How Twin Peaks taught me to see patterns everywhere, and why getting the mystery wrong was sometimes more interesting than getting it right.
Spoiler warning: This article discusses both seasons of the original 1990–91 Twin Peaks television series, including its central murder mystery and finale.
At some point during the second season of Twin Peaks, I became suspicious of the hospital food.
Not the hospital itself, although there was plenty to be suspicious of there. People were experiencing visions, slipping into comas, receiving mysterious injections and behaving in increasingly inexplicable ways. But it was the food that caught my attention. It kept being mentioned, seemingly more often than any reasonable television programme needed to mention hospital catering. By the fifth reference I'd recorded, I was convinced something was going on.
I began wondering whether the townspeople were being drugged. Perhaps the food was the delivery mechanism. Perhaps the strange visions weren't supernatural at all, but the consequence of some elaborate experiment. The more I thought about it, the more details appeared to support the possibility. Medication, injections, altered behaviour, memory disturbances, people seeing things that shouldn't exist. It was beginning to resemble a conspiracy, and I had constructed much of it around the dietary arrangements of a fictional hospital.
This wasn't how I'd intended to watch Twin Peaks. I'd approached the series as a psychologist interested in its cultural influence, its characters and the peculiar relationship between the ordinary and the uncanny. By the time I finished the original two seasons, I'd developed theories involving comas, spirit possession, temporal anomalies, ancestral identities, supernatural forces and an extraterrestrial chess game being played with the inhabitants of a small American town.
Some of those theories were more plausible than others. One or two turned out to be surprisingly close to what the programme eventually revealed. Others quietly collapsed beneath the weight of contradictory evidence. The interesting question wasn't whether I'd solved Twin Peaks, but why certain explanations had felt so convincing in the first place.
And, perhaps more importantly, what was happening psychologically when an apparently insignificant detail began to look like evidence.
When everything starts to look like a clue
One of the peculiar achievements of Twin Peaks is its ability to make the viewer suspicious of ordinary things. A deer head falls from a wall. Somebody mentions a fish in a percolator. A woman carries a log. A man is unusually enthusiastic about coffee. Initially, these moments are funny, eccentric or vaguely unsettling. But as the programme develops, the viewer learns that seemingly absurd details might have significance beyond their immediate context.
That changes the way we watch.
During the first season, I found myself paying attention to conversations and background characters that I'd probably have ignored in another programme. I wasn't merely following the investigation into Laura Palmer's murder. I was trying to understand the entire town, partly because almost anything might turn out to matter. A peculiar remark, an unusual expression or an apparently irrelevant encounter could conceivably become important several episodes later.
By Season Two, that attentiveness had become something closer to suspicion. The programme introduced more overtly supernatural experiences, including prophetic messages, unexplained appearances and increasingly elaborate connections between people and events. Every new development expanded the range of possible explanations. If spirits could inhabit human beings, perhaps they could inhabit animals. If visions could reveal things that hadn't happened yet, perhaps time wasn't behaving conventionally. If somebody appeared in a dream before another person encountered them, perhaps the dream wasn't entirely private.
The hospital food was hardly the most compelling evidence available. But it had one quality that made it difficult to ignore: repetition.
Once I'd noticed the references, each subsequent mention seemed increasingly unlikely to be accidental. I wasn't simply hearing another comment about food. I was hearing what felt like another confirmation that the programme wanted me to notice it.
That is where the psychology becomes interesting, because the significance of an observation depends partly on what we already expect it to mean.
The predictive brain and the suspicious dinner tray
One influential account of perception proposes that the brain is continually generating predictions about incoming information and updating those predictions as new evidence becomes available. Rather than passively recording everything around us, perception involves an interaction between sensory information and expectations about what we're likely to encounter (Clark, 2013).
This framework is often called predictive processing. It doesn't mean we simply hallucinate whatever we expect to see, or that every judgement can be reduced to a prediction-error calculation. The theory is more nuanced, and its broader claims remain subject to debate. But it offers a useful way of thinking about why context changes the significance we assign to information.
