Why is the duckling following Konrad Lorenz around the lab? Lorenz is a man, not a duck. His leg is the only thing the duckling has imprinted on. There are four answers, all of them true at once, and confusing them is one of the fastest ways to misunderstand any animal behavior — including human behavior.
The first answer points at what is happening right now. The duckling is following because the leg is moving and the duckling's nervous system is wired to track moving objects of approximately the right scale. Stop the leg and the duckling stops. Move it again and the duckling moves. The trigger is immediate, present, mechanistic.
The second answer points at the duckling's history. The duckling is following Lorenz because there was a critical period in early life — a window of hours after hatching — during which the duckling's nervous system locked onto whatever moving object it saw. Lorenz happened to be in the room. If a duck had been in the room instead, the duckling would have locked onto her. The mechanism that did the locking developed in this duckling over its short life. A different developmental trajectory would have produced different following.
The third answer points at the function. Following the first moving object you see in life keeps you close to your mother in the wild — and a duckling that is not close to its mother gets eaten by predators, freezes overnight, fails to learn what to forage on, and does not become anybody's ancestor. Imprinting is for staying alive. The behavior solves an adaptive problem.
The fourth answer points at deep history. The imprinting mechanism in this duckling exists because, generation after generation across millions of years, ducklings whose nervous systems imprinted on the first moving object they saw left more descendants than ducklings whose nervous systems didn't. The phylogenetic history of duck-following accounts for the existence of the mechanism in the first place. No history, no mechanism. No mechanism, no following.
These are Tinbergen's four whys.1 Nikolaas Tinbergen formulated them in 1951 as the four issues that any complete explanation of an animal behavior has to address. Buss treats Tinbergen's framework as one of the foundational gifts ethology gave to evolutionary psychology, and the framework continues to organize the discipline's methodology.1
The four whys correspond to four kinds of causation, and each one answers a different question about a behavior.
Immediate influences are about what triggers the behavior right now. What stimulus elicited it? What internal state was the organism in? What sensory or hormonal cue tipped the response? When a person flinches at a sudden loud noise, the immediate cause is the air-pressure change reaching the eardrum, the auditory pathway routing it to the amygdala, the amygdala signaling the motor system. This is the proximate mechanism in its narrow sense — the causal chain in the present moment.2
Developmental influences are about what shaped this organism over its lifetime so that this stimulus produces this response. What experiences during ontogeny calibrated the response? Was there a critical period? Did learning play a role, and if so, learning of what? The flinch at a loud noise gets stronger after one experience of being startled by a similar noise; the calibration occurred during this organism's life. The mechanism that does the calibration is itself an adaptation, but the specific tuning of this organism's flinch is developmental.2
Function is about what the behavior is for. What adaptive problem does it solve? How does it contribute to survival or reproduction? Flinching at loud noises functions to interrupt ongoing behavior fast enough to mobilize for a possible threat — predator approach, falling object, weapon use, attack. The function explains why the mechanism exists in the form it does, not how it operates.2
Phylogenetic origins are about deep history. What is the evolutionary sequence that built the mechanism? What ancestor first had it, and what selection pressures shaped it across millions of years? Loud-noise flinching is found in mammals broadly and in many birds and reptiles, suggesting deep-time origins in vertebrate predator-avoidance machinery. Different lineages tuned it differently — fish flinch differently than humans — but the underlying mechanism has a long history.2
Buss flags the same four levels using slightly different language: immediate influences, developmental influences, function, and evolutionary/phylogenetic origins.2 The wording matters because researchers from different disciplines often use only one of the four levels and treat it as the complete explanation. Cognitive scientists default to immediate causation. Developmental psychologists default to developmental influences. Evolutionary psychologists default to function and phylogeny. Each is right about its level. Each is wrong if it claims to be the whole picture.
A separate concept Tinbergen contributed is the fixed action pattern — a stereotyped behavioral sequence triggered by a specific stimulus and run to completion once started.3 Show certain male ducks a plastic facsimile of a female duck and you get a rigid courting sequence. The sequence is not flexible mid-stream; the duck does the whole routine. Fixed action patterns let ethologists carve continuous behavioral streams into discrete units for analysis. Many human behaviors do not run as fixed action patterns — they're more flexible — but the concept gave the field the tools to investigate when and why automaticity shows up.
