The mid-twentieth-century psychological paradigm was dominated by an unyielding commitment to the general-process view of learning. Within this intellectual milieu, comparative psychologists and radical behaviorists operated under the premise of equipotentiality: the axiom that the laws governing associative learning and instrumental conditioning were fundamentally universal, applying with equal force across all vertebrate species, all sensory modalities, and all arbitrary motor topographies. Whether an animal was required to depress a stainless-steel lever, peck an illuminated plastic key, navigate a spatial maze, or jump across an electrified hurdle, learning was conceptualized as the mechanical stamping-in of stimulus-response connections via reinforcement, or the gradual reduction of primary and secondary drives. The biological identity of the experimental subject was treated largely as an irrelevant parameter, an interchangeable biological substrate whose ecological heritage and evolutionary history could be effectively washed away through standardized laboratory deprivations and habituation regimens.
This theoretical consensus, championed by figures such as Edward Thorndike, B.F. Skinner, and Clark Hull, faced an escalating crisis when confronted with the empirical paradoxes of aversive conditioning. While hungry or thirsty animals readily acquired novel, arbitrary motor habits to obtain nutritive rewards, animals placed in aversive paradigms—specifically avoidance conditioning, where a designated action prevented the delivery of an electric shock—exhibited wild, inexplicable behavioral disparities. Certain avoidance behaviors were mastered with startling, one-trial immediacy, defying the gradual learning curves mandated by reinforcement theory. Conversely, ostensibly trivial instrumental responses, such as pressing a lever to forestall shock, proved nearly impossible for rodents to acquire, even across thousands of training trials. The prevailing frameworks could neither account for why some responses were acquired instantly nor why other, seemingly simple operants generated persistent behavioral breakdowns and catastrophic response failures.
It was within this climate of empirical anomalies that the American comparative psychologist Robert C. Bolles delivered a decisive paradigm shift. In his landmark 1970 paper published in Psychological Review, titled “Species-Specific Defense Reactions and Avoidance Learning,” Bolles bypassed the artificial apparatus of traditional operant reinforcement theory by introducing an evolutionary and ethological architecture. Bolles argued that animals do not enter the laboratory as empty behavioral vessels capable of learning any arbitrary response via negative reinforcement. Instead, when confronted with imminent danger, natural selection dictates that an animal cannot afford the lethal luxury of trial-and-error instrumental learning. Instead, predatory pressures have endowed every species with an innate, pre-organized behavioral repertoire designed to maximize immediate survival: the species-specific defense reactions (SSDRs). Bolles’ groundbreaking insight fundamentally destabilized behavioral orthodoxy, dismantling the artificial boundaries separating European ethology from American comparative psychology, and redefining our contemporary understanding of fear, defense, and the evolutionary architecture of the brain.
1. Historical Context: Traditional Avoidance Learning and Reinforcement Theory
1.1 The Dominance of Operant and Instrumental Conditioning Paradigms
The foundational bedrock of early American learning theory rested on the premise that all behavior could be dissected into discrete, quantifiable associative linkages formed between environmental antecedents and behavioral outputs. This doctrine found its primary early codification in Thorndike’s Law of Effect, which posited that responses accompanied or closely followed by satisfaction to the animal would, other things being equal, be more firmly connected with the situation, such that when the situation recurred, the responses would be more likely to recur. Crucially, Thorndike assumed that the physical form of the response was arbitrary; the organism was viewed as functionally malleable, capable of linking virtually any motor movement within its physiological capacity to any perceptible sensory cue, provided the appropriate satisfying state of affairs ensued.
This mechanistic perspective reached its operational zenith with Skinner’s formulation of radical behaviorism. Skinner relegated internal cognitive and phylogenetic considerations to the conceptual periphery, arguing that behavior should be analyzed strictly through functional relationships established between measurable responses and environmental reinforcement schedules. The operant conditioning paradigm assumed cross-situational malleability: a rat lever-press was fundamentally equivalent to a pigeon key-peck, representing an interchangeable operant unit whose probability of emission was governed entirely by its consequences. Operants were viewed as arbitrary behavioral tokens, selected precisely because they had no pre-existing, hardwired biological meaning in relation to the reinforcement delivered.
Simultaneously, Clark Hull sought to formalize these behavioral observations into an axiomatic, mathematically rigorous deductive system. Central to Hull’s framework was the drive-reduction hypothesis, which posited that learning occurred exclusively when an instrumental response coincided with the reduction of an internal biological need or drive state. In appetitive contexts, food or water consumption reduced primary hunger or thirst drives ($D$). When this logic was extended into the domain of aversive conditioning, electric footshock was conceptualized as a primary aversive drive, while fear or pain was viewed as an internal emotional excitation whose abatement acted as the primary or secondary reinforcing event responsible for habit formation.
This theoretical uniformity dictated a rigid methodological orthodoxy across laboratory research. To ensure rigorous experimental control and eliminate uncontrolled ethological variables, psychologists isolated organisms within standardized test environments. Rodents were placed into Skinner boxes, uniform operant chambers, or simple T-mazes, equipped with automated levers, sensitive contact keys, and stainless-steel grid floors capable of delivering scrambled electrical shock. The apparatus was explicitly designed to divorce the animal from its natural ecology, operating under the unstated assumption that the architecture of learning was entirely neutral to the evolutionary relevance of both the experimental enclosure and the motor demands imposed on the subject.
1.2 Mowrer’s Two-Factor Theory and Its Theoretical Vulnerabilities
As researchers attempted to model avoidance behavior—wherein an organism emits an action during a warning stimulus, thereby preventing an aversive event from occurring—they confronted a profound philosophical and mechanistic dilemma. How could the non-occurrence of an event (the absent electric shock) serve as a causal, reinforcing consequence capable of stamping in an instrumental habit? To resolve this paradox of negative reinforcement, Orval Hobart Mowrer introduced the influential two-factor theory, attempting to synthesize classical Pavlovian conditioning and instrumental operant learning into a cohesive explanatory model.
Mowrer’s formulation posited two distinct and sequentially dependent operational stages. In the first stage, classical conditioning took precedence: the initially neutral warning cue (the conditioned stimulus, or CS, such as an auditory tone or visual light) was repeatedly paired with the noxious unconditioned stimulus (US, an electric shock). Through contiguous temporal pairings, the CS acquired aversive emotional valence, becoming capable of eliciting an unconditioned fear response now conditioned to the warning cue. In the second stage, instrumental conditioning emerged: the organism emitted a behavioral response that terminated the warning CS. Because the CS was now a conditioned fear-evoking stimulus, its immediate cessation produced a sudden reduction in the internal emotional state of fear. This fear reduction functioned as a potent internal reinforcer, mechanically strengthening the preceding instrumental motor habit.
Despite its theoretical elegance, two-factor theory contained deep internal vulnerabilities that empirical investigations repeatedly exposed. Chief among these was the foundational premise that the termination of the warning signal was the strictly necessary reinforcing event for avoidance acquisition. Substantial empirical data began to accumulate demonstrating that animals could successfully acquire active avoidance responses even under experimental protocols where the warning CS did not terminate concurrently with the motor response, or where CS termination was systematically delayed. If fear reduction through signal termination was the sole motor of reinforcement, avoidance learning under such conditions should have been theoretically impossible.
Furthermore, two-factor theory faced a devastating theoretical crisis regarding the non-extinction of avoidance habits. According to Mowrer’s model, once an animal reliably executed an avoidance response, it no longer experienced the electric shock. Consequently, every successful avoidance trial necessarily represented an extinction trial for the classical Pavlovian association: the CS was presented repeatedly without the US. Logically, the conditioned fear response to the CS should have rapidly extinguished over non-reinforced presentations. As fear diminished, the secondary reinforcing capacity of CS termination should have collapsed, causing the instrumental avoidance behavior to unravel. In reality, animals frequently exhibited remarkable, near-permanent resistance to extinction, performing hundreds or thousands of avoidance responses with no signs of behavioral deterioration, long after all behavioral and physiological markers of overt fear had vanished.
1.3 Emergence of Anomalies in Aversive Conditioning Research
By the late 1960s, the empirical cracks in traditional avoidance learning frameworks had expanded into insurmountable chasms. Researchers across diverse laboratories observed dramatic, irreconcilable rate disparities across different response typologies. In theory, if avoidance was governed by general-process instrumental principles, any simple operant that an animal could reliably emit should have been conditionable at roughly comparable rates, adjusting only for minor differences in biomechanical complexity.
The reality was starkly contradictory. If a laboratory rat was placed into an avoidance paradigm where the required response was a simple lever-press—an action routinely mastered within minutes in food-reinforced paradigms—the animal almost universally failed to acquire the avoidance contingency. Even when subjected to hundreds of shocks over dozens of daily sessions, rats rarely developed stable bar-pressing avoidance behavior. Instead, they exhibited behavioral disorganization, profound agitation, or catatonic immobility, often passively enduring the electric shocks despite having the mechanical capability to depress the lever.
In striking contrast, if the experimental paradigm was configured such that the required avoidance response was an active locomotor escape—such as leaping over a hurdle, sprinting down a linear alleyway, or running inside a modified wheel—acquisition was rapid, often approaching asymptotic performance within one to three trials. This extreme disparity could not be explained by motor complexity: depressing a stationary metal bar requires significantly less overall kinetic energy and fewer complex behavioral sub-components than coordinating a full-body ballistic leap over a suspended physical barrier.
