, which is indeclinable and neuter.
Apparently, in Latin, the word came to be declined, but retained the neuter gender of the Greek noun. It is supposed to have also developed a consonant-stem declension by analogy with Greek -ma neuter nouns, gaining a stem paschat- with unetymological -t- that was used to form genitive singular paschatis, genitive plural paschatum etc.
The first-declension forms: sometimes feminine—ever neuter?
It's fairly easy for me to find examples of the first-declension forms: e.g. "pascharum" in "pascharum diem" seems to be a first-declension genitive plural. But it's harder to find evidence supporting the idea that these forms are ever grammatically neuter rather than grammatically feminine.
In fact, it seems that it must have been possible for pascha to be a feminine noun in at least some situations, because the accusative singular form pascham seems to exist: this could not be neuter because it would violate the rule that neuters are always identical in nominative and accusative.
And in accordance with the form pascham, two sources list pascha, paschae as a feminine noun, and not as a neuter noun (reserving the neuter gender for the paradigm of pascha, paschatis): Lewis and Short and Félix Gaffiot.
However, as I mentioned at the start, I have also found sources that say that the noun could be neuter when declined according to the first declension:
Nouvelle méthode pour apprendre facilement la langue latine, 12th ed., by Claude Lancelot (? et al?) (1761) says
Pascha, est du Neutre. : Pascha próximum, Pâque prochain; & se décline de la premiere ou de la troisiéme: Pascha, æ, Pascha, ătis. [...]
les Grecs l'ont fait Neutre, parce qu'ils l'ont pris comme indéclinable [...] les Latins les ont suivis dans le Genre, quoiqu'ils ayent décliné ce nom, ou de la premiere, ou de la troisiéme
The bolded part says "the Latins followed the Greeks [in making the gender neuter] whether they inflected the noun according to the first or third declension".
I found a text that gives an entire declension in the neuter (for the singular only), but it looks a bit suspicious. Unfortunately, I don't have full access to the text; I just saw a page through Google Books. The text is given in Donati Graeci: Learning Greek in the Renaissance, by Federica Ciccolella (2008); if I understand correctly, the Latin is the original, so it shouldn't have been erroneously influenced Greek grammar. Ciccolella says "version a [...] probably originated as a simple word-for-word translation of the Latin textbook for Greeks who wanted to learn Latin. During the fifteenth century, this version, perhaps originally written in the interlinear spaces of a Latin text, became an independent grammar and was used to learn Greek; apparently, Greek Donatus a did not undergo the process of adaptation of Latin morphology to the "target language" that led to the composition of Donati in modern languages" (p. xvi).
Nominativo hoc Pascha, genitivo huius Paschae, dativo huic Paschae, accusativo hoc Pascha, vocativo o Pascha, ablativo ab hoc Pascha, pluralia non habet.
(Donatus graecus a, "De Nomine", lines 53-55, p. 271 in Donati Graeci)
The corresponding lines in Greek (53-55, p. 270) use the indeclinable form πάσχα, but as I mentioned above, it doesn't seem like the Greek translation could have influenced the Latin text because the Latin is supposed to be the original.
The statement that plural forms do not exist seems to be false for at least some stages of the language, since as I mentioned above, the genitive plural form Pascharum seems to have been used sometimes.
Perhaps there is only one instance of a neuter noun of the first declension: viz., pascha—the passover
(p. 12, A Commentary on the Eton Latin Grammar, by Richard Haynes, 1843)
The kind of evidence that I think would work
Is it true that pascha can be a "neuter noun of the first declension", or are these three sources wrong? Aside from the declension explicitly given by the quoted Donatus, I suppose that the only way to clearly show that it is a neuter noun is to find an example in text where it triggers neuter agreement: for example, if Paschae in the genitive singular were modified by an adjective ending in -i. Are there any such examples when the word has a clearly first-declension form?
- I specified "substantive" because I had heard of some common-gender adjectives ending in -a in the nominative singular that allegedly could modify neuter nouns; e.g. advena (mentioned in an answer by fvogel) and apparently also verna (see this passage in Beiträge zur Griechischen und Römischen Literaturgeschichte, Volume 2). I'm not sure whether these unusual adjectives were defective in some cases/numbers; it seems plausible to me (although I haven't checked) that they were defective at least in the neuter nominative/accusative plural, because it seems to have been fairly common for even the more typical third-declension adjectives of a single ending to lack a neuter plural in -a.