On the occasion of Steven Soderbergh’s announced retirement from filmmaking, and having gone out on a limb in praising his approach to his art and craft having seen fewer than half of his films, I decided to go back and fill in at least the most important gaps in my knowledge of his filmography. I started by taking in his two most recent features, “Magic Mike” (which I had missed last year) and “Side Effects.” And, apropos particularly of the latter, I found myself meditating on Soderbergh’s relationship to genre expectations, and how his approach winds up affecting his audience.

Genre expectations are something every writer of narrative art has to be cognizant of, because they shape the reader’s (or the audience’s) response. If you fulfill those expectations in a familiar way, your audience will likely get bored, though they might experience the familiarity as pleasing – old-fashioned sitcoms in particular depend on the pleasure of repetition. If you ignore them, your audience will likely be lost – they will likely experience the art as “bad” because they don’t understand what they are supposed to pay attention to and why.

A lot of interesting narrative art frustrates the reader’s or audience’s narrative expectations in ways that draw them more fully into the story. The classic example would be Hamlet. Hamlet is, on its surface, a traditional revenge tragedy. Before Hamlet has met the ghost, he knows that his uncle has usurped his own place on the throne, and on top of that he has other “issues” with his mother’s hasty re-marrying. Then Hamlet meets the ghost and gets his mission: revenge. Most of the rest of the play consists of Hamlet delaying in executing this mission, drawing more and more attention to himself in the process, ending in a general massacre that leaves the country in the hands of a foreign dynasty.

In Shakespeare’s source material, the prince delays for a practical reason: the murder is known. Amleth pretends to be mad so that his uncle, Feng, doesn’t kill him the way he killed his father. He uses his wits to evade the various traps Feng lays for him and, when the time is ripe, achieves his revenge. Shakespeare removes this motivation: in his play, the murder is hidden, and Claudius makes Hamlet his heir. Hamlet’s madness, and his delay, become puzzles. Along with other structural choices – particularly the decision to embed the Hamlet revenge drama within a drama about Fortinbras’s revenge for the killing of his own father, and embedding within it another drama, about Laertes avenging the death of his father, Polonius, the drama becomes, in part, a meta-narrative about the meaninglessness of revenge, and, by extension, of all action. But it only works at doing that because Shakespeare preserved the overall shape and familiar narrative conventions of the revenge tragedy. Hamlet looks like he is cunningly plotting revenge, and so we think we understand what kind of play we are in – which is why we keep watching. But as we think about what he is doing – and he helps us by thinking for us in soliloquy – we realize that what he’s doing doesn’t, actually, make sense, and furthermore that he knows it. And thus a hoary revenge tragedy long-since mocked for its absurd plot elements becomes the vehicle for creating the first modern existential tragic hero. Which he wouldn’t be if we didn’t think, at the start, that he was the hero of a familiar revenge tragedy.

Why am I going on about this? Soderbergh is known, as a director, as someone acutely aware of genre expectations, and someone who puts himself at their service, but who uses them to mislead the audience, and thereby achieve unexpected effects. He does this frequently with music – in “The Informant,” for example, the soundtrack misleads the audience into thinking that Matt Damon is playing a comic hero – that he’s a kind of Inspector Clouseau, incompetent but with a grandiose sense of his own importance (which, arguably, is what Matt Damon’s character genuinely thinks he is). That makes the twist, when we realize that his character is some kind of extravagant con artist, much more effective – because we were expecting, with eager anticipation, a very different plot.

“Side Effects” plays a similar game with the audience, but even more comprehensively. For the first third, maybe even the first half of the film, we think we’re in a movie about the social and ethical dilemmas of psychopharmacology. Rooney Mara, whose husband, Channing Tatum, has just been released from prison (for insider trading), is struggling with depression. After an attempted suicide, she starts seeing a psychiatrist, Jude Law, who, along with talking therapy, puts her on various medications. Nothing seems to work, until he puts her on Ablixa, recommended by her former psychiatrist, Catherine Zeta-Jones. Suddenly, the fog of depression lifts, she’s able to sleep at night, her sex drive is back – it looks like she’s making great progress. But there is a nasty side effect: she starts sleep-walking, even setting the table and playing music while unaware of her husband trying to wake her. In spite of the side effects, she doesn’t want to give up the drug because of the good it’s doing at relieving the depression.

And then, one afternoon while chopping vegetables in her sleep, she stabs her husband to death.

Law, the good psychiatrist, works to get her acquitted at trial, since, in his view, she can’t be held accountable for her actions committed entirely without conscious intention. But, as the D.A. points out to him, if it wasn’t her fault, then it was his, for prescribing the medication that caused her to commit the crime. And, even as Law succeeds in keeping Mara out of prison (she’s sent to a psychiatric facility until certified not to be a danger to herself or others), he is duly punished – fairly or unfairly – for her crime: he loses his other patients, his partners cut him loose, the drug companies drop him from paid positions organizing clinical trials, etc.

It feels, even after the murder, like we are in a film about psychopharmacology and how it messes with our sense of intentionality, and hence with our ability to talk about right and wrong actions. Then Law starts to notice holes in Mara’s story, holes he never noticed before, and becomes convinced that she is faking it – and still, we might be in a movie in which Law is driven nuts by the imputation of responsibility to him of her actions. That’s certainly the movie that Law’s old partner, Peter Friedman, thinks he is in, the movie that Law’s wife, Vinessa Shaw, thinks he is in. Until he gets definitive proof, at which point we finally know we’re in a Hitchcockian psychological thriller.

