There is a theory running around that the Latin first declension descended from an uncountable plural declension in an older language, hence its similarity to the second declension neuter plural. Or was it the other way round?