A visitor to a zoological garden witnesses chimpanzee behavior