Consider an ordinary television drama in which several characters complain about hospital food. We might reasonably interpret the repetition as a running joke, a familiar observation about hospital life or simply a feature of scenes occurring in the same setting. In Twin Peaks, however, we've been taught that peculiar repetitions might matter. That expectation changes the possibilities we consider.
The same information can therefore support quite different interpretations depending on the framework through which we're viewing it.
My initial suspicion about the hospital food created a new expectation. Subsequent references seemed to fit the theory, while other medical details began to acquire potential significance. The injections were no longer simply treatments. Comas were no longer simply medical conditions. Even changes in the programme's atmosphere became potentially relevant.
I eventually developed a theory that Cooper hadn't properly recovered from being shot at the end of Season One. Perhaps much of Season Two was taking place within his altered consciousness, with fragments of conversations, hospital routines and television programmes being incorporated into an elaborate dream.
The theory had a certain appeal. It explained the sudden increase in surrealism and offered a naturalistic alternative to the supernatural mythology that was rapidly developing. It also accounted for why the hospital seemed to loom so large in my interpretation.
Unfortunately, it had problems. The programme had included dreams and inexplicable experiences before Cooper was shot. We continued watching events involving people and circumstances beyond his apparent knowledge. And the theory was sufficiently flexible that almost anything could be explained as another product of his imagination.
The more I tried to make it fit, the more I risked turning an explanation into a device for avoiding contrary evidence.
Confirmation bias is more complicated than simply wanting to be right
Confirmation bias is frequently discussed as though it were nothing more than a stubborn refusal to acknowledge facts. In reality, the processes involved are considerably more varied.
Nickerson (1998) reviewed evidence showing that expectations and existing beliefs can influence how people seek, remember and interpret information. We may notice evidence consistent with an idea more readily, give it greater weight or evaluate competing explanations using different standards. None of this necessarily requires deliberate dishonesty or an especially strong emotional investment in being correct.
My hospital-food theory illustrates a comparatively harmless version of the problem. Once I'd identified food as potentially significant, further references became unusually memorable. I was less interested in how frequently the programme mentioned other mundane subjects, or whether the medical events I'd connected to the theory were meaningfully associated with meals.
I had identified a pattern without establishing how unusual that pattern actually was.
There's another important complication. Researchers Joshua Klayman and Young-Won Ha (1987) argued that some behaviour traditionally described as confirmation bias is better understood as a positive test strategy. People often investigate a hypothesis by examining cases where they expect its proposed property to be present.
This isn't automatically irrational. Under appropriate conditions, positive testing can be a useful and efficient way of acquiring information. The difficulty arises when the strategy is poorly suited to the problem, particularly when the observations supporting a hypothesis would also be expected under several alternative explanations.
If I suspect that hospital food is causing strange behaviour, noticing another character discussing hospital food isn't necessarily a useful test. Food might be mentioned whether or not anybody is being drugged. A better test would ask whether the people experiencing unusual symptoms had a particular exposure that unaffected characters lacked, or whether there was a consistent relationship between the suspected cause and its effects.
Similarly, my coma theory would become more informative if it generated predictions that distinguished it from an ordinary narrative containing supernatural events. Simply identifying another bizarre scene wouldn't do the job. Both explanations could accommodate bizarre scenes, and one of them could accommodate almost anything.
The important question wasn't whether I could find details that fitted my theory. It was whether those details were more likely to occur if the theory were true than if it were false.
That's considerably harder work than collecting suspicious dinner trays.
Finding a signal in the noise
Signal detection theory provides another useful framework for understanding what happens when we're looking for clues. Originally developed to examine how observers distinguish meaningful signals from background noise, it separates the ability to discriminate between the two from the decision threshold used to declare that a signal has been detected (Stanislaw, 2018).
Imagine being presented with a series of faint sounds and asked whether a particular tone is present. Some judgements will be correct detections, or hits. Others will be false alarms, in which noise is mistaken for a signal. A person who adopts a more permissive threshold may identify more genuine signals, but also produce more false alarms.
Watching a mystery involves a comparable judgement, although narrative interpretation is substantially more complicated than a laboratory detection task.
