I agree with @MixingManiac's advice and analogies. Good stuff.
to expand on his advice a bit... it's not necessarily the harmonic or frequency content of the material as much as it is repetition and leitmotif via sound design. "Subconscious" sounds and not "literal" sounds are likely the way to go in your situation. It does indeed sound like you're about to have a lot of fun and freedom to experiment. I'd definitely take advantage of that type of situation as much as possible. If the directors work is quality then it'll likely end up being a really good demo/showpiece.
My experience example would be a horror film I worked on where I decided upon several core source material sounds and a couple specific processing techniques to use as the basis and representation for the presence of all of the "evil" in the movie (mostly because it was a ghostly presence as opposed to a specific entity). Limiting my sound sources allowed me to focus on very specific things without having to waste time digging through stock libraries. I got to make all my own field recordings and make it 100% original. I was then able to focus more on the process and application over how it was necessarily going to be perceived. In effect, I was creating and forcing the way it was to be perceived. If you want an example you can hear, then you can check out the clip below.
Here's some info on the sound sources and techniques:
In this case, all of the sounds started with field recordings I made of my daughter laughing and talking as a happy 2.5yr old does, prepared and tortured violin and piano (typically using a very jagged toothed metal file), a bunch of creaking doors and a couple other random things. I then recorded them to a Studer 2" analog tape machine at 30ips and recorded them back into Pro Tools at 15 and 7.5ips to get a more natural timestretching and pitch shift. I then left some of the sounds as is regular speed with no further DSP wankery, maybe all it required was just good placement, trimming, fades or reversing. Some was more extensive going through DSP FX chains that were 10 plug-ins deep or maybe some very extreme spectral based time stretching (or a combination of both). Another example in there is that I used violin plucks/pitch harmonics drenched in reverb for the blood droplets. Some of the door creeks are used as is with no processing, that was how I recorded them. Or how I used the file scraping on piano string recordings to represent the abrasiveness of the concrete or the grinding teeth. The opening sound is a single word from my daughter put through extreme spectral timestretching (which I use a lot). It almost takes on the characteristics of a ship hull resonating from tension.
What it mostly goes to show is that you don't necessarily have to use recordings of traditionally creepy things to use as creepy sounds (ie: literal or cliche). In the case of my daughter laughing and rambling on in the way a 2yr old randomly does when they're happy it was just the opposite. It was now the context I used it in and the repetition that bred familiarity. So over the course of the film, once you start to hear these similar sounds happening you subconsciously started to associate them with the fact that some evil stuff was about to happen. As is very often the case, sometimes creepy and dissonant sounds can represent love better that lush warm synth pads or wind chimes. Mostly because it typically better represents the feeling of longing and depth better than the aforementioned cliche sounds. Sometimes it might require using those cliche sounds in a new way or context like I did with the creaking door sounds (which are admittedly a bit cliche for a horror film) and processing then using analog or spectral timestretching. The important part is that I just didn't use the door sounds for a physical door being shown on the screen. I used it as a tension sweetener.
A good Hollywood example would be the "Chi chi chi, ha ha ha" vocal sound effect from the Friday the 13th series. You hear that specific sound and you know what's up. Of course it'd be really silly if Jason went around making that sound with his mouth, thus it's a subconscious sound and informs the listener about the emotion they should be feeling through it's repetitive usage and it's selective placement. Unlike using Darth Vaders iconic leitmotif style breathing as an example, which is very well a sound that is supposedly actually happening due to a breathing apparatus in his suit and therefore not a subconscious sound that doesn't actually happen in real life (like Jason making funny breathing delay effects with his mouth).
Hope that helps!
I can't necessarily think of any other examples off the top of my head right this second and sorry for only using horror films as my reference (it's just that they're typically so damn good at stuff like this). I'm sure some other people will have some other examples of subconscious sound effects.