Lorenz himself anticipated the move from animal ethology to evolutionary psychology when he wrote that "our cognitive and perceptual categories, given to us prior to individual experience, are adapted to the environment for the same reasons that the horse's hoof is suited for the plains before the horse is born, and the fin of a fish is adapted for water before the fish hatches from its egg."4 The line is the bridge from imprinting in ducks to evolved psychological mechanisms in humans. The same four-why framework applies to both.
Take human sexual jealousy as a case where keeping the four whys apart pays off.5
Why is this person feeling jealous right now? At the immediate level: their partner is standing close to someone else, leaning in, eye contact, light touching. Cues match the input pattern the jealousy mechanism is tuned to detect. The amygdala is firing. Cortisol is rising. The person's attention has narrowed. This is the present-tense causal chain.
Why does this person feel jealous in this particular intensity, with this particular response set, when other people in similar situations feel jealous differently? Developmental level: this person's attachment style was calibrated by early caregiver relationships, by previous betrayals or stable trust, by cultural messages absorbed during adolescence about partner loyalty. Different developmental trajectory, different jealousy phenotype. The mechanism is species-typical; the calibration is individual.
Why does the human species have a jealousy mechanism in the first place? Functional level: the mechanism evolved to deter partner infidelity, to deter rival mate-poaching, to maintain pair bonds, and (specifically for men, on the sexual-vs-emotional dimension Buss documents) to combat paternity uncertainty. The function explains why selection built the mechanism.
Why does human jealousy take this form rather than some other form? Phylogenetic level: human jealousy descends from primate mate-guarding mechanisms, with specific human elaborations driven by long-term pair bonding, paternity uncertainty under concealed ovulation, and male provisioning. The deep history explains the architecture.
Each of these is a real cause. None of them is the cause. A clinician treating a jealous person addresses the immediate level (de-escalate the current situation) and the developmental level (work on attachment patterns, communication, trust history). A theorist trying to explain why humans are jealous addresses the functional and phylogenetic levels. The clinician who only knows function and phylogeny offers nothing useful in the moment. The theorist who only knows the immediate situation cannot explain why the response exists at all.
Take a second case. Why are children afraid of the dark?6 At the immediate level, low-light environments increase visual uncertainty and the brain's threat-detection system upweights ambiguous stimuli. At the developmental level, fears of darkness emerge around two to three years of age in coordination with the child's increasing ability to leave the parent's side and the increasing relevance of nocturnal predator and stranger threats. At the functional level, fear of darkness keeps young children near caregivers at night, when they are most vulnerable. At the phylogenetic level, fear of darkness traces to ancestral predator pressures from nocturnal hunters across primate evolution and through human history. A pediatrician handling night terrors addresses immediate (sleep environment) and developmental (age-appropriate reassurance). An evolutionary psychologist explains why the fear exists at all (function and phylogeny). Both are right about different things.
Take a third case to illustrate where conflation breaks the explanation. Why does the man hit his wife when she comes home late? At the immediate level, a specific argument escalated to a specific moment of physical aggression. At the developmental level, the man witnessed similar violence in his own family of origin and never developed the impulse-control machinery to interrupt anger before it reached behavior. At the functional level, intimate partner violence functions as a coercive control mechanism — male sexual jealousy enforcing fidelity through threat — and Buss documents this extensively in the conflict-between-the-sexes chapter.7 At the phylogenetic level, the underlying mate-guarding-through-aggression machinery descends from primate male-coalition aggression and from the specific selection pressures faced by human males under paternity uncertainty.
Now watch what happens when someone collapses the levels. A defender of the man might say "he hit her because of his childhood — it's not his fault." That's the developmental level offered as an excuse for the immediate behavior. A critic might say "he hit her because men are evolutionarily designed to be violent." That's the phylogenetic level offered as an excuse for individual choice. Both are wrong about what each level can do. The developmental cause does not erase moral responsibility for the immediate action. The phylogenetic cause does not determine the individual outcome. The four whys are separate causes operating on different aspects of the behavior, and treating them as one collapses the explanatory frame.
The hardest tension in the four-whys framework is the one most readers of evolutionary psychology trip on: the naturalistic fallacy. Identifying a phylogenetic or functional cause of a behavior does not justify the behavior. "Evolved" is not a synonym for "good" or "right" or "we have to accept it." The four whys are causal levels. The question of what to do with the resulting human equipment is a separate normative question that the framework does not answer.