If we treat every unusual detail as potentially important, we'll probably identify more genuine clues. We'll also spend a considerable amount of time investigating incidental details, jokes, atmospheric flourishes and coincidences. If we're too sceptical, we risk missing the very connections the story has carefully constructed.
And Twin Peaks creates a particularly difficult environment for making that distinction.
An ordinary crime drama generally provides some sense of which information belongs to the investigation. Twin Peaks repeatedly crosses the boundaries between the relevant and the irrelevant, the psychological and the supernatural, the absurd and the meaningful. Sometimes a bizarre experience really does contain information. Sometimes a seemingly meaningful development appears to lead nowhere. Occasionally, a man simply wants to talk about coffee.
There's an important limitation to the analogy, however. A television programme isn't random noise. Writers, directors and editors deliberately choose what to include. Repeated imagery may be intentional even when it has nothing to do with the plot's eventual explanation. A detail can be meaningful as atmosphere, characterisation, comedy or symbolism without being a clue to the mystery.
This distinction became increasingly important during my viewing. Something could be deliberately placed in the programme without supporting the particular theory I'd constructed around it.
The question wasn't simply whether a pattern existed, but what kind of significance, if any, that pattern possessed.
Why our minds connect things that don't belong together
Psychologists sometimes use the term illusory pattern perception to describe the identification of meaningful relationships between stimuli that are random or unrelated. The broader tendency to perceive patterns or connections where none are warranted is often discussed under the term apophenia.
Whitson and Galinsky (2008) conducted six experiments examining how feelings of reduced control affect the perception of illusory patterns. Their findings suggested that experiencing a lack of control can increase people's tendency to perceive patterns in ambiguous information, including visual noise, financial data and descriptions of conspiracies.
Later research by van Prooijen, Douglas and De Inocencio (2018) found associations between illusory pattern perception and beliefs in conspiracy theories or supernatural phenomena. Across several studies, the researchers examined responses to random sequences, ambiguous visual material and information concerning extraordinary claims.
These findings shouldn't be stretched into the claim that anyone who enjoys solving fictional mysteries is engaging in conspiratorial thinking. The studies concern particular psychological processes and, in some cases, beliefs about the real world. A viewer entertaining an unusual explanation for a deliberately mysterious television programme is doing something importantly different from asserting that an unsupported conspiracy actually exists.
Nevertheless, the research highlights a mechanism relevant to both situations: the human capacity to organise ambiguous information into coherent, meaningful structures.
During my viewing, that capacity became increasingly productive. I wondered whether BOB was inhabiting Leland, whether certain victims had been selected as potential hosts, whether the lodges represented opposing supernatural forces and whether the inhabitants of Twin Peaks were pieces in a larger game.
Some of those interpretations connected details the programme would later bring together explicitly. Others depended on assumptions for which I had little evidence.
One particularly questionable theory involved Harold Smith. When Donna first arrived at his house, I immediately decided that we'd found Laura's killer. I subsequently wondered whether he was the grandson of the man Leland remembered from his childhood, potentially resolving an apparent problem with the killer's age.
The initial impression was strong, but the subsequent evidence made Harold considerably less convincing as a suspect. His possession of Laura's diary gave him a connection to her, although access to important information wasn't the same thing as responsibility for her death. As his story developed, I revised my judgement.
The experience was a reminder that a coherent explanation can feel compelling long before it has earned that confidence.
There was also something useful about getting it wrong. The theory changed as new information became available, rather than requiring the programme to conform to my initial impression.
The inconvenient problem of sometimes being right
If this were simply a story about seeing patterns where none existed, it would have a rather unsatisfying ending.
Because I also got something right.
Long before the final episode explicitly identified the Red Room as a waiting room, I'd begun wondering whether it might be a transitional space between the White and Black Lodges. The idea developed from several observations: Laura appeared to have experienced the Red Room before Cooper encountered it; Cooper appeared much older in his original dream; and the room seemed difficult to locate within ordinary chronological or physical reality.
I wondered whether it resembled purgatory, not necessarily in a strictly religious sense, but as a place between different states of existence. It might be somewhere people passed through rather than a final destination.
