Autoshaping resides in the gap between nature and nurture. It has been posited that autoshaping represents interactions between organism and environment, phylogeny and ontogeny, and respondent and instrumental processes. Each of these represents a different level of analysis to the puzzle of autoshaping.
Though the formal study of autoshaping largely began in the late 1960s, the existence of the phenomenon may have been foreshadowed by Darwin's theory of evolution (1859). Darwin posited natural selection as the mechanism whereby species-specific morphogenesis and behavior would need to show environmental adaptation (i.e., functionality) with regard to subsistence and reproduction. In simple terms, as long the new structure or behavior did not impair the animal's relative ability to compete for basic resources, it would continue to exist. Influenced by Darwin's work, William James (1890) similarly implied the existence of autoshaping in discussions of instinct. According to James, an instinct was defined as "the faculty of acting in a such a way as to produce certain ends, without foresight of the ends, and without previous education in the performance" (p. 383). But instincts were not to be considered immutable stimulus-response relations; they were to be considered "blind" to the resultant consequences of the action on the first occurrence of the behavior, after which they could be "disguised" or "modified." Hence, fixed action patterns, an interchangeable term for instinct used by etholo-gists, may be more or less fixed depending upon the effect of the behavior as well as the species under consideration. James implicated the existence of a process whereby innate, hard-wired behavior might interact with and be modified by resultant environmental stimuli.
Lorenz (1957), an early ethologist, posited the more widely held view that due to the simplicity of the nervous system of lower animals, constraints on stimulus perception and response are more likely than in humans and that those responses would be adaptive to the survival of the animal. This view on instinct proposed a mechanism whereby the animal perceived a stimulus that released a species-specific response (e.g., pecking) designed to provide a specific consequence (e.g., food). This paradigm also adhered to the assumption that instinctive responses were unlearned, yet were modifiable, although the modification would only be found in the offspring. Lorenz postulated that the fixed action pattern released by a specific stimulus should be referred to as an instinct; all supporting, orienting, or learned behaviors maintaining or modifying an instinct are to be considered appetitive responses. But, in practice, the line between instinctive and appetitive behaviors remained blurred, perhaps because the etiology of instincts or phylo-genically predisposed fixed action patterns was not well understood.
Better understood are ontogenic models for acquiring behavior within the life of the animal. Two specific forms of learning, classical and operant conditioning, appear relevant to autoshaping. In the aforementioned example with autoshaped pecking in pigeons, it was originally thought that innate aspects of the bird provided for, or predisposed the bird for, rapid shaping via reinforcement of successive approximations of pecking. However, introducing terms like innate aspect or predisposition weakened the scientific explanation, as those terms were not operationally defined, did little to advance the understanding of the data, and were usually tautological (i.e., based on circular reasoning). Brown and Jenkins (1968) were the first to report that non-contingent food presentation temporally contiguous with key illumination resulted in pigeon pecking. Furthermore,
Williams and Williams (1969) conducted the first example of omission training with pigeons, whereby the presentation of food was contingent upon the nonoccurrence of pecking. Under an omission training model, behavior under operant control would cease or become greatly reduced. However, the pigeons continued to exhibit pecking over many trials without food. This study underscored the implausi-bility that autoshaping was maintained by contingent reinforcement with food, even if intermittently or supersti-tiously. This prompted researchers to investigate the possibility that key pecking was classically conditioned.
The rationale for considering classical conditioning as the mechanism of action for autoshaping stems from the fact that within each operant there resides the potential for simultaneous classical conditioning (for in-depth discussion, refer to texts by Davis & Hurwitz, 1977; Honig & Stad-don, 1977; Rachlin, 1976; Schwartz, 1989). Due to the stimulus properties of consequences, particularly primary consequences, neutral stimuli that reliably precede and predict delivery may become conditioned. In other words, rein-forcers and punishers may also serve as unconditioned stimuli-unconditioned response (US-UR) pairs, inadvertently creating conditioned stimuli (CS) and conditioned responses (CR). In the example with pigeons, the food pellet was contingently delivered upon pecking at the key when illuminated. This food pellet, both a potential reinforcer and paired US-UR, might allow the light inside the key to become a CS that elicits a key-pecking response (CR) that closely approximates a normal unconditioned eating response (UR). This model fits the data well, as autoshaped behaviors closely approximate the normal phylogenic response released by the goal stimulus. In a further testing of this model, noncontingent delivery of the food maintained key pecking as long as the illumination preceded and was temporally contiguous to the food delivery, that is, CS continued to evoke the CR when it reliably predicted the US-UR delivery (Brown & Jenkins, 1968). Later, Jenkins (1977) altered the predictability of the CS so that it no longer preceded the food delivery. Classical conditioning extinction curves were noted, as were spontaneous remission curves when contiguity was reestablished. Jenkins also noted that maintenance of the pecking response was best when both contiguity and contingency were in place (i.e., classical and operant conditioning may be additive processes).
In summary, autoshaping appears to be primarily a function of classical conditioning in that underlying US-UR relations are a requisite condition. However, operant consequences may also serve as US-UR pairs, allowing the occurrence of classical conditioning. Autoshaping per se only manifests when operant training appears to be overriding US-UR patterns, or in the terms of James and Lorenz, attempting to modify instinctive fixed action patterns for obtaining goals. Hence, behaviors exhibited during auto-shaping continue to defy simple categorization and precise etiologic explanation.
Brown, P., & Jenkins, H. (1968). Auto-shaping of the pigeon's key peck. Journal of the Experimental Analysis of Behavior, 11, 1-8.
Darwin, C. A. (1859). The origin of species by means of natural selection. London: John Murray.
Davis, H., & Hurwitz, H. M. B. (1977). Operant-Pavlovian interactions. New York: Wiley.
Hergenhahn, B. R., & Olson, M. H. (1997). An introduction to theories of learning (5th ed.). Upper Saddle River, NJ: Prentice Hall.
Honig, W. K., & Staddon, J. E. R. (1977). Handbook of operant behavior. Englewood Cliffs, NJ: Prentice Hall.
James, W. (1890). Principles of psychology (reprint 1990). Birmingham: Smith Peter.
Jenkins, H. (1977). Sensitivity to different response systems to stimulus-reinforcer and response-reinforcer relations. In H. Davis & H. M. B. Hurwitz (Eds.), Operant-Pavlovian interactions (pp. 47-66). New York: Wiley.
Lorenz, K. (1957). Companions in the life of birds. In C. Schiller (Ed.), Instinctive behavior. New York: International Universities Press.
Rachlin, H. (1976). Behavior and learning (pp. 83-128). San Francisco: W. H. Freeman.
Schwartz, B. (1989). Psychology of learning and behavior (3rd ed.). New York: W. W. Norton.
Siegel, R. K. (1978). Stimulus selection and tracking during urination: Autoshaping directed behavior with toilet targets. Journal of Applied Behavior Analysis, 10(2), 255-265.
Williams, D., & Williams, H. (1969). Auto-maintenance in the pigeon: Sustained pecking despite contingent non-reinforcement. Journal of the Experimental Analysis of Behavior, 12, 511-520.
David B. Hatfield Devereux Cleo Wallace, Colorado Springs, CO
See also: Operant Conditioning
Was this article helpful?