Depends on whether

But, but, unicorns don't exist! — Agent Smith

can be prevented from arising, at least within discourse referring to the story. You might say that a feature of fantasy as a specific genre is that such a protest can be suppressed indefinitely. The discourse (the variety of things it's helpful to say about the story) can be shielded from normal influences and standards of conduct. Selectively, that is. Shielded from normal standards of evaluating existence claims, but thereby enabled to apply normal standards of inference about certain events in the story.

But fiction generally, and even fantasy to some extent, seems to have things to say about things outside of it. Consequently, interpretations of sentences, such as your formal paraphrase of a sentence, are likely to be judged with some degree of reference to real-world criteria.

Then your question, how the fictional sentence should be interpreted, is fair, and I think there are two kinds of answer: half-measure and full-measure.

Half-measure is some way of relativising the statement to the story. To talk about the fiction-related discourse from outside. E.g. qualify statements by way of disclaimers like "in the story" or "fictionally speaking".

Full-measure is to seek to reconcile the truth of the fictional statement with that of factual statements. The popular way to do so is to treat the statement as on a par with conditional or hypothetical statements: so that they might paraphrase along the lines of, e.g. "suppose for the sake of argument that there exists an x such that..."; or "consider the set of possible worlds in which there exists an x such that...".

Less well known is Goodman's approach, in which the story is acknowledged literally false, but allowed to be metaphorically true in a manner that has the novel advantage of being about the real world.

Your formulation has at least one unicorn. You need some additional structurer to rule put multiple unicorns.

Why not R(a), where "a" is the unicorn's proper name, with the domain set to range over mythical creatures?

Muchas gracias. I'll need some time to process that. It's not a surprise at all, now that I think of it, that logic (being reality-oriented) is so fiction-unfriendly. Care to comment on my reply to Banno below.



"A unicorn" I thought means "there is at least one unicorn"

Ra= Aaron (the unicorn) is running?

What about Descartes' cogito then?

If nonexistent things like Aaron the unicorn can run then cogito ergo sum is false.

"a unicorn is running" — Agent Smith

Logically, it's a simple claim: there is a unicorn, it is running. There is no problem with it at all except it happens to be false concerning this world. And there is no problem either with "In Narnia there is a unicorn that runs" which can be true of Narnia, a fictional world.

Why agonise over this stuff? We understand - you understand that unicorns are mythical creatures and that is the reason for using it rather than a rabbit.

You ask good questions mon ami, good questions. It's just that fictional stories are logically rich and I would much prefer it if logic could be used on them; alas this wish, like all my other wishes, won't come true.

The Logical Status of Fictional Discourse

It's fiction by intent. "A Unicorn is Running" pretends to make an assertion. So long as we keep the domains of discourse clear, there should be no problem.

Nothing to do with the cogito.

It's fiction by intent. "A Unicorn is Running" pretends to make an assertion. So long as we keep the domains of discourse clear, there should be no problem.

Nothing to do with the cogito — Banno

Yep, that seems to be the most logical option.

I tried translating "If I think THEN I exist" into predicate logic but all I can manage is nonsense like this:

$\forall x(Tx \to \exists y(y = x))$. If I instantiate using d = Descartes, I get $Td \to \exists y(y = d)$. In English, if Descartes thinks then there exists something that is identical to Descartes.

In propositional logic there's no issue.

If the light blips then atoms exist

$\forall x(Bx \to \exists x(Ax))$

If x thinks then a thinker x exists

$\forall x(Sx \to \exists x(Rx))$

,

You might use a free logic.