The best thing about the film is the way in which Soderbergh builds his trap for us. For much of the earlier part of the film, we’re “with” Mara, and both the way it is shot and the way it is scored encourage us to believe the story we are being told, about a woman with terrible depression. It’s not just a matter of tricking us about her character – it’s also a matter of tricking us about what genre of film we’re in, which is to say, in part, what kind of moral universe we are in: a world in which intentionality and fault are fuzzy concepts because we understand too well their chemical origins, and in which medical professionals elide easily into repositories for our discarded moral sense, the people who are ultimately responsible for our actions, though they are, in reality, just human beings like ourselves.

But the shift of genre upends this world. In a psychological thriller, your only hope of remaining sane (and, indeed, alive) is to absolutely trust your own perceptions – to be psychologically isolated and morally ruthless in pursuit of a righteous end that only you really believe in. You have to embrace, in other words, a hypertrophied version of who Jude Law is at the beginning: a (male) doctor with the power, literally, to control the minds of his (female) patients.

I bring up gender here because when the genre shifts we become aware of how deeply-woven into the narrative is a Hitchcockian misogyny. Mara and Zeta-Jones, who hatch and execute the murder plot between them, are two exemplary female villain types: Mara the deceptively vulnerable manipulator, Zeta-Jones the ice queen. And to top it off, they are lesbian lovers. But it goes well beyond them. Shaw plays Law’s wife as utterly unsympathetic to his plight before he starts acting crazy – indeed, she is more furiously condemnatory of his behavior in the Mara case than anyone, with the possible exception of Law’s female partner, Laila Robins, who seethes and rages where the other partner, Friedman, shows at least a trace of empathy. The only two women who seem to have normal human feelings, in this movie, are Tatum’s mother (Ann Dowd) and Mara’s Attorney (Sheila Tapia). Not coincidentally, both are made up and played to lack even the slightest hint of sex appeal.

The murder plot, by the way, is wildly implausible, even for a genre film. Mara’s character is undermotivated – there’s no evidence from her performance that she genuinely loves Zeta-Jones (Zeta-Jones is besotted one), but if she’s just after money why doesn’t she stick with Tatum? Or dump him and find another Wall Street hotshot? Why murder? Then the plot is insanely risky for her. First, she must commit the murder successfully while pretending to be asleep. If Tatum disarmed her (not implausible – he’s Channing Tatum, and he’s just survived a prison term) the plot would be sunk. Then, she must get off with an insanity defense. Her own lawyer says that’s 100-to-1 against. Why would she take those odds? And all this not to get Tatum’s money (he’s broke) but to pull off an insider-trading scam related to Ablixa. But the trades were executed by Zeta Jones, a psychiatrist who prescribes Ablixa! The SEC would be all over her if she put on a massive short in that stock – they would assume she had inside information about, for example, really bad side effects. Why take these insane risks?

And then, as the coup-de-grace, once Law fully sees through to the truth, he has a ridiculously easy time of it manipulating the two women and trapping them in turn. He uses his power as a doctor to cut off communication between them, convinces each of them that the other one has betrayed her, uses Mara to convict Zeta Jones and then, again, uses his apparently unlimited authority as her physician to trap her.

Precisely because I had been prepared by the first third of the film to believe that I was in a morally-complex and real universe, it was impossible for me to ignore these absurdities. Instead, the absurdities redounded upon the genre. The latter half of the movie made the most sense as Law’s revenge fantasy. He is able to reassert emotional and physical control over his life, as a doctor and as a man, but only by obliterating everything he knows about human psychology and operating in a psychological thriller world. But it isn’t directed that way. It’s directed as if we are supposed to be along for this particular ride, buying into this as a “solution” to all the problems thrown up at the start.

The penultimate scene in the movie has Jude Law restored to the life he loved. He has his wife and stepson back (he lost them along the way, ostensibly because Zeta Jones falsely smeared him as having an affair with Mara, but underneath we understand that she’s walking out because Law’s become a loser), and we see him picking up his son at school and bringing him to their black SUV – which Shaw is driving. And I thought, why is Shaw driving? Shouldn’t Law be back in the driver’s seat? After all, he’s put the scary women in their place again.

My instinct is that Soderbergh intended this to be a kind of hedge. See, this movie isn’t really as classically misogynistic as it appears; Law’s a sensitive New Age guy at heart, with a great relationship with the kid, and comfortable letting his wife drive. But that’s not the effect it had on me. Rather, by returning us to the world of the first half of the film, it deepened the misogyny of the overall narrative by not letting us see it at the end. If the whole murder plot can only be understood by a self-conscious audience as a male revenge fantasy, projecting his error and failure as a doctor back onto the patient, then I want to see what impact indulging in that fantasy has had upon the doctor’s character. The answer appears to be: none.

In effect, this is the opposite of what happens in Hamlet, another revenge story with, shall we say, problems with its women. Rather than take a genre and explode it from within, Soderbergh’s movie uses a genre to close down the tough questions thrown up by the “realistic” story. But it’s not so easy to tamp down questions once they are thrown up. Try to repress them too hard, and they may blow up in your movie’s face. Which is fine if that’s what you intend, but I didn’t get the sense, from Soderbergh’s direction, that he did.