A second tension runs between the levels themselves. Researchers gravitate toward the level their training equips them for and tend to treat the others as either wrong or trivial. Behaviorists and cognitive scientists for much of the twentieth century treated the immediate level as the only level that mattered. Evolutionary psychologists have sometimes been accused of treating function and phylogeny as the only levels that matter. The corrective is the framework itself: a complete explanation needs all four. Buss endorses this directly while writing primarily about function and phylogeny because those are the levels evolutionary psychology specifically contributes.
A third tension shows up in clinical applications. Patients want explanations that help them change. The proximate levels (immediate cause, developmental cause) feel actionable. The ultimate levels (function, phylogeny) feel fatalistic. A patient with severe sexual jealousy who hears "this is your evolved mate-guarding mechanism" may feel less responsible for the behavior, not more, even though that is not what the explanation entails. Clinicians have to manage this carefully.
A fourth tension is about the boundaries between levels. Where does developmental influence end and functional explanation begin? When a critical period sets up a phenotype that then has fitness consequences, the developmental story and the functional story tangle. Tinbergen's framework keeps them analytically separate, but real causal chains run through all four levels at once.
A fifth tension is about whether the framework applies cleanly to humans. Some critics argue that human behavior has so much developmental and cultural plasticity that the four-why framework, which works well for stereotyped animal behaviors, distorts the human case. Buss's response is that the framework still applies — the levels are about kinds of causation, not about how rigid the behavior is — but the relative weight on each level shifts for humans. Developmental and immediate influences carry more weight in human cases than in pigeon imprinting. Function and phylogeny still operate; they just have to be read against more cultural-developmental noise.
Tinbergen's 1951 framework was a methodological intervention in the ethology of his time, designed to keep mechanism, development, function, and phylogeny from being confused or collapsed into each other. Mayr later sharpened the same distinction into the proximate-vs-ultimate causation pair: proximate causes (immediate mechanism, developmental history) explain how a behavior happens, ultimate causes (function, phylogeny) explain why the behavior exists. Buss adopts both framings.8 Tinbergen's four-level scheme is more granular; Mayr's proximate/ultimate split is the simpler frame most working researchers use day to day. The two are compatible.
Where Buss adds emphasis is on the methodological insistence that evolutionary psychology specifically contributes to the ultimate-cause levels. Other branches of psychology address proximate causes — what stimulus triggered this, how did this person develop. Evolutionary psychology asks why the underlying mechanism exists at all, and what it was selected to do. Buss treats this not as a turf claim but as a corrective to a discipline that, through most of the twentieth century, treated proximate explanations as if they were complete.9
Lorenz, Tinbergen's collaborator and the discoverer of imprinting, had already anticipated the framework's relevance to human cognition. The 1941 quote — about cognitive categories being adapted to the environment "for the same reasons that the horse's hoof is suited for the plains before the horse is born" — is the bridge between animal ethology and evolutionary psychology, and Buss includes it explicitly.4 Lorenz was making the strong claim that human cognition is also imprinted-like, also selected, also has functional and phylogenetic explanations on top of immediate and developmental ones. The four-why framework, originally developed for ducks and sticklebacks, scales up to humans without modification.
Where Buss is silent — and where the vault has work to do — is on the question of how exactly to weight the four levels in any given case. The framework says all four matter. It does not say which matters most when. For some behaviors, the dominant cause is immediate (most reflexes). For others, developmental (most attachment phenotypes). For others, functional (most basic emotions). For others, phylogenetic (most cross-species universals). The work of attribution is left to the case-by-case investigation. Buss demonstrates the work in his own analyses but doesn't formalize the meta-rule.
A behavioral-mechanics operator who can decompose a target's behavior at all four levels at once has more leverage than one who can only see the immediate trigger. The reason is that different intervention points correspond to different levels.
Immediate-level interventions are tactical. Present a cue that the target's mechanism is tuned to read. Smile, touch the arm, mirror posture, lower the voice. The operator is intervening at Tinbergen's level one — directly triggering the mechanism with the input it was built to detect. Hughes's elicitation framework is mostly immediate-level work. Lieberman's fast-pitch is mostly immediate-level work. The interventions work because they hit live mechanisms with cues those mechanisms read automatically.
Developmental-level interventions are slower and require more access. A cult that recruits adolescents and shapes their identity over years is intervening at Tinbergen's level two — calibrating the developmental history that will shape future automatic responses. Religious upbringing intervenes here. Ideological education intervenes here. The operator does not get the immediate response right now; the operator builds the disposition that will produce particular immediate responses years later. Cult recruitment and influence architecture in totalitarian institutions both work at this level.