These persistent anomalies precipitated intense dissatisfaction within segments of experimental, cognitive, and comparative psychology. Researchers began to argue that the laboratory configurations utilized by strict behaviorists had created a profound illusion of generality by confining animals to artificially impoverished, biologically arbitrary tasks. The realization grew that the theoretical frameworks of Hull, Skinner, and Mowrer had systematically ignored the biological reality of the organism, operating under an erroneous mechanistic dogma that treated the animal’s nervous system as an arbitrary associative switchboard completely detached from evolutionary pressures.
2. Robert Bolles and the Ethological Turn in Learning Psychology
2.1 Bolles’ Academic Trajectory and Theoretical Shift
Robert C. Bolles emerged as a pivotal figure situated directly at the intersection of rigorous quantitative experimental psychology and biological naturalism. Trained in the mid-twentieth-century traditions of experimental behavior analysis and comparative psychology, Bolles initially devoted significant intellectual energy to exploring motivational constructs, formal drive concepts, and dynamic mathematical models of animal behavior. His early work, including his authoritative 1967 treatise Theory of Motivation, engaged deeply with Hullian drive theory, examining how internal physiological deficits translated into overt motor activity.
However, through his exhaustive empirical assessments of motivation and aversive learning, Bolles grew increasingly disillusioned with the mechanistic artificiality of traditional reinforcement paradigms. He recognized that the prevailing drive-reduction and operant models were perpetually forced to invent post-hoc auxiliary assumptions—such as conditioned emotional responses, fractional anticipatory goal responses, or hypothetical internal feedback cues—simply to explain away the persistent empirical failures of avoidance conditioning. Bolles recognized that these theoretical contortions were required because psychologists were viewing animal behavior through an inverted lens: they were attempting to force the animal’s natural behavior into the narrow parameters of artificial laboratory apparatuses, rather than designing experiments that accounted for the organism’s evolutionary history.
This paradigm shift culminated in his 1970 theoretical manifesto, “Species-Specific Defense Reactions and Avoidance Learning,” published in the Psychological Review. This paper marked an intellectual departure from the dominant behaviorist paradigm. Bolles abandoned the premise that avoidance learning was an instance of instrumental reinforcement, proposing instead that avoidance behavior was governed almost exclusively by the activation, elicitation, and hierarchical reorganization of evolutionary adaptations that predated the animal’s introduction to the laboratory by millions of years.
2.2 Integrating Evolutionary Biology with Experimental Psychology
Bolles’ theoretical synthesis was driven by the integration of European ethological principles—pioneered by Nikolaas Tinbergen and Konrad Lorenz—into the experimental frameworks of North American psychology. Ethology had long emphasized that animal behavior could only be meaningfully comprehended within the context of an organism’s natural environment, evolutionary history, and adaptive survival needs. Ethologists recognized that animals possess complex, hardwired behavioral structures, such as fixed action patterns, sign stimuli, and innate releasing mechanisms, which operate autonomously without requiring prior individual reinforcement history.
Bolles realized that when an experimental psychologist placed an albino Norway rat (Rattus norvegicus) onto an electrified grid floor, the animal did not perceive the enclosure as a neutral laboratory apparatus designed for the objective measurement of arbitrary operants. Instead, the sudden immersion in pain, intense light, open space, and novel sensory stimuli activated the ancient, conserved survival mechanisms of a small, nocturnal, prey mammal. The rodent brought its ancestral phylogenetic history directly into the testing chamber. The laboratory box was, in the perceptual and biological reality of the rat, an encounter with a lethal predatory crisis.
Natural selection, Bolles argued, serves as the primary architect of defensive motor programs. In predatory encounters, natural selection operates under asymmetric, unforgiving mathematical parameters: a single error results in biological death. An animal cannot afford the luxury of engaging in slow, trial-and-error instrumental discovery when confronting a predator. The individual that relies on reinforcement to gradually discover that running away reduces pain or fear will inevitably be consumed before any reinforcement contingency can be registered or consolidated within its nervous system. Consequently, evolution had to equip organisms with pre-formed, highly effective defensive behaviors that are elicited rapidly upon the detection of threat, completely invalidating the tabula rasa assumption of operant theory.
2.3 The Concept of Evolutionary Preparedness in Learning
Bolles’ theoretical formulation did not emerge in total isolation; it aligned with a broader, quiet revolution occurring across psychological science that highlighted biological constraints on learning. Concurrently, Martin Seligman was developing his continuum of biological preparedness, arguing that organisms are phylogenetically prepared to learn certain associations vital to their survival, unprepared for neutral associations, and contraprepared against associations that violate their ecological reality.
This perspective was powerfully reinforced by the revolutionary findings of John Garcia and his colleagues in the domain of conditioned taste aversion (the Garcia Effect). Garcia demonstrated that rats could acquire profound, enduring aversions to novel gustatory tastes paired with delayed gastrointestinal illness across a single trial, even when hours intervened between the ingestion and the sickness. Crucially, the animals could not associate the same visceral illness with audiovisual cues, nor could they associate electric cutaneous shock with gustatory cues. This double dissociation dealt an empirical blow to the equipotentiality hypothesis, demonstrating that sensory modalities are hardwired to process specific, biologically relevant environmental contingencies.
Bolles expanded this biological critique into the motor domain. If Garcia proved that sensory input processing was biologically constrained by evolutionary preparedness, Bolles demonstrated that motor output options under aversive conditions were governed by the same evolutionary logic. An operant could no longer be conceptualized as an arbitrary motor token that could be arbitrarily conditioned to any consequence. In aversive contexts, an animal’s available behavioral repertoire was constrained to specialized, phylogenetically selected behavioral survival routines. The scientific objective shifted from formulating universal, content-free laws of learning to mapping how specialized evolutionary adaptations interact with immediate environmental demands.
3. Core Principles of the Species-Specific Defense Reactions (SSDR) Model
3.1 Defining the Species-Specific Defense Reaction Repertoire
At the center of Robert Bolles’ theoretical model was the identification and formalization of what he designated as Species-Specific Defense Reactions (SSDRs). Bolles defined an SSDR as an innate, hardwired, phylogenetically conserved behavioral motor pattern triggered automatically by novel, threatening, or abruptly changing environmental events, as well as by conditioned stimuli that signal impending pain or danger. These responses do not represent newly constructed operant habits learned through reinforcement; rather, they are pre-packaged evolutionary responses embedded within the neurobehavioral architecture of the species.
For small terrestrial mammals, particularly the rodent models pervasive in experimental psychology, the primary, non-negotiable SSDR repertoire consists of three predominant defensive topographies:
- Flight: High-velocity, directional locomotive evasion characterized by rapid sprinting, bounding, and vertical wall-climbing or jumping, designed to place spatial distance between the organism and the localized threat source.
- Freezing: Complete somatic motor arrest, characterized by an immobile, crouched posture in which only obligatory respiratory movements are maintained, accompanied by profound muscle tension, autonomic vigilance, and the absolute suppression of non-defensive exploratory activity.
- Threat and Defensive Aggression: Overt, agonistic behavioral routines typically elicited under extreme spatial proximity to a threat, featuring upright defensive postures, boxing, lateral displays, the baring of incisors, ultrasonic vocalizations, and retaliatory biting.
Bolles emphasized the behavioral inflexibility of these innate responses upon initial confrontation with danger. When an aversive stimulus (such as an intense unconditioned footshock) or a conditioned warning stimulus (a tone or light previously paired with shock) is presented, the animal does not engage in plastic, flexible, exploratory motor actions to discover an arbitrary off-switch. Instead, the organism is seized by an innate defense system. The neurobehavioral architecture instantly deploys an SSDR from its pre-existing repertoire, precluding the immediate emission of arbitrary motor patterns.
The evolutionary rationale underpinning this mechanistic inflexibility is rooted in the mathematical calculus of ecological predation. Trial-and-error learning is biologically unviable under predatory attack. A wild rat hunted by a barn owl (Tyto alba) or a red fox (Vulpes vulpes) does not have the operational leeway to experiment with novel, creative, arbitrary behaviors to escape death. If the animal does not instantaneously freeze to eliminate visual and auditory detection, or execute ballistic flight toward shelter, its genetic lineage terminates. Natural selection has executed the trial-and-error process over evolutionary time, filtering out any organisms that hesitated to deploy innate defensive actions in favor of exploratory behavioral experimentation.
3.2 Hierarchical Organization of Defensive Behaviors
Crucially, Bolles did not characterize the SSDR repertoire as an undifferentiated, random collection of reflexive actions. Instead, he conceptualized defensive behavior as organized within an environmental, dynamic hierarchy. At any given moment, the specific SSDR manifested by an animal is governed by the structural layout and physical affordances of the immediate environment, alongside the spatial and temporal proximity of the perceived threat.
The primary determinant of the behavioral hierarchy is the presence or absence of clear environmental escape affordances. If an animal detects an open exit route, a salient physical opening, a darkened shadow, or an elevated safe ledge within its visual field, flight immediately ascends to the apex of the defensive hierarchy. The animal will execute continuous, high-intensity locomotive bursts oriented directly along these escape trajectories. Under these conditions, flight dominates the behavioral landscape, actively suppressing all alternative defensive choices.
Conversely, if the environment is strictly enclosed, offering no discernible escape corridors, no spatial polarity, and no avenues for physical withdrawal, flight is rendered functionally ineffective. Under these architectural constraints, flight drops down the hierarchy, and freezing ascends to become the dominant SSDR. The animal halts all somatic locomotion, collapses into a low-profile posture against the substrate, and initiates immobility. Freezing remains dominant as long as the threat persists at a distance or remains ambient throughout the environment.
