100. Without sensible signs, man could not even conceive abstract ideas. In fact, what are abstract ideas? They are simply qualities of beings contemplated by the mind in their ideality, and apart by themselves; they are mental conceptions. Now, where are the objects of such ideas to be found? Nowhere but in the mind itself....
...Now, how is this mode of conceiving possible? I answer:—
102. By means of an external sign, a sign which by holding the place of whiteness apprehended by the mind, gives it an existence also outside the mind; a sensible sign of the idea which is not sensible; in short, a word directed to single out the whiteness from among the other objects that surround it so long as it is perceived along with the bodies in which it really exists.
Here we have the essential feature of linguistic or semiotic rationalism: we have abstract ideas despite not sensing them; we cannot create abstract ideas by pure force of reason; we must have sensible signs of abstract ideas in order to think of abstract ideas. (Rosmini insists fairly strongly elsewhere on the 'abstract' qualifier; he does not hold that this is true of words that do nothing but mark common sensible experience.) In the New Essay Concerning the Origin of Ideas (1830), he summarized it quite straightforwardly (at section 521):
Humanity has no existence outside of the mind; it cannot therefore attract our attention, unless in some sensible sign which, being external to the mind, holds the place of that idea, giving it, as it were, subsistence. It is impossible, then, for the mind to conceive abstract ideas, that have no realities corresponding to them, unless it be moved thereto by sensible signs which may take the place of those realities and represent, or, to speak more properly, raise them before it.
To return to the Theodicy. On the basis of semiotic rationalism, we move on to get an indirect formulation of the traditionary argument itself:
From this it is plain that external signs were necessary to man in order that his mind might associate and bind up abstractions with them. But he could not invent those signs by himself, for the reason that to invent, he must already have been possessed of abstractions, which, nevertheless, he could not acquire save by means of words. God, therefore, imparted to him a language; that Supreme Instructor taught him the use of some words, in which the abstractions, contemplated together with them, might, so to speak, appear outwardly subsistent. These words could attract to themselves the attention of the mind, and determine it to fix itself on special qualities apart from the objects in which they exist. All this in accordance with the general law, that the human mind must primarily be moved to act by the impressions made on the sense by external objects.
Let's call the claim that inventing signs requires that one already possess abstractions that can only be acquired by means of signs, the Prior Possession Principle (PPP). It plays a key role here, since it is what motivates the rejection of the idea that human beings could invent their own language to speak about abstractions. To invent a language suitable for abstract ideas, one would already need abstract ideas; to develop abstract ideas, there would need to be a language already in existence. Thus the human mind must receive language from the outside, by a teacher; and from this one can get to a first teacher, who cannot be human or have a mind limited in the way human minds are.
Later, however, in the Psychology (1846), Rosmini will pull back from PPP:
We have elsewhere expressed the opinion that men could never by themselves have come to think and designate pure abstractions, for the reason that had they not in nature any stimulus to do so, and hence we inferred the divine origin of this portion of language. Since then, we have given this subject more mature reflection, and now that proof does not seem to us irrefragable. Let us, therefore, distinguish between the question of fact and that of simple possibility. As to the fact, there can be no doubt that the first man received his impulse to speech from God Himself, Who, by speaking to him first, communicated to him a portion of language. The arguments that prove this we shall set forth in another place. But confining ourselves to the simple metaphysical question of whether the human family (not an isolated man) might possibly in the course of time have come to think at least a few abstracts, designating them at once and with one and the same complex operation, by words or other signs, it seems to us now that it may be answered in the affirmative, and that we have discovered the stimulus we had vainly sought for before, by which the human understanding might have been moved.
(I'm not sure what 'other place' Rosmini has in mind here.) This is not, of course, a rejection of semiotic rationalism; Rosmini immediately goes on to say that "it is undoubtedly necessary that we should be able to find in real nature something to link the abstracts to, so as to serve as a natural sign of them; because it is only on this condition that the attention of the human mind can rest on them and seize them." What has changed is that he thinks that human beings who are already using a rudimentary language already (at least sometimes) have (some) such signs available. This happens by metonymy: 'hand', which you can just use to label your hand, becomes used for the things you can do with it (thus, although this is recognized only in retrospect when we have produced the abstract idea, hand+power), because it's much easier to keep using the same words than to keep inventing new ones. Once you are doing this, though, you already have the sign and therefore can form the abstraction. And thus effectively, given the sign 'hand' you can get the idea of power, and it's even the case that over a long time you might use it almost solely to mean power.
If we think of the traditionary argument as having certain basic elements, we can say that Rosmini accepts Deficiency of Empiricism and Necessity of Language (and thus upholds the semiotic rationalism of the traditionalist; but over time he came to think that teaching (in the traditionary sense of receiving one's language from another) was not strictly necessary, although there were reasons to think that, in fact, some language use does trace back to divine teaching. This is in part, however, because he holds that language has different 'portions', and it is only one portion -- the portion of language used for reasoning with abstract ideas -- for which the question is even raised. He continues to hold that a single human being could not possibly invent language for discussing abstract ideas. But he later comes to think that as long as a group of interacting human beings had developed a rudimentary language for labeling things they actually sense, this would actually suffice: laziness, so to speak, would lead to using labels loosely, and this loose use of signs serves as the semiotic basis for abstraction. This would not convince a hardcore traditionalist, I think; I think the traditionalist would argue that Rosmini is already attributing to pre-abstract thought things that could only be possible with abstract ideas. But the route he takes, appealing, in effect, to figurative language to get us beyond mere labeling of things we can sense, and thus to provide a way around PPP, is an interesting one.