Functional-level analysis tells the operator which mechanism to target in the first place. If you understand that male sexual jealousy evolved to combat paternity uncertainty, you know that fragments of input suggesting partner infidelity will produce maximal jealousy response. If you understand that disgust evolved as disease-avoidance, you know that contamination cues will produce maximal moral revulsion against an out-group target. The functional level guides which buttons exist to be pressed.
Phylogenetic-level analysis tells the operator which buttons are species-wide and reliable across cultures. If a mechanism is human-universal because it descends from deep mammalian or primate machinery, the operator can deploy techniques that target it across cultural contexts without recalibration. If a mechanism is recent or culturally tuned, the operator has to adapt for context. Buss's cross-cultural data — 37 cultures, 10,047 participants — is functionally a map of which mechanisms are deep and reliable versus which are culturally variable.
The insight neither domain generates alone: behavioral-mechanics catalogs read as bag-of-tricks until each technique is tagged with which level it operates on. Some techniques (immediate-level cues to deep mechanisms) work near-universally and cheaply. Some (developmental shaping) require enormous time and institutional capture but produce durable change. Some (cultural-specific cues to recently calibrated mechanisms) work only in specific contexts. The four-why framework lets the operator choose the right tool for the goal and lets the target understand which kinds of resistance work for which kinds of attack.
A second handshake runs to clinical psychology and the "what should I do about my jealousy/anger/anxiety" question. The four whys give a concrete answer: address each level separately, because they respond to different interventions. The immediate-level response (the activated jealousy in this moment) is calmed by present-moment grounding, recognition of the activation, slowing the chain before behavior. The developmental-level response (why I am especially prone to this) is worked on through therapy, attachment repair, and revisiting the formative experiences. The functional-level response (what the mechanism is for) is recontextualized through education — understanding that the jealousy is doing something the mechanism was built to do helps the patient stop treating it as personal pathology. The phylogenetic-level response (why humans have this at all) is mostly about acceptance — the mechanism is part of being human, and asking it to go away entirely is not the right ask.
This connects to the somatic-trauma theory hub and to identity architecture pages in the vault. Most of those pages work at the immediate and developmental levels — they describe what activates a wounded response, and how the wound was set up. The four-why framework would add functional and phylogenetic context to those pages: this is not just your wound, this is human equipment with a long history operating in modern conditions it was not built for. The insight neither domain generates alone: clinical work and EP both describe the same machine from different sides. Tinbergen's framework is the joint that makes them speak the same language.
The Sharpest Implication.
Most of what people argue about when they argue about human behavior is people unknowingly using different levels of Tinbergen's framework and treating their level as the only level. Whether someone is "responsible" for an angry outburst depends on which level you're addressing. At the immediate level, the outburst was caused by the specific provocation, and the person could in principle have responded differently — yes, responsible. At the developmental level, the person's anger response was calibrated by formative experiences they did not choose — partly responsible, partly shaped. At the functional level, the anger mechanism is doing what it was built for — neutrally not "responsible" in any moral sense, just operating. At the phylogenetic level, anger as a human capacity exists because it was selected for — not the individual's responsibility at all.
A debate between someone holding the immediate level and someone holding the phylogenetic level looks like a moral disagreement and is not. They're describing different parts of the same elephant. The fight is structurally identical for almost any contested human behavior — addiction, infidelity, violence, prejudice, conformity, creativity. People argue about whether the behavior is "free choice" or "wired in" without recognizing that both are partly true at different levels.
Knowing this lets you do something useful in your own thinking. When you find yourself defending or attacking a behavior — your own or someone else's — name which level you're standing on. Then check whether the strongest objection is coming from a different level. Often it is. The disagreement is dissolved when both sides recognize they're addressing different causal questions.
Generative Questions.
If clinical practice mostly addresses immediate and developmental levels, and evolutionary psychology mostly addresses function and phylogeny, what would integrated four-level practice look like? Some traditions — psychodynamic depth analysis, certain contemplative practices — already do this implicitly. Could they do it explicitly?
The four levels carry different felt weights when applied to one's own behavior versus someone else's. Most people read their own anger at the immediate level (the situation made me) and someone else's anger at the developmental or phylogenetic level (they're a violent person, men are violent). What does the framework look like when applied symmetrically?
If a complete explanation of a human behavior requires all four levels, and almost no public discourse achieves all four, what does that say about public discourse? Most political and moral arguments are level-collapsed. Could a culture that learned the four whys argue better, or would it just argue more?