Finally, defensive threat displays and retaliatory aggression occupy the terminal tier of the SSDR hierarchy. This terminal response is deployed when an organism is physically cornered, when escape is impossible, and when the threat closes the distance to make physical contact. When flight has failed and freezing has been rendered obsolete by immediate physical proximity, the animal transitions explosively into agonistic striking, tooth-baring, and biting. Thus, the behavioral output of the threatened animal is an ecologically organized sequence determined by external environmental affordances interacting with internal evolutionary priorities.
3.3 The Mechanism of Response Selection: Reinforcement vs. Elimination
Perhaps the most radical conceptual argument advanced in Bolles’ 1970 paper was his outright rejection of the reinforcement construct in avoidance conditioning. For decades, theorists had debated whether avoidance learning was mediated by the reinforcement of drive reduction (Hull), conditioned fear termination (Mowrer), or cognitive feedback. Bolles bypassed this entire theoretical framework by making a counter-claim: avoidance learning does not involve the instrumental reinforcement of novel behaviors at all.
Traditional reinforcement theory posited that when an animal successfully performed an operant that terminated or prevented an electric shock, an active consequence was delivered: the cessation of aversiveness or the delivery of safety, which acted retroactively upon the nervous system to increase the habit strength or emission probability of that response. Bolles rejected this foundational assumption. He argued that negative reinforcement does not strengthen an operant. Instead, what psychologists mistakenly label “avoidance learning” is fundamentally a process of rapidly eliminating ineffective species-specific defense reactions from the animal’s behavioral hierarchy.
When an animal is introduced to an avoidance paradigm, it immediately deploys its top-ranked SSDR, determined by its evolutionary heritage and the apparatus layout. If that top-ranked SSDR happens to coincide with the arbitrary response selected by the experimenter—such as running out of a shock compartment into a safe compartment—the animal executes the response instantly, and avoidance is “learned” in a single trial. In reality, no new learning has taken place; the experimenter simply selected a task that matched the animal’s pre-existing, top-ranked evolutionary adaptation.
If, however, the top-ranked SSDR fails to terminate or prevent the aversive shock—such as when a rat freezes on an electrified grid floor where running is required, or when running fails to turn off the shock because the experimenter demands a lever-press—the experience of continued pain suppresses and eliminates that specific SSDR from the hierarchy. The animal is forced to drop to the next behavioral alternative in its defensive repertoire. Therefore, learning in an aversive context is entirely subtractive: it is the rapid, successive pruning of unreinforced, ineffective innate defense reactions until an effective behavior is exposed. If the experimenter demands an arbitrary motor topography that does not exist within the animal’s SSDR hierarchy, the animal will cycle continuously through its innate defense reactions, breaking down into freezing, panic, or aggression, never acquiring the required instrumental habit.
4. Methodological Paradigms in Bolles’ Avoidance Conditioning Experiments
4.1 The One-Way Avoidance Apparatus and Design
To provide rigorous empirical substantiation for the SSDR hypothesis, Robert Bolles and his students developed and systematically compared distinct experimental apparatuses that varied the degree of biological naturalism and ecological coherence provided to the subject. The foremost among these was the one-way active avoidance apparatus, a design configured to maximize the spatial and sensory polarity between danger and safety.
The classic one-way avoidance apparatus consisted of an elongated rectangular enclosure divided into two distinct chambers separated by a partition or a drop-gate. The starting compartment, or shock chamber, was structurally defined by unpainted metal walls, bright illumination, and an unyielding stainless-steel grid floor wired to a high-voltage, scrambled shock generator. In contrast, the second chamber—the goal or safe compartment—was constructed with matte black wooden walls, dim illumination or total darkness, and a solid, non-electrified wooden or plastic floor. This structural dichotomy established clear sensory contrast and environmental affordances.
The experimental protocol was straightforward. A rat was placed into the electrified starting chamber. After a predetermined warning interval (the CS, often a buzzer, tone, or the mere visual presence of the compartment doors opening), the grid floor was energized. The animal was required to traverse the spatial threshold from the shock compartment into the safe compartment. Crossing into the safe compartment instantly prevented or terminated the electric shock, and the animal was permitted to remain in the safe environment for an inter-trial rest interval before being manually removed by the experimenter and returned to the starting chamber for the subsequent trial.
The empirical results produced by the one-way active avoidance paradigm were immediate and conclusive. Rather than exhibiting the gradual, protracted learning curves predicted by Thorndike’s Law of Effect or Hull’s mathematical habit-strength equations, rodents exhibited near-instantaneous acquisition. The latency to cross from the shock compartment to the safe compartment dropped from several seconds on the initial trial to fractions of a second by the second or third trial. The trial-to-criterion metric (the number of trials required to achieve nine or ten consecutive avoidance responses) was frequently as low as one to three trials. Bolles noted that this performance did not reflect an incremental associative process; the rats displayed complete, asymptotic avoidance execution almost immediately, because the experimental design aligned with the animal’s primary SSDR: unidirectional, ballistic flight away from danger toward an identified, dark, enclosed sanctuary.
4.2 The Shuttlebox and Two-Way Active Avoidance Paradox
The profound alignment between innate ethology and experimental design observed in the one-way apparatus highlighted the experimental paradox generated by the two-way active avoidance paradigm, commonly referred to as the shuttlebox. In this configuration, the apparatus consisted of two identical compartments separated by a small hurdle or a narrow opening. Both compartments possessed identical physical characteristics: identical grid floors, identical wall colors, and identical lighting.
The fundamental operational contingency of the shuttlebox was that the animal had to move continuously back and forth between the two compartments. On trial one, compartment A delivered the shock, and the rat escaped into compartment B. However, on trial two, compartment B was designated as the shock compartment, and the rat was required to run back into compartment A—the precise location where it had just experienced painful electrical shock moments earlier. The spatial polarity was completely eliminated; every location within the apparatus was simultaneously an area of sanctuary and an area of trauma.
Under these two-way shuttlebox parameters, experimental psychologists observed severe, persistent acquisition deficits. Laboratory rodents exhibited protracted, highly variable learning curves, frequently requiring hundreds of trials to attain modest levels of avoidance performance, with a substantial percentage of subjects failing to master the contingency entirely. Rather than smoothly transitioning from the warned compartment to the alternate chamber, rats placed in the shuttlebox displayed profound behavioral vacillation. They approached the central hurdle, abruptly halted, retreated, engaged in prolonged freezing episodes, or exhibited abortive crossings where they thrust their forepaws into the adjacent chamber only to recoil back into the shock zone.
Bolles’ SSDR model provided a direct explanation for this two-way avoidance paradox. The shuttlebox contingency forced the animal into an intense biological conflict between competing evolutionary defense reactions. While the onset of the warning cue elicited the flight SSDR, that flight was blocked because the only available escape trajectory led directly into a territory marked by contextual fear. In wild rodent ecology, entering an unfamiliar or recently dangerous territory is suppressed by innate freezing and avoidance. The animal was trapped within a behavioral impasse: the urge to flee the current shock chamber was directly counteracted by the innate urge to freeze and avoid entering an environment associated with trauma. The protracted learning curves observed in shuttleboxes were not reflections of slow associative habit formation; they were mathematical artifacts of the animal slowly, painfully overcoming its innate biological resistance to entering a known danger zone.
4.3 Running Wheels and Treadmills as Methodological Baselines
To isolate the locomotive components of flight from the confounding spatial properties of goal boxes and contextual safe chambers, Bolles and his contemporaries utilized automated running wheels and motor-driven treadmills as experimental baselines for aversive conditioning. In the running wheel paradigm, the subject was placed inside a freely rotating, wire-mesh circumference. The operational demand was purely locomotive: upon the presentation of a conditioned warning stimulus, the animal was required to rotate the wheel a predetermined number of degrees or maintain a specific running velocity to forestall the onset of grid shock delivered through the wheel’s perimeter.
The introduction of the running wheel decoupled the defensive motor act from spatial re-orientation. The animal was not required to navigate a spatial labyrinth or choose between ambiguous geographic compartments; it was merely required to run. Under these pure locomotive conditions, the acquisition of active avoidance was practically instantaneous. Upon the very first exposure to shock, the animal engaged in high-velocity sprinting, which directly rotated the wheel and terminated the shock. When the conditioned warning stimulus was presented on subsequent trials, the tone or light acted as an immediate trigger, eliciting running behaviors with low latencies.
Quantitative analyses of velocity, burst frequency, and kinematic acceleration confirmed that shock-elicited and warning-elicited locomotion in a running wheel shared identical structural properties with natural predatory flight. The running was not shaped through gradual instrumental approximations; it was emitted in complete, high-intensity bursts from its initial presentation. This provided confirmation for Bolles’ premise: when an experimental procedure allows the animal to deploy forward locomotion as its defensive operant, avoidance conditioning is essentially effortless, because running constitutes the prioritized, default mammalian defense mechanism against environmental threat.
5. The Triad of Innate Defense: Flight, Freezing, and Aggression
5.1 Flight Dynamics: Locomotion, Jumping, and Escape Topographies
The evolutionary survival of small mammals under predatory threat relies fundamentally on the rapid, coordinated deployment of flight. The ethological kinematics of flight involve far more than simple ambulation; they represent a total, neurobiologically coordinated mobilization of the organism’s musculoskeletal and autonomic systems. Under sudden aversive stimulation, a rodent displays high-velocity burst running, characterized by prolonged strides, rapid ground contact times, and explosive leaps directed toward vertical surfaces or visual openings.