When the programme eventually called it a waiting room, the connection was satisfying precisely because it emerged from earlier uncertainty.
But what should we conclude from that success?
Certainly not that the method was infallible. I had produced a considerable number of interpretations, and the fact that one was subsequently supported doesn't validate the others. Nor does recognising the intended function of a fictional location establish that all the reasoning leading to the prediction was sound.
Still, dismissing the successful prediction as mere luck would be equally unhelpful.
The Red Room theory drew on recurring narrative features, made a distinction between different possible kinds of space and generated an interpretation that the programme subsequently supported. It was a reasonable hypothesis, even if other explanations remained possible when I first proposed it.
This is the central difficulty with pattern recognition: the cognitive processes that produce false connections are closely related to those that allow us to discover genuine ones.
We can't solve that problem by simply refusing to see patterns. Human reasoning depends on recognising relationships, forming expectations and making inferences from incomplete information. The challenge is learning how to evaluate those inferences without becoming excessively committed to them.
And that requires something more demanding than intuition, even when intuition occasionally turns out to be right.
When the mystery followed me into a dream
Towards the end of Season Two, I had a dream that appeared to provide an explanation for the entire programme.
The two peaks represented opposing powers or realities. The giant and the dancing man seemed to occupy corresponding positions within some larger cosmic structure. The town's inhabitants were pieces being moved in a game they didn't understand. There might be an extraterrestrial or other-dimensional struggle taking place beyond ordinary human perception.
What unsettled me wasn't simply the content of the dream. It was the feeling that the explanation might have come from somewhere else. Perhaps a podcast had been playing in the background. Perhaps I'd encountered the idea previously and forgotten its source. For a moment, the dream felt less like an interpretation I'd constructed and more like information I had somehow acquired.
That experience offers another useful connection to psychological research.
Johnson, Hashtroudi and Lindsay (1993) developed a framework for understanding source monitoring: the processes through which we attribute memories, thoughts and knowledge to their origins. We routinely distinguish between things we've witnessed, imagined, heard from somebody else or inferred ourselves. These judgements generally work well, but they aren't infallible.
The subjective familiarity of an idea doesn't guarantee that we've correctly identified its source. An imagined explanation can feel remembered, while something genuinely encountered may later be mistaken for an original thought.
My dream couldn't establish whether I'd independently arrived at the cosmic-chess theory or encountered some version of it elsewhere. The feeling that I'd somehow been spoiled was interesting, but it wasn't evidence that a spoiler had actually reached me.
It was also difficult to ignore the irony.
I'd spent two seasons watching Dale Cooper take dreams seriously as potential sources of information. Now I'd had my own dream about Twin Peaks and was attempting to determine how much interpretative authority it deserved.
For the record, I didn't begin throwing rocks at suspects. Cooper and I have rather different approaches to fieldwork.
The difference between a good theory and a good story
There is something seductive about a theory that explains everything.
If the hospital food caused the visions, several mysteries might suddenly become connected. If Cooper never recovered from being shot, the programme's changing reality might be explained through altered consciousness. If every character were a pawn in a supernatural conflict, the apparently disconnected storylines might belong to a single elaborate design.
The appeal is understandable. A unified explanation reduces uncertainty and replaces a collection of unrelated events with a coherent structure. Research on conspiracy beliefs has identified the desire to understand and explain one's environment as one of several motivations that can contribute to the appeal of such accounts (Douglas et al., 2017).
But coherence is not the same thing as evidence.
A theory that explains every possible outcome may eventually explain very little, particularly if contradictory information can always be dismissed as another part of the conspiracy. When I began interpreting practically every medical reference as support for the coma hypothesis, I was approaching that problem.
The useful questions were simpler. What would I expect to observe if this explanation were correct? Would I expect the same observations if a competing explanation were correct? What evidence would actually make me reconsider?
Those questions don't guarantee accurate conclusions. They do, however, provide some protection against the tendency to fall in love with our own explanations.
