After Sophie's Choice won piles of undeserved acclaim from critics who, one assumes, were too chickenshit to criticise a movie about the Holocaust, Alan J. Pakula lay silent for a long time. Three years and some months passed until his next movie came out, the longest gap in his career as either director or producer at that point; perhaps he needed to rest up from having pushed out something as big and ambitious as Sophie's Choice on a relatively quick production schedule. Certainly, I do not imagine that he spent all that time developing his next project, making sure the screenplay was perfect and spending a great deal of time thinking about how he wanted to shoot it; and if that is where he spent those years, it was time wasted. For he broke his silence with Dream Lover, a film that is so... you know what, actually, as I sit here thinking about it, it's still a better movie than Sophie's Choice. At the same time, it's pretty inane and possesses one of the least-dramatic plots of any nominal thriller that I have ever seen. So continued Pakula's swift descent into banality.

Let me say in the same breath that I don't think Dream Lover is therefore an impersonal, careless hack-job - the more Pakula I see, the more I am positively convinced that this conception that he was just a post-'70s washout who cranked out movies for the paycheck (a belief so very common among people who rightly view his 1990s films as weak, flavorless things) is not based at all in reality. If you were dropped in front of Dream Lover wholly without context, that would be a defensible position to hold, but we are not here coming in without context, but with nine previous Pakula movies to explain his predilections, and viewed in that light, Dream Lover is quite plainly a work that exemplifies one of the most important recurring themes in his career: a fascination with the science of psychology, and a detached, clinical treatment of characters undergoing some kind of emotionally traumatic experience that resembles the notes from a psychiatric session more than a legitimate screenplay. If Dream Lover doesn't work remotely as well in those terms as did Klute or Starting Over, to name the two Pakula films to most prominently fit that mold, that has a great deal to do with the fact that in place of a truly great actress like Jane Fonda or a fun movie star like Burt Reynolds, Dream Lover has to make do with Kristy McNichol.

Man, do you remember the huge Kristy McNichol fanbase back when she was a teen star? Well then I'm really sorry for you. I only got to know McNichol in her years of decline, after she was diagnosed with depression (later found to actually be bipolar disorder). I do not want to diminish the reality of how awful this condition must have been for a young woman in her position, but this isn't a therapy session: it's a film review. And in that capacity, I would propose that having pronounced psychological issues did nothing at all good for McNichol's acting. Oh, she does the work mechanically well, I suppose: which doesn't come as a surprise, given how well Pakula was as a director of actors. She delivers the lines without sounding stilted, she moves around without obviously trying to hit her marks. But in terms of actually putting anything like emotion into the part, well as they say, the light's on, but nobody's home. Near the end of the movie, McNichol is called upon to play her character in a waking dream state, basically a glazed-over fugue, where she does little but stare absently and keep her tiny little mouth closed in a half-smile. I uncharitably thought as I was watching it that this was the first time in the movie I actually believed her character, although I couldn't necessarily tell that there was a difference without cues from the dialogue.

But let us not assign all the blame to poor McNichol. There are a great many flaws just in the screenplay, by a certain Jon Boorstin. His erratic career took him into TV and eventually nature shows after this, his debut work as a writer; his sole film credits prior to this were as associate producer of All the President's Men, and as the director's personal assistant on The Parallax View, which suggests to me that Pakula came up with a the notion for a movie, but he didn't want to trouble himself with writing it, and so farmed it out to his little lackey. The results are about what you'd expect from that: very functional expository dialogue, and since the movie is rotten with contemporary science theories, there's quite a lot of exposition to be had.

The biggest problem with the screenplay, though, is its incredibly tedious lack of conflict. Here's what happens: Kathy Gardner (McNichol) is studying flute while living with a domineering father (Paul Shenar). It so happens that she snags a prestigious scholarship to jazz improvisation for one month in New York City, and after some brief strain between herself and her dad, he lets her go. Things in New York are just swell, up until the night that a creepy guy that Kathy met her first day in her apartment (he was looking for the previous tenant) comes in and attempts to torture and rape her. In the struggle that follows, she manages to grab his knife, and stab him to death. A clear-cut case of self-defense, but Kathy's father doesn't see it that way, and forcefully suggests that she should claim to have blacked out during the attack.

Immediately following this event, Kathy's dreams - formerly replete with turn-of-the-century British garden parties - are filled with replays of this attack, over and over again. Not able to get sleep, Kathy eventually finds her way to a sleep disorder research center, where she becomes the willing guinea pig for Michael Hansen (Ben Masters), who instantly recognises in her a 1-in-10,000 opportunity to study lucid dreaming in humans. And for the rest of the movie - and while it's only 104 minutes long, there's a whole lot still to go at this point - the film consists of two things: hearing Hansen explain the latest and greatest in dream theory, or wondering if Kathy will ever be able to get a proper sleep again. Eventually, Hansen's experiments go wrong, and Kathy enters a sleepwalking state where she carries around a knife and doesn't try to stab anybody. Still, it's a kind of action, relative to the rest of the movie.

That's seriously the main conflict: between Kathy and her dreams. Pakula and Boorstin wager everything on the audience finding mid-'80s dream research to be the height of compelling cinema, and while I can't say if it was even borderline acceptable in the mid-'80s, by God it's a waste of time 23 years later. So many possible plots are presented and abandoned: Kathy's deception of the police, her uncomfortably quasi-sexual relationship with her father, the tension generated by wondering when somebody will realise that she's actually a really shitty jazz flautist (though it was the '80s). All of it dropped in favor of the scintillating depiction of a housecat that walks around while dreaming; a silly thing to have a housecat do, since they're already so cute when they sleep and their little paws twitch and they smack their little lips.

With a screenplay that could not possibly be any more boring - it's not even particularly bad in a way that's fun to make fun of - all that is left is for the film to have some kind of technical competence, and it does, in a vague way. At any rate, it's a better piece of film directing than Sophie's Choice, while still being infinitely worse than anything else in Pakula's career to that point. His treatment of Kathy's dreams is the only really interesting part of the film, and even these sequences are undeniably redolent of the artsier breed of horror movie (if Pakula were alive today to tell me that he'd never seen a Dario Argento film prior to making Dream Lover, I'd call him a liar to this face). This film re-teamed Pakula with Sven Nykvist, absurdly, and there is a certain cod-Bergman feeling to the way that the scenes are lit and framed; it feels like one of that Swedish master's chamber dramas (especially Through a Glass Darkly, though there are at least a couple direct lifts from Cries and Whispers), which would be profoundly offensive if it weren't so damn funny. The hell of it is, you can tell that Pakula and Nykvist actually though about what they were doing; it is quite deliberately-composed, and I suppose that if I bought the concept of the movie, I'd find its grim, underlit interiors to be quite a striking evocation of Kathy's mental break. This much, I will at least say: it has fantastic sound design, particularly in the dream sequences. I cannot name another director who has such consistently good sound design in his movies; this was his third film in a row to utilise Anthony J Ciccolini in that capacity, and the best-sound of the three. If the movie ever really works as a thriller - if the mood ever picks up out of dullness - it is solely because of the skilled way that dead silence is intermixed with tiny blasts of sound.

This is one of the most soundly-ignored films in Pakula's career; it has every reason to be. If I believed in the existence of Kristy McNichol completists, those are the only people to whom I could recommend this film, which is dully acted, and even more dully written, and though it looks pretty damn fine, the cinematography presumes to belong to a much better film. But you know what? It doesn't whore out the Holocaust. So at least we can safely say that the director was on an upswing.