Flight topographies are governed by innate sensory biases, most notably thigmotaxis: the biological tendency of an organism to maintain close physical contact with vertical walls, perimeters, and physical conduits. In natural environments, open spaces represent zones of vulnerability to avian raptors and mammalian carnivores. Therefore, when flight is elicited, rodents do not run into open centers; they sprint along perimeters, dive into shadowed crevices, and attempt to scale vertical boundaries through ballistic jumping. Bolles observed that in laboratory enclosures with open ceilings, rats exposed to shock would frequently leap several times their own body height in an attempt to hook their paws over the chamber walls.
Crucially, the activation of the flight state exerts powerful inhibitory control over all alternative behavioral systems. When the neural circuits governing flight are engaged, appetitive foraging, sexual receptivity, territorial scent marking, and grooming are completely suppressed. This mutual exclusivity ensures that the animal’s energetic resources are directed toward spatial evasion. Flight efficiency is the primary metric of biological survival; a delay of milliseconds in executing directional locomotion can mark the difference between genetic persistence and death.
5.2 Freezing Behavior: Motor Inhibition and Autonomic Vigilance
While flight represents the active kinetic phase of the defensive triad, freezing behavior constitutes its motionless counterpart. Ethologically, freezing is defined with rigorous precision: it is the complete absence of all observable somatic movement, excluding only the shallow, rapid muscular expansions and contractions of the ribcage necessary for respiration. The posture is stereotyped: the animal assumes a crouched, low-profile stance, with its center of gravity pulled close to the substrate, limbs drawn tightly beneath the torso, and vibrissae pinned back against the snout.
Far from representing a state of passive exhaustion or behavioral apathy, freezing is an active, demanding state of motor inhibition paired with heightened autonomic vigilance. Physiologically, the animal undergoes pronounced changes, including sustained bradycardia (a sudden decrease in heart rate), intense peripheral vasoconstriction (which shunts blood away from the skin surface toward the deep skeletal musculature to minimize hemorrhage in the event of an attack), and profound muscle tension. The sensory systems operate in hyper-vigilance: the eyes remain wide, tracking microscopic movements, while the auditory pathways are tuned to detect low-frequency footsteps or raptor wing beats.
The functional role of freezing in natural ecosystems is to prevent visual and auditory detection by movement-sensitive ecological predators. Predators such as raptors, owls, and carnivores possess sensory systems tuned to detect motion. An animal that moves, even slightly, immediately triggers predatory strike sequences. By eliminating motion, the frozen prey organism blends into the visual texture of the substrate, leveraging natural camouflage. Bolles recognized that this profound innate adaptation had been fundamentally misunderstood by decades of behaviorist researchers. In standard experimental literature, immobility on an electrified grid or in an avoidance apparatus was routinely mischaracterized as “passive avoidance learning,” as though the rat had rationally deduced an instrumental rule that staying still prevented punishment. In reality, the animal was gripped by an unlearned, innate SSDR triggered automatically by the presence of inescapable contextual fear.
5.3 Defensive Aggression and Ultrasonic Vocalization
The final tier of the innate mammalian defense triad is defensive aggression. Unlike predatory aggression, which is calculated, quiet, and directed toward consumption, defensive aggression is frantic, loud, and designed to inflict immediate physical pain or distraction upon an attacking predator to facilitate last-ditch escape. Ethologically, this state manifests in upright defensive postures, where the animal rears onto its hindquarters, spreads its forepaws wide to shield its vulnerable ventral abdomen, bares its sharp incisors, and tracks the approaching threat with defensive lunges and retaliatory bites.
This dynamic was brought into clear experimental focus by studies on pain-induced fighting, a phenomenon thoroughly documented by Nathan Azrin and his colleagues in the 1960s. When two laboratory rats were placed together in a small enclosure and subjected to electric footshocks, they did not attempt to cooperate, run in parallel, or freeze side-by-side. Instead, the delivery of shock triggered stereotypic, explosive physical combat between the conspecifics. The animals immediately assumed upright boxing postures, striking, grappling, and biting one another. Azrin initially interpreted this as an unconditioned reflexive response to pain. Bolles incorporated these observations into his SSDR model, recognizing that pain-induced fighting was not an arbitrary reflex, but the direct activation of the terminal SSDR: when an animal is trapped in close physical proximity to an inescapable threat, the innate behavioral program defaults to defensive attack.
Accompanying this physical aggression are specialized non-visual communications, specifically ultrasonic vocalizations (USVs). Threatened rodents emit sustained vocalizations in the 22-kHz frequency range, completely imperceptible to human ears without specialized heterodyne acoustic detectors. These 22-kHz alarm calls, often lasting for hundreds of milliseconds, are produced during sustained freezing and defensive posturing. In wild ecologies, these vocalizations serve a dual evolutionary role: they act as warnings to genetically related conspecifics occupying nearby burrow systems, and they signal to a predator that it has been detected, potentially reducing the predatory advantage of surprise. Aggression and alarm calling represent the emergency boundaries of the SSDR hierarchy, deployed when spatial escape affordances have collapsed.
6. Experimental Discrepancies: Running vs. Bar-Pressing Avoidance
6.1 The Lever-Press Avoidance Failure
The theoretical divergence between traditional operant conditioning and Bolles’ species-specific defense reactions model is illustrated by the lever-press avoidance failure. Throughout the 1950s and 1960s, the operant conditioning chamber—the Skinner box—served as the gold standard for measuring associative learning. Because rats could be trained within minutes to depress a small metal lever to receive a pellet of sucrose or a drop of water, it was assumed that applying an avoidance contingency to the same lever would produce identical learning curves. The animal merely had to depress the lever during a warning tone to prevent an electric shock.
The empirical reality was an experimental failure. When subjected to discrete-trial or free-operant (Sidman) lever-press avoidance schedules, laboratory rats proved almost incapable of mastering the task. Animals subjected to thousands of shock trials across weeks of testing failed to achieve basic criterion levels of avoidance performance. To obtain even marginal lever-press avoidance acquisition, experimenters were forced to resort to complex behavioral shaping regimens: installing massive, protruding response levers that occupied half the chamber, punishing all non-lever behaviors, or introducing artificial wooden platforms that physically constrained the animal within centimeters of the response lever.
The SSDR model exposed the theoretical error underpinning this design. Depressing a delicate, stationary metal lever requires fine motor coordination, deliberate tactile exploration, and focused paw manipulation. However, the presentation of a conditioned fear stimulus or the onset of an electric shock triggers high-arousal SSDRs. The rat’s physiological and behavioral systems are flooded with commands to either run, leap, or freeze. Fine motor coordination and manipulative paw exploration are suppressed. Requiring a terrified rodent to gently press a lever to prevent a shock is ethologically equivalent to asking a human running from an apex predator to thread a needle to unlock an escape door. The motor demands of the experimental operant directly contradicted the innate behavioral repertoire of the organism.
6.2 The Running Response: Rapid Mastery and Low Error Rates
When the arbitrary lever-press was replaced with a response that matched the animal’s natural defensive repertoire, the empirical picture transformed completely. When running served as the required avoidance operant—whether in a one-way runway, an automated running wheel, or a linear escape track—avoidance acquisition occurred rapidly and with low error rates.
Under these locomotive conditions, the gradual, incremental learning curves characteristic of appetitive operant shaping were absent. Rats did not require successive approximations; they did not need to be shaped from orientation to approach, and finally to motor execution. Instead, subjects exhibited step-function transitions to mastery. On trial one, the animal might experience shock and run forward; on trial two or three, the onset of the warning signal elicited a complete, adult, asymptotic running response that carried the animal to safety before shock onset. The trial latency distributions shifted overnight from multi-second delays to sub-second responses.
This empirical divergence provided validation for Bolles’ theoretical thesis. Running did not succeed because it was a mechanically simpler operant for the rat’s skeletal system to execute than pressing a bar; running succeeded because running was the innate flight SSDR. The experimental apparatus simply provided an environmental affordance that allowed the animal’s hardwired defensive repertoire to successfully interface with the experimenter’s programmed contingency. The animal did not learn a new motor habit; it merely deployed an ancient evolutionary survival program that the experimenter had accommodated.
6.3 Response Competition and the Concept of Behavioral Incompatibility
To provide a formal, quantitative account of these experimental discrepancies, Bolles introduced the concept of response competition and behavioral incompatibility. In any aversive conditioning paradigm, the probability of observing a specific target operant ($R_t$) is an inverse function of the strength and frequency of competing species-specific defense reactions ($R_{ssdr}$):
$$P(R_t) = f\left(\frac{E(R_t)}{\sum E(R_{ssdr})}\right)$$
In this dynamic, freezing operates as an active and potent competitor against any required manipulative, stationary, or complex operant. When a rodent is exposed to a warning signal paired with electric shock, freezing behavior is elicited by contextual fear conditioning. Because freezing involves motor inhibition, its execution is physically, kinematically incompatible with any instrumental action requiring spatial displacement or somatic motion, such as walking over to a lever, rearing up, and pressing it down with a forepaw.
This response competition framework explained the phenomenon of conditioned suppression, originally documented by William K. Estes and B.F. Skinner in 1941. When an appetitively trained rat is happily pressing a lever for food, the introduction of a conditioned aversive cue immediately halts all lever-pressing behavior. Traditional behaviorists interpreted this conditioned emotional response (CER) as an internal affective state of fear that non-specifically suppressed operant drive. Bolles demonstrated that the suppression was behavioral and mechanical: the CS elicited freezing, and an animal that is frozen cannot press a lever. The lever-pressing was not suppressed by a loss of appetitive interest; it was displaced by an incompatible SSDR that assumed total dominance over the animal’s motor systems.