And there is a final complication worth acknowledging. Twin Peaks is fiction. Its mysteries don't necessarily have to operate like empirical problems, and its creators aren't obliged to provide a single consistent account of every event. Some unanswered questions may be deliberate. Others may reflect the practical realities of television production, changing narrative priorities or ideas introduced after earlier episodes were written.
It's possible to approach a fictional mystery intelligently without insisting that its entire world must ultimately yield to one explanatory system.
In fact, some of the programme's most interesting moments may resist that demand.
Simply Put: The Food Was Never The Point
Looking back, I don't particularly mind that my theories about hospital food, widespread drugging and Cooper's coma didn't survive the series.
They were genuine attempts to make sense of information available at the time. Some were playful, others became elaborate, and a few probably deserved to be abandoned considerably earlier than they were. But they reveal something interesting about the relationship between attention, expectation and belief.
The programme gradually changed what I considered worth noticing. Repetition became suspicious. Eccentricity became potentially meaningful. Ordinary details began to acquire significance because they appeared within a world where conventional explanations could no longer be taken for granted.
Sometimes that approach helped. I recognised connections between BOB, Leland and Laura before their full implications became clear. I began considering the Red Room as a transitional space before the programme identified it as such. At other times, I spent rather too long entertaining the possibility that a fictional hospital was conducting an elaborate pharmacological experiment.
The distinction wasn't that one set of ideas emerged from intelligent reasoning and the other from defective thinking. Both involved attention, interpretation, inference and uncertainty.
What mattered was whether I could keep the explanation provisional, consider what might contradict it and revise it when the evidence changed.
There's a temptation, after finishing a mystery, to rewrite our experience so that everything seems obvious in retrospect. We remember the successful predictions, forget the inconvenient ones and quietly reconstruct a version of ourselves that understood the story all along.
Fortunately, I'd been recording my thoughts as I watched. There was no opportunity to pretend I'd never suspected Harold Smith or worried about the hospital food. The evidence against my own retrospective cleverness was sitting there in writing.
And perhaps that's what made the exercise worthwhile.
The interesting thing about a mystery isn't simply whether we arrive at the correct answer. It's how we decide which possibilities deserve our confidence while we don't yet know what that answer is.
Twin Peaks made that process unusually visible. It offered enough genuine connections to reward curiosity and enough ambiguity to expose its weaknesses.
As for the hospital food, I remain open to new evidence.
Preferably something more substantial than another mention of the pudding.
References
Clark, A. (2013). Whatever next? Predictive brains, situated agents, and the future of cognitive science. Behavioral and Brain Sciences, 36(3), 181–204. https://doi.org/10.1017/S0140525X12000477
Douglas, K. M., Sutton, R. M., & Cichocka, A. (2017). The psychology of conspiracy theories. Current Directions in Psychological Science, 26(6), 538–542. https://doi.org/10.1177/0963721417718261
Johnson, M. K., Hashtroudi, S., & Lindsay, D. S. (1993). Source monitoring. Psychological Bulletin, 114(1), 3–28. https://doi.org/10.1037/0033-2909.114.1.3
Klayman, J., & Ha, Y.-W. (1987). Confirmation, disconfirmation, and information in hypothesis testing. Psychological Review, 94(2), 211–228. https://doi.org/10.1037/0033-295X.94.2.211
Nickerson, R. S. (1998). Confirmation bias: A ubiquitous phenomenon in many guises. Review of General Psychology, 2(2), 175–220. https://doi.org/10.1037/1089-2680.2.2.175
Stanislaw, H. (2018). Signal detection theory and its applications. In Oxford bibliographies in psychology. Oxford University Press. https://doi.org/10.1093/obo/9780199828340-0209
van Prooijen, J.-W., Douglas, K. M., & De Inocencio, C. (2018). Connecting the dots: Illusory pattern perception predicts belief in conspiracies and the supernatural. European Journal of Social Psychology, 48(3), 320–335. https://doi.org/10.1002/ejsp.2331
Whitson, J. A., & Galinsky, A. D. (2008). Lacking control increases illusory pattern perception. Science, 322(5898), 115–117. https://doi.org/10.1126/science.1159845
Table of Contents
