In Berys Gaut and Dominic Lopes, The Routledge Companion to Aesthetics (2001) 

This document is a summary of Currie. 
My personal comments are in red. 
These comments have been added to help students understand his arguments.

Some works of art invite us to imagine something. There are two basic kinds of imagining. [Notice that imagining is our topic here, not imagination! Imagining is something that we do; it's an event. If I imagine that Tom Sawyer is lost in a cave, my imagining this situation is an event that takes place at a specific time. So imaginings are events with specific durations.] 

• The Recreative Imagination: As an event, imagining simulates another sort of event without being a full, literal recreation of it. For example, the event of visually imagining X is a simulation of visually perceiving X. 
• The Creative Imagination: An event in which one produces ideas that defy expectations and conventions. [As opposed to generating something new by a trial-and-error process, or by mere variation on existing things, creative imagining is supposed to happen mentally and without determining reasons.] 

Everybody uses recreative imagination, and it is not itself evidence that one can or does employ creative imagination.

Representational content in art engages recreational imagination, but complex art may demand creating imagining of the audience (they must produce fresh interpretations that are not obvious). The production of original art requires creative imagination. The rest of this analysis focuses on recreative imagination.

Imagination and Fiction

Representational art (e.g., paintings with a recognizable subject, most literature, films, etc.) demands recreative imagination. 

Kendall Walton contends that this art furnishes prompts, guiding games of make-believe. What it prompts us to imagine, it "makes fictional." What is prompts us NOT to imagine, it makes fictional the opposite. Most works are neutral about most things. Sometimes this neutrality is important, as when Shakespeare's Hamlet remains neutral about his motives. On this view, being fictional is explained by imagining, not imagining by the fictional. [Imagining is the more basic category]  On this theory, it is "fictionally true" that Hamlet is Danish, and "fictionally false" that Hamlet is 75 years old. To be fictional is not equated with being false.

Real people and places are fictional when they appear in fiction (when we are prompted to imagine them in certain ways), but what of fictional characters who aren't real? Since we can come up with all sorts of "empty names," (e.g., Zeus, Vulcan), this is not a special problem involving imagination and fiction.

Photographs are an interesting case. If their aim is to "induce belief" (to make us believe that something is to be seen in the photo), then they do not seem to be fictions. For they do not prompt us to imagine.

For Walton, all photos are fictions. Like fictional literature, they present us with something to understand AND prompt us to imagine things conforming to the descriptions provided. But we can understand without following the prompt. Photographs do not work the same way; we cannot understand them except by imagining that the two dimensions are three. Since all photos and all films require imagining, they are fictions. [Remember that imagining makes fiction, not the other way around.]

But some people (e.g., some persons with autism) have an ability for visual recognition without being able to engage in imagination, so perhaps pictures do not automatically involve recreative imagination. In that case, pictures aren't automatically fictions.

Perhaps we should impose a fiction/non-fiction distinction on photographs, films, and even paintings. The fictions will be those intended to prompt recreative imaginings.

Belief

Notice the opening claim about recreative imaginings. An event simulates another sort of event, NOT that something "in the head" simulates objects. Visual imagining simulates visual perception, NOT that an imagined battle somehow "copies" a real battle. The parallel is between events.

Literature often prompts non-visual imagining. What does this event simulate? One view is that it can be more like believing a piece of information (a proposition) than it's like visualizing something. We make inferences when reading stories, and there is no difference in the inferential processes with fiction and non-fiction. (Holmes is in London and it's the 19th century, so if he's in Chicago in the next chapter, some time must have passed.) So narrative fiction introduces imagining that copies believing (without fully being a case of belief).

Is it a REAL case of believing? Not as we normally understand belief, as including a disposition to ACT in certain ways [you don't really believe you're going to the movies at 7:00 p.m. if it's 6:59 p.m. and you've made no effort to get there]. At best, imagining a story is true is a "weak" version of belief.

But perhaps it is not ANY sort of belief. There is "something wrong' [very wrong!] with someone who believe P and not-P simultaneously. One has an epistemic defect in one's belief system. [If you believe that you're eating tuna and believe that tuna is a fish, AND you believe that you are eating chicken and not fish, then there is a big defect somewhere in your "epistemic condition." You need help.]

But with fiction, I imagine P when I believe not-P. But there is no epistemic defect in doing so. Therefore imaging is not a kind of belief.  [I do not believe that a person named Hamlet talked to the ghost of his father, the deceased king. But when I read the play, I imagine that Hamlet talks to the ghost of his father, the deceased king. And I'm not in a condition of epistemic defect when I do this. So imagining isn't really subject to the conditions that apply to beliefs.] At best, imagination is simulated belief. 

Emotion

Does this help us to solve one of the classic problems about fiction?

These three propositions all seem plausible:

1. We fear for characters in fictions who are in danger.
2. To fear for someone we must believe they are in danger.
3. We do not believe in the dangers described in fictions.

Problem: Any pair of the three will exclude the third.

Solutions:

• Accept the irrationality of our response to fiction.
• Deny 1 by saying that we substitute a real object for whom we fear.
• Deny 2 by saying that thoughts, not beliefs, are sufficient.
• Deny 1 by saying that there's never any real fear. Just as we fictionally believe, we only fictionally fear.
• Deny 2 by saying that belief is not a condition for fear.
• Embrace 3 but substitute the idea that we can simulate/imagine believing, and we do something parallel with 2, creating a functional parallel to fear in 1. This is Currie's solution.

The problem is not just one of emotions and fiction: we can also recast it in terms of belief, where the puzzle is how we can believe something about a character and desire that it be otherwise, when we do not believe there is any such person.

So desire seems to have an imaginative counterpart, not just belief and emotion.