7. The Role of Classical Conditioning and Innate Predispositions
7.1 Pavlovian Contingencies as Triggers for Innate Repertoires
A critical dimension of the SSDR model is its reimagining of the functional relationship between Pavlovian conditioning and instrumental behavior. In traditional two-factor theory, classical conditioning was viewed as a passive associative process that endowed an arbitrary neutral stimulus with fear properties, which then served as an internal drive to power instrumental learning. Bolles, conversely, reframed the conditioned stimulus (CS) through an ethological lens: the CS was not merely an occasion-setter or an associative node; it functioned as an innate sign stimulus or a releaser.
When an initially neutral auditory tone or visual cue is repeatedly paired with an aversive unconditioned stimulus (US, such as footshock), classical conditioning does not simply create an abstract cognitive association between two stimuli. Instead, through temporal contiguity and contingency, the CS acquires the biological capacity to tap directly into the organism’s subcortical defense circuits. The CS becomes an ethological surrogate for a predator’s warning signals—such as the snap of a twig, the sudden cast of a shadow, or the low-frequency vibration of an approaching carnivore.
Consequently, the conditioned response (CR) in an aversive paradigm is not an arbitrary learned behavior; its topographical form is heavily constrained by the animal’s phylogenetic history. When the CS is engaged, it releases an innate defensive motor program. If the experimental arrangement is compatible with that program, the CR appears rapidly and is cataloged by the experimenter as successful avoidance learning. If the arrangement is incompatible, the CR manifests as freezing or disoriented flight, disrupting the experimenter’s intended task. The form of the conditioned response is dictated by evolution, not by the experimenter’s arbitrary reinforcement schedule.
7.2 Instinctive Drift and Structural Behavioral Constraints
Bolles’ theoretical formulations directly complemented the landmark discoveries of Keller and Marian Breland, two students of B.F. Skinner who transitioned from academic behavior analysis to commercial animal training. In their 1961 classic, “The Misbehavior of Organisms,” the Brelands documented how dozens of diverse animal species, when trained on standard operant conditioning schedules to perform novel behaviors for food reinforcement, systematically experienced behavioral breakdowns over repeated trials. Pigs trained to deposit wooden coins into a piggy bank began dropping the coins, rooting them along the ground, tossing them into the air, and stepping on them; raccoons trained to drop tokens into a box began persistently rubbing the coins together, dipping them into the box, and pulling them back out in an endless washing routine.
The Brelands coined the term instinctive drift to describe this phenomenon: the progressive intrusion of phylogenetically older, hardwired evolutionary food-procurement routines that gradually overrode, eroded, and replaced the conditioned operant behaviors. No matter how much reinforcement was withheld, the innate biological behaviors could not be suppressed. The animal’s evolutionary instincts drifted into the gap between stimulus and reinforcement, demolishing the operant habit.
Bolles recognized that what the Brelands had documented in the appetitive domain was identical to the operational realities of the aversive domain, albeit amplified by threat urgency. While instinctive drift in appetitive conditioning might take hundreds of trials to slowly erode an operant, in aversive conditioning, the biological constraints operate with immediate, explosive force. From trial one, the animal’s phylogenetic defense systems dominate the behavioral field. Operant learning does not slowly erode into instinct; rather, instinct establishes the structural boundaries within which any behavioral modification can occur. Ethology and behavioral conditioning were reconciled: conditioning is not an alternative to biological evolution, but a specialized evolutionary adaptation designed to fine-tune pre-existing innate repertoires.
7.3 Species Differences in SSDR Manifestations
The concept of species-specific defense reactions is inherently comparative. The term itself—species-specific—demands that defensive repertoires vary across evolutionary taxa, reflecting the unique ecological niches, predatory pressures, and morphological specializations of each organism.
Consider the stark comparative differences between avian and mammalian defense reactions. While a laboratory rat’s primary SSDRs under shock consist of horizontal running, perimeter freezing, and biting, an avian subject, such as a pigeon (Columba livia), exhibits an entirely divergent behavioral architecture under aversive stimulation. An electric shock delivered to a pigeon does not elicit paw manipulation or terrestrial sprint routines; it triggers violent, immediate wing-flapping, vocal squawking, and attempts to gain vertical altitude via flight, followed by high-perch tonic immobility. Consequently, trying to train a pigeon to execute an avoidance response by pecking a small plastic key—the standard operant response in appetitive feeding experiments—is an exercise in experimental frustration. Pigeons cannot easily learn to key-peck to avoid shock; however, they can learn to flap their wings to avoid shock in a single trial, because wing-flapping is their primary, hardwired SSDR.
Even within the class Mammalia, ecological divergence dictates distinct defensive hierarchies. A nocturnal burrowing rodent, such as the Norway rat, relies on thigmotactic perimeter running and diving into subterranean cavities, where freezing in darkness provides high crypsis against visual hunters. Conversely, a diurnal, cursorial mammal adapted to open savannahs or plains—such as an ungulate or an open-country hare—possesses a defensive hierarchy almost exclusively dominated by open-field sprint velocities, zig-zagging maneuvers, and jumping, with somatic freezing suppressed except in neonates. Any universal behavioral law that ignores these evolutionary, ecological, and morphological specializations is scientifically unviable; the animal’s ecological niche is the definitive blueprint that dictates its defensive reactions.
8. Safety Signals and Feedback Stimuli in Avoidance Learning
8.1 The Functional Architecture of the Safety Signal
While Robert Bolles argued that traditional reinforcement theory was fundamentally flawed in asserting that shock omission strengthens avoidance behaviors, he recognized the necessity of explaining how avoidance performance stabilizes and how an animal ceases its panic-induced SSDR activation. To resolve this mechanistic question, Bolles formulated the safety-signal hypothesis, introducing an exteroceptive and proprioceptive sensory architecture into the defensive landscape.
Bolles defined a safety signal as any environmental stimulus, sensory cue, or internal proprioceptive feedback event that reliably and unambiguously correlates with a prolonged period free from danger or shock delivery. In a one-way avoidance runway, the sensory properties of the safe compartment—the dark black walls, the textured solid floor, the spatial distance from the shock grid—act as potent, salient safety signals. The instant the animal traverses the threshold into this safe zone, it is immersed in these sensory cues.
The functional role of the safety signal is neurobehavioral: it acts as a physiological inhibitor of central fear states. When the animal encounters a validated safety signal, the activation of the underlying defense system is attenuated. The safety cue halts the execution of high-arousal SSDRs, terminating the explosive flight bursts and thawing the tonic motor freeze. Upon encountering safe cues, the animal transitions from defensive panic to behavioral stabilization, shifting into low-arousal exploratory sniffing, orientation, and autogrooming. The safety signal stabilizes behavior by turning off the defensive state.
8.2 Feedback Stimuli vs. Shock-Termination Reinforcement
To demonstrate that avoidance acquisition was driven by safety feedback rather than the retroactive reinforcement of shock termination, Bolles and his colleagues designed experiments that decoupled shock-omission from explicit feedback stimuli. In these studies, rats were placed in aversive paradigms where the required response was notoriously difficult to acquire, such as the lever-press or a stationary wheel-turn.
In standard conditions, pressing the lever simply prevented the scheduled shock; there was no change in the auditory or visual environment. Under these conditions, the animals consistently failed to acquire the avoidance response. However, when the experimenters introduced an immediate, salient feedback stimulus—such as an instantaneous auditory click, a brief illumination flash, or the mechanical drop of a protective partition that occurred the microsecond the lever was depressed—the rats acquired the lever-press avoidance response rapidly and reliably.
This empirical demonstration dealt a blow to traditional drive-reduction theory. In both conditions, the objective consequence regarding the primary aversive drive was identical: the electric shock was omitted. If the non-occurrence or termination of shock was the actual reinforcing event, both groups should have learned at identical rates. The dramatic acceleration of learning in the feedback group proved that the immediate sensory stimulus was the critical operational variable. Feedback stimuli do not work by reinforcing an operant in the Hullian sense; they work by serving as immediate, highly conditionable safety signals that terminate fear, quiet the central nervous system’s threat circuits, and suppress the freezing SSDR, releasing the animal to emit subsequent adaptive behaviors.
8.3 Conditioned Inhibition and Extinction Dynamics
Within contemporary learning theory, Bolles’ safety-signal framework merged with Robert Rescorla‘s formalization of conditioned inhibition. A conditioned inhibitor is an environmental cue ($CS^-$) that signals the absence of an unconditioned stimulus that would otherwise be expected based on contextual cues ($CS^+$). Through rigorous summation and retardation-of-acquisition tests, safety signals were demonstrated to function as formal conditioned inhibitors of fear.
This operational insight resolved the puzzle regarding the non-extinction of avoidance habits that had undermined Mowrer’s two-factor theory. Recall Mowrer’s dilemma: if an avoidance response consistently prevents the shock, the animal undergoes classical extinction, fear must collapse, and the avoidance response must unravel. Bolles’ safety-signal model provided a different explanation: the successful execution of the avoidance response consistently exposes the animal to the safety signal ($CS^-$). Because the safety signal acts as a potent conditioned inhibitor of fear, it protects the contextual cues and warning stimuli from undergoing normal extinction dynamics.
Furthermore, the persistent availability of clear, validated safety signals plays a protective psychological role in preventing the onset of learned helplessness—the debilitating syndrome identified by Martin Seligman and Steven Maier in which animals exposed to inescapable shock become incapable of learning subsequent escape tasks. When animals possess behavioral access to reliable safety signals, they maintain stress resilience and autonomic regulation, because the safety cue provides temporal predictability, allowing internal somatic recovery between defensive activations.
9. Critique of Two-Factor Theory and Secondary Reinforcement
9.1 Deconstructing Mowrer’s Secondary Reinforcement Postulate
The foundational core of Mowrer’s classical two-factor theory rested upon the concept of secondary reinforcement: the premise that the internal emotional reduction of fear was functionally equivalent to primary drive reduction, mechanically stamping in the preceding instrumental motor habit. Robert Bolles launched a critique of this postulate, arguing that the theory was logically circular and empirically inaccurate.
The logical circularity was direct: How do we know the animal is afraid of the CS? Because it performs the avoidance response. Why does it perform the avoidance response? Because it is motivated to reduce its fear. This circular reasoning insulated the theory from falsification, allowing theorists to invent hypothetical fluctuations in internal fear states without requiring independent empirical verification.
When researchers began measuring real-time physiological and autonomic correlates of fear—such as blood pressure, heart rate, pupillary dilation, and corticosteroid secretion—the empirical foundation of two-factor theory fractured. Substantial evidence revealed that during the early phases of avoidance training, an animal does indeed exhibit pronounced autonomic fear responses to the warning CS. However, as the avoidance response becomes well-established and asymptotically stable, these autonomic markers of fear disappear. Experienced animals execute avoidance responses smoothly, effortlessly, and with low heart rates, displaying zero behavioral agitation or somatic panic.
This created a paradox for two-factor theory: the avoidance response was operating at its highest, most stable efficiency precisely when fear was minimal. If fear reduction was the sole secondary reinforcer responsible for the execution of the motor habit, the disappearance of fear should have caused the immediate collapse of the behavior. The physiological dissociation between autonomic fear indexes and instrumental avoidance execution revealed that avoidance is not maintained by the ongoing, momentary reduction of felt fear.
9.2 Bolles’ Expectancy Theory of Avoidance Learning
Having dismantled the reinforcement and secondary drive-reduction accounts of aversive behavior, Bolles turned to cognitive psychology to formulate an alternative explanatory model: the expectancy theory of avoidance learning. Deeply influenced by Edward Tolman‘s purposive behaviorism, Bolles argued that animals in learning paradigms do not acquire mechanical, reflexive stimulus-response (S-R) habits. Instead, they acquire cognitive representations of environmental contingencies, which Bolles formalized as expectancies.
In Bolles’ cognitive architecture, avoidance acquisition involves the encoding and integration of two distinct, highly organized propositional structures:
- Stimulus-Stimulus (S-S) Expectancies: The internal cognitive mapping that a specific environmental cue predicts another environmental event. In an avoidance chamber, the animal learns that the warning CS reliably predicts the delivery of the aversive US (Tone $\rightarrow$ Shock). This classical expectancy is acquired rapidly and generates the central threat state that elicits initial SSDRs.
- Response-Stimulus (R-S) Expectancies: The cognitive representation of the consequences of the organism’s own behavioral actions. Specifically, the animal learns that emitting a particular motor response ($R_1$) predicts the non-occurrence of the shock and the immediate onset of the safety signal (Run $\rightarrow$ Safety / No Shock), whereas emitting an alternative response ($R_2$, such as freezing) predicts the continued delivery of shock ($R_2 \rightarrow$ Shock).
This theoretical reformulation replaced automatic reflexology with cognitive-evaluative structures. The animal does not run because a mechanical S-R habit has been stamped into its neural circuitry by a retroactively acting fear reduction. The animal runs because it possesses an explicit cognitive expectancy: it evaluates that running will lead to an environment characterized by safety cues and shock omission. This Tolmanian perspective accounted for behavioral flexibility, the immediate cessation of avoidance upon changes in contingency, and the failure of habits to conform to rigid S-R reinforcement laws.
9.3 The Elimination of the Reinforcement Construct in Defense Paradigms
The logical endpoint of Robert Bolles’ theoretical trajectory was a recommendation for comparative psychology: the complete elimination of the reinforcement construct from defensive behavioral paradigms. Bolles argued that the term “reinforcement” had become an intellectual crutch, a circular label that gave the illusion of explanation without providing mechanistic insight.
In appetitive conditioning, reinforcement might maintain practical descriptive utility: a food pellet reliably increases the rate of bar-pressing in a hungry rat. But in the domain of defense and survival, the invocation of negative reinforcement through shock-termination or fear-reduction was unnecessary. As Bolles demonstrated, an animal’s defensive trajectory through an aversive experiment can be explained by two factors:
- The innate elicitation and dynamic hierarchical reorganization of species-specific defense reactions.
- The cognitive acquisition and updating of S-S and R-S expectancies.
No retroactive stamping-in, no hypothetical drive reductions, and no mechanical habit strengths are required to account for the empirical data.
By abandoning the unnecessary baggage of the reinforcement construct, Bolles achieved theoretical parsimony. Avoidance learning was rescued from the mechanistic contortions of early behaviorism and reunited with the principles of evolutionary biology and cognitive ethology. The animal was restored as an active, cognitive agent equipped with millions of years of evolutionary adaptations, navigating an environment governed by predictable empirical contingencies.
10. Neurobiological Mechanisms Underlying Species-Specific Defense Reactions
10.1 Amygdalar Circuitry and Threat Detection
While Robert Bolles developed the SSDR model primarily through rigorous behavioral and ethological experimentation, modern systems neuroscience has validated his theoretical framework at the cellular, synaptic, and circuit levels. At the core of the brain’s threat-detection architecture is the amygdala, a complex nuclear structure located deep within the medial temporal lobe that coordinates the acquisition of fear expectancies and the deployment of innate defensive repertoires.
The processing of aversive environmental cues initiates within the lateral nucleus of the amygdala (LA). The LA serves as the sensory convergence hub, receiving dual projections regarding conditioned warning stimuli (CS) and unconditioned aversive footshocks (US). Sensory inputs arrive via two routes: a rapid, low-resolution subcortical pathway direct from the sensory thalamus (the “low road,” facilitating near-instantaneous emergency response) and a slower, detailed pathway from the primary sensory cortices (the “high road,” providing complex perceptual evaluation). Long-term potentiation (LTP) occurring at glutamatergic synapses within the lateral amygdala encodes the classical Stimulus-Stimulus (S-S) expectancy identified by Bolles, transforming an initially neutral sensory cue into an innate threat trigger.
Once threat associations are consolidated within the lateral amygdala, excitatory projections are routed through the basal amygdala to the central nucleus of the amygdala (CeA), the primary operational output switchboard for defensive responding. The central amygdala does not project randomly across the brain; its distinct subnuclei orchestrate downstream effectors. The medial division of the central amygdala ($CeA_m$) sends dense, descending axonal projections directly into the brainstem, hypothalamus, and midbrain. In striking validation of Bolles’ separation between threat acquisition (classical conditioning) and motor output (SSDR deployment), neurobiologists have demonstrated that the amygdala does not construct the specific motor topographies of defense; rather, it acts as a top-level trigger, releasing specialized motor programs hardwired within subcortical effectors.
10.2 The Periaqueductal Gray (PAG) and Motor Pattern Selection
The neural structure responsible for executing and switching between the specific motor patterns of Bolles’ SSDR hierarchy is the periaqueductal gray (PAG), an anatomically continuous column of grey matter surrounding the cerebral aqueduct in the midbrain. Modern functional tracing and optogenetic dissections have demonstrated a dissociation between the distinct functional columns of the PAG, mapping onto the SSDR triad:
- Ventrolateral PAG (vlPAG): The neural substrate mediating freezing behavior. Direct descending projections from the medial central amygdala synapse onto inhibitory interneurons and projection neurons within the vlPAG. Activation of this pathway projects down to the rostral ventromedial medulla and premotor pathways in the spinal cord, causing complete somatic motor arrest (freezing), paired with bradycardia, respiratory deceleration, and non-opioid conditioned analgesia. When the vlPAG is optogenetically stimulated, an animal freezes, regardless of environmental context.
- Dorsolateral and Lateral PAG (dlPAG/lPAG): The neural substrate mediating flight and explosive escape. Ascending projections from spinal nociceptive pathways and descending inputs from the superior colliculus and basal amygdala synapse within the dlPAG. Excitation of this region triggers ballistic running, leaping, high-velocity locomotive sprinting, tachycardia, hypertension, and explosive autonomic mobilization.
The PAG operates as the brain’s innate neural substrate for the SSDR hierarchy. The transition between the columns is mediated by biochemical and neurochemical modulation. High concentrations of endogenous opioids, substance P, and corticotropin-releasing factor (CRF) within the midbrain act as chemical switches that bias the system toward either freezing (vlPAG) or flight (dlPAG). Under the influence of imminent proximal threat or physical pain, the dlPAG suppresses vlPAG activity via collateral inhibition, switching the animal’s motor output from motionless freezing to explosive panic flight.
10.3 Prefrontal-Striatal Circuits in Avoidance Control
While the amygdala-PAG axis governs the reflexive elicitation of hardwired SSDRs, the execution of learned, instrumental avoidance and the integration of safety signals requires the recruitment of higher-order cortical and basal ganglia loops. This transition from brainstem defense to adaptive behavioral control is mediated by the medial prefrontal cortex (mPFC), specifically the infralimbic (IL) and prelimbic (PL) cortices, alongside the dorsal and ventral striatum.
The prelimbic cortex (PL) is critically involved in the expression of conditioned fear, sending excitatory projections to the basolateral amygdala to sustain defensive output. Conversely, the infralimbic cortex (IL) is essential for the suppression of fear and the consolidation of safety signal associations. The IL projects directly to the intercalated (ITC) cell masses within the amygdala—a specialized meshwork of GABAergic inhibitory interneurons that gate transmission between the lateral and central amygdalar nuclei. When a safety signal is recognized, the IL fires, exciting the ITC neurons, which in turn silence the central amygdala, terminating the activation of the downstream PAG and thawing the freezing response.
Concurrently, the acquisition of Bolles’ Response-Stimulus (R-S) expectancies and the successful execution of active avoidance tasks relies on the dorsal striatum and the nucleus accumbens. The successful avoidance of shock and the subsequent presentation of a validated safety signal triggers an immediate, transient burst of dopamine release from the ventral tegmental area (VTA) into the nucleus accumbens. This dopaminergic signaling does not represent simple hedonic reward; rather, it functions as an aversive prediction error signal, encoding the unexpected non-occurrence of pain and stabilizing the neural representations that suppress panic. Modern systems neuroscience demonstrates that Bolles’ cognitive expectancies and SSDR hierarchies are physical realities written into the prefrontal-striatal-amygdalar-midbrain architecture of the mammalian brain.
11. Comparative Ethology and Cross-Species Generalizability of SSDRs
11.1 Predator-Prey Interactions in Natural Ecosystems
The definitive validation of Robert Bolles’ species-specific defense reactions model required venturing beyond the artificial confines of university laboratories and evaluating defensive dynamics within wild ecosystems. Field studies of predator-prey dynamics—tracking wild rodents interacting with natural predators such as owls, raptors, snakes, and small carnivores—have provided empirical support for the ecological reality of the flight-freeze-fight hierarchy.
In wild ecologies, the selection pressures operating on defensive speed are measured in single-digit milliseconds. An owl diving from a tree canopy relies on acoustic silent flight and surprise; if a foraging rodent detects the acoustic rustle of feathers and does not deploy freezing within a fractional window, it is captured. Conversely, if a snake approaches a burrow system, freezing offers no protection against thermal sensing and chemoreception; under these ecological constraints, the rodent deploys high-velocity leaping and kicking of sand or debris—specialized, highly localized SSDRs tailored to defeat serpentine predatory techniques.
Importantly, comparative research has highlighted profound differences between wild-type animals and domesticated laboratory strains. Decades of selective breeding within sterile laboratory cages have inadvertently produced behavioral dampening in standard laboratory rats (such as the Sprague-Dawley or Wistar strains). While domestic rats retain the basic architecture of the SSDR triad, their activation thresholds are substantially elevated, their freezing durations are attenuated, and their flight topographies are less explosive than those of wild Rattus norvegicus. When wild rats are introduced into experimental avoidance apparatuses, their SSDRs are expressed with primal, explosive intensity: they scale smooth vertical walls, execute immediate attacks against human experimenters, and demonstrate acquisition of one-way avoidance tasks in a single trial, demonstrating that the laboratory anomalies documented by Bolles were softened reflections of raw evolutionary survival programs.
11.2 Avian and Piscine Defensive Adaptations
The architectural principles of Bolles’ SSDR model are not restricted to terrestrial mammals; they represent universal survival design patterns running across diverse phylogenetic lineages. Consider the defensive adaptations of teleost fish, which occupy a three-dimensional aquatic environment characterized by hydrodynamic pressure waves and rapid aquatic predators.
In fish, the primary, non-negotiable SSDR against predatory attack is the C-start escape response. Mediated by a pair of giant reticulospinal neurons in the hindbrain known as the Mauthner cells, the C-start is a ballistic, high-velocity escape reflex initiated within five to ten milliseconds of a localized hydrodynamic disturbance. The fish’s body contracts into a tight “C” shape, followed by an explosive contralateral tail-beat that propels the animal away from the threat vector. Attempting to condition a fish to execute an arbitrary operant (such as gently nudging a target ring with its snout) to avoid an electrical pulse is notoriously difficult; however, when the required avoidance operant matches the Mauthner-cell-mediated C-start burst, acquisition is instantaneous.
Similarly, across avian species, the phenomenon of tonic immobility (death feigning or “thanatosis”) represents an extreme, terminal SSDR. When an avian subject is physically caught by a predator and flight is completely foreclosed, the animal enters a catatonic state characterized by waxy muscular flexibility, unresponsiveness to painful stimuli, and profound motor arrest. In wild ecologies, many terrestrial predators possess strong motor programs to kill moving prey, but will temporarily relax their grip or set down an animal that appears completely dead to manage other prey or survey the environment. The moment the predator looks away, the bird breaks its tonic immobility and launches into ballistic flight. Across fish, reptiles, birds, and mammals, the SSDR concept holds: defensive behaviors are phylogenetically pre-packaged adaptations tuned to the ecological morphology of the species.
11.3 Primate and Human Manifestations of SSDRs
The evolutionary continuity of species-specific defense reactions extends directly into higher primates, including human beings. While human beings possess an expansive neocortex capable of abstract prospective reasoning and complex technological problem-solving, under conditions of sudden, proximal physical threat, our behavioral architecture collapses into ancient, conserved subcortical defense programs.
In non-human primates, SSDRs are visible in complex alarm-calling systems and collective antipredator routines. Vervet monkeys (Chlorocebus pygerythrus) possess distinct, acoustically differentiated alarm vocalizations that elicit specific, hardwired defensive motor patterns: a leopard alarm causes monkeys to immediately run up into the thinnest branches of trees where heavy carnivores cannot follow; an eagle alarm causes monkeys to look up and dive into dense low-lying brush; and a snake alarm causes monkeys to stand upright on two hind legs and scan the surrounding grass. These responses are not arbitrary instrumental behaviors shaped by individual reinforcement; they are prepared, species-specific evolutionary routines.
In humans, the manifestations of SSDRs are evident in the clinical phenomenology of panic, acute trauma, and phobic avoidance. When confronted with an immediate, terrifying crisis—such as a mass shooter, a natural disaster, or a sudden violent assault—humans rarely engage in deliberate, rational, cognitive evaluation. Instead, the nervous system undergoes somatic freezing (tonic motor inhibition, hyper-vigilance, inability to vocalize), followed by frantic, disoriented flight (stampeding along perimeters toward perceived exits), or defensive combat. Furthermore, human phobias are distributed along lines of evolutionary preparedness: humans readily develop debilitating, non-extinguishing phobias toward ancestral threats (spiders, snakes, heights, enclosed spaces, darkness) with minimal or zero conditioning, while rarely developing phobias toward modern lethal objects (electrical outlets, automobiles, firearms) despite far higher empirical injury rates. Bolles’ SSDR framework provides the foundational ethological model for understanding human survival behaviors and psychiatric anxiety disorders.
12. Contemporary Implications and Legacy of Bolles’ Theoretical Framework
12.1 Michael Fanselow’s Predatory Imminence Continuum Theory
The most direct, influential theoretical descendant of Robert Bolles’ SSDR model is the predatory imminence continuum (PIC) theory, formulated and developed by Bolles’ former student, Michael S. Fanselow. Fanselow recognized the brilliance of Bolles’ model but sought to expand its predictive power by providing a quantitative, ecologically grounded metric that could mathematically predict which specific SSDR an animal would display at any given second.
The predatory imminence continuum posits that an animal’s defensive behavioral repertoire is dynamically determined by the psychological, physical, and temporal distance separating the organism from a predator. Fanselow segmented the defensive continuum into three distinct operational phases:
- Pre-Encounter Defense: The threat is non-localized, abstract, or ambient. The animal is in a territory where predators have historically been encountered, but no specific threat has been detected. The behavioral output is characterized by increased vigilance, cautious thigmotactic exploration, meal-pattern reorganization (consuming food in rapid, high-volume bursts to minimize exposure time), and the exploitation of protective shelters.
- Post-Encounter Defense: A specific predator or validated threat cue has been detected within the perceptual field, but direct physical contact has not yet occurred. The predator may not have detected the prey. In this zone, the dominant behavioral topography is freezing. The animal becomes somatically motionless to eliminate visual and acoustic detection, accompanied by profound bradycardia and opioid-mediated conditioned analgesia.
- Circa-Strike Defense: The predator has detected the prey, closed the physical distance, and initiated physical attack or contact. Freezing is immediately abandoned, as crypsis is now obsolete. The animal transitions into explosive, non-directional panic flight, ballistic jumping, high-pitch vocalizations, and retaliatory biting and defensive aggression designed to inflict injury or create a momentary window for physical escape.
The predatory imminence continuum represents the modern mathematical and ethological evolution of Bolles’ hierarchical SSDR model. It has become a foundational architecture in modern behavioral neuroscience, providing a quantitative laboratory framework that links precise neurobiological structures (prefrontal cortex for pre-encounter; amygdala and vlPAG for post-encounter; dlPAG and superior colliculus for circa-strike) to real-world ecological survival behaviors.
12.2 Impact on Clinical Models of Anxiety, Trauma, and PTSD
Robert Bolles’ deconstruction of avoidance learning and his formulation of the SSDR and safety-signal hypotheses have transformed clinical psychiatry and experimental psychopathology. Contemporary models of Post-Traumatic Stress Disorder (PTSD), panic disorder, and specific phobias understand these conditions not as irrational, broken cognitive thoughts, but as evolutionary defense reactions locked in maladaptive, chronic operational states.
In patients suffering from PTSD, the neurobiological threat-detection circuitry is held within a permanent circa-strike or post-encounter mode. Traumatic triggers (visual sights, sudden acoustic bangs, olfactory cues reminiscent of trauma) bypass conscious cortical appraisal, descending directly into the amygdala and brainstem to unleash somatic SSDRs: motor freezing (tonic immobility, dissociative detachment, emotional numbing), frantic flight (sudden agoraphobic panic, running from public spaces), or defensive rage and combativeness. The persistent physical symptoms of PTSD—hyper-arousal, resting tachycardia, startle hyper-reflexia—are the autonomic correlates of an SSDR repertoire trapped in continuous activation.
Furthermore, Bolles’ safety-signal framework explains the persistent failure of traditional cognitive and extinction therapies. In disorders such as Obsessive-Compulsive Disorder (OCD) and severe panic disorder, patients engage in relentless safety-seeking behaviors (checking locks, hand washing, carrying medication, avoiding open spaces). Clinicians historically viewed these behaviors through standard operant reinforcement models. Through Bolles’ framework, we understand that these safety-seeking behaviors function as formal conditioned inhibitors ($CS^-$). By deploying these safety behaviors, the patient terminates fear in the short term, but prevents the underlying conditioned fear cues from undergoing authentic extinction. Therapeutic innovations, such as Exposure and Response Prevention (ERP), work precisely by eliminating these safety signals, forcing the patient to remain in the presence of conditioned cues until the central threat state collapses and the nervous system can update its expectancies.
12.3 The Paradigm Shift: From Strict Behaviorism to Evolutionary Behavioral Science
The enduring legacy of Robert C. Bolles lies in his central role in dismantling the artificial, mechanistic dogmas of strict mid-twentieth-century behaviorism. By demonstrating that the laws of learning are subordinated to the laws of natural selection, Bolles exposed the fundamental error of the general-process view of learning. He shattered the myth of the interchangeable organism, proving that the evolutionary history of the subject is an active variable present in every psychological experiment.
Bolles’ 1970 paper in Psychological Review served as an intellectual catalyst that tore down the ideological wall separating American experimental psychology from European ethology. He helped birth contemporary evolutionary psychology, cognitive ethology, and behavioral neuroscience. Today, no neuroscientist or psychologist designs an aversive experiment, interprets an avoidance curve, or investigates a fear circuit without acknowledging that the animal’s motor capabilities, sensory biases, and learning trajectories are framed by its species-specific defense reactions.
In retrospect, Bolles did not merely explain why rats could not press levers to avoid electric shocks; he forced the psychological sciences to recognize that organisms are not passive, malleable associative machines awaiting environmental reinforcement. Animals are evolutionary actors, engineered through millennia of predatory struggle, carrying within their nervous systems the behavioral blueprints of their ancestors. Learning is not the mechanical stamping-in of arbitrary consequences; it is the adaptive fine-tuning of an already sophisticated, phylogenetically conserved survival repertoire. In reconnecting learning theory with evolutionary biology, Robert Bolles redefined our understanding of animal and human behavior.
Conclusion
The scientific trajectory charted by Robert C. Bolles represents an intellectual turning point in the history of psychology and behavioral neuroscience. By interrogating the empirical failures of traditional avoidance conditioning—anomalies that the prevailing operant, instrumental, and drive-reduction frameworks of Thorndike, Skinner, Hull, and Mowrer could neither assimilate nor resolve—Bolles exposed the profound hazards of viewing animal behavior through the artificial lens of mechanistic equipotentiality. His realization that laboratory organisms do not enter testing enclosures as tabula rasa entities, but as biologically structured products of relentless natural selection, fundamentally altered the theoretical architecture of comparative psychology.
The Species-Specific Defense Reactions (SSDR) model successfully replaced the artificial construct of negative reinforcement with an ethological and cognitive framework. In demonstrating that aversive learning is driven by the dynamic activation, hierarchical deployment, and rapid elimination of hardwired defense reactions (flight, freezing, and defensive threat) paired with the cognitive encoding of stimulus-stimulus and response-stimulus expectancies, Bolles provided a parsimonious, predictive, and biologically grounded account of survival behavior. His insights solved long-standing paradoxes: the rapid mastery of one-way avoidance, the agonizing acquisition deficits of the two-way shuttlebox, the intractable failure of lever-press avoidance schedules, and the non-extinction of stable defensive behaviors.
More than half a century after the publication of his seminal 1970 paper, Bolles’ theoretical architecture continues to guide empirical discovery. His formulations laid the direct foundation for Michael Fanselow’s predatory imminence continuum, provided modern systems neuroscience with the conceptual map required to untangle the amygdalar-periaqueductal gray defense networks, and revolutionized clinical psychiatry’s approach to trauma, panic, and anxiety disorders. Ultimately, Robert Bolles’ enduring monument is the complete reunification of learning theory with evolutionary biology. He proved that to understand how an organism learns to survive tomorrow, science must first understand how its ancestors survived the millions of years that came before.
References
- Azrin, N. H., Ulrich, R. R., Hutchinson, R. R., & Norman, D. G. (1964). Effect of shock duration on shock-induced fighting. Journal of the Experimental Analysis of Behavior, 7(1), 9-11. https://doi.org/10.1901/jeab.1964.7-9
- Bolles, R. C. (1967). Theory of motivation. Harper & Row.
- Bolles, R. C. (1970). Species-specific defense reactions and avoidance learning. Psychological Review, 77(1), 32-48. https://doi.org/10.1037/h0028589
- Bolles, R. C. (1971). Species-specific defense reactions. In F. R. Brush (Ed.), Aversive conditioning and learning (pp. 183-233). Academic Press.
- Bolles, R. C. (1972). Reinforcement, expectancy, and learning. Psychological Review, 79(5), 394-409. https://doi.org/10.1037/h0033120
- Bolles, R. C., & Collier, A. C. (1976). The effect of predictive cues on freezing in rats. Animal Learning & Behavior, 4(1), 6-8. https://doi.org/10.3758/BF03211975
- Bolles, R. C., & Fanselow, M. S. (1980). A perceptual-defensive-recuperative model of fear and pain. Behavioral and Brain Sciences, 3(2), 291-301. https://doi.org/10.1017/S0140525X0000491X
- Breland, K., & Breland, M. (1961). The misbehavior of organisms. American Psychologist, 16(11), 681-684. https://doi.org/10.1037/h0040090
- Estes, W. K., & Skinner, B. F. (1941). Some quantitative properties of anxiety. Journal of Experimental Psychology, 29(5), 390-400. https://doi.org/10.1037/h0062283
- Fanselow, M. S. (1986). Associative vs. topographical accounts of the immediate shock-freezing deficit in rats: Implications for the behavioristic concept of habit. Quarterly Journal of Experimental Psychology Section B, 38(1), 15-39.
- Fanselow, M. S. (1994). Neural organization of the defensive behavior system responsible for fear. Psychonomic Bulletin & Review, 1(4), 429-438. https://doi.org/10.3758/BF03210947
- Fanselow, M. S., & Lester, L. S. (1988). A functional behavioristic approach to aversively motivated behavior: Predatory imminence as a determinant of the topography of defensive behavior. In R. C. Bolles & M. D. Beecher (Eds.), Evolution and learning (pp. 185-212). Lawrence Erlbaum Associates.
- Garcia, J., & Koelling, R. A. (1966). Relation of cue to consequence in avoidance learning. Psychonomic Science, 4(3), 123-124. https://doi.org/10.3758/BF03342209
- Hull, C. L. (1943). Principles of behavior: An introduction to behavior theory. Appleton-Century-Crofts.
- LeDoux, J. E. (2000). Emotion circuits in the brain. Annual Review of Neuroscience, 23(1), 155-184. https://doi.org/10.1146/annurev.neuro.23.1.155
- Lorenz, K. (1937). The companion in the bird’s world. The Auk, 54(3), 245-273. https://doi.org/10.2307/4078077
- Maier, S. F., & Seligman, M. E. (1976). Learned helplessness: Theory and evidence. Journal of Experimental Psychology: General, 105(1), 3-46. https://doi.org/10.1037/0096-3445.105.1.3
- Mowrer, O. H. (1947). On the dual nature of learning—a re-interpretation of “conditioning” and “problem-solving”. Harvard Educational Review, 17(2), 102-148.
- Mowrer, O. H. (1960). Learning theory and behavior. John Wiley & Sons. https://doi.org/10.1037/10802-000
- Rescorla, R. A. (1969). Conditioned inhibition of fear resulting from negative CS-US contingencies. Journal of Comparative and Physiological Psychology, 67(4), 504-509. https://doi.org/10.1037/h0027299
- Rescorla, R. A., & Solomon, R. L. (1967). Two-process learning theory: Relationships between Pavlovian conditioning and instrumental learning. Psychological Review, 74(3), 151-182. https://doi.org/10.1037/h0024475
- Seligman, M. E. (1970). On the generality of the laws of learning. Psychological Review, 77(5), 406-418. https://doi.org/10.1037/h0029790
- Sidman, M. (1953). Avoidance conditioning with brief shock and no exteroceptive warning signal. Science, 118(3058), 157-158. https://doi.org/10.1126/science.118.3058.157
- Skinner, B. F. (1938). The behavior of organisms: An experimental analysis. Appleton-Century-Crofts.
- Thorndike, E. L. (1911). Animal intelligence: Experimental studies. Macmillan. https://doi.org/10.5962/bhl.title.55243
- Tinbergen, N. (1951). The study of instinct. Oxford University Press.
- Tolman, E. C. (1932). Purposive behavior in animals and men. Century Company.
- Tovote, P., Fadok, J. P., & Lüthi, A. (2015). Neuronal circuits for fear and anxiety. Nature Reviews Neuroscience, 16(6), 317-331. https://doi.org/10.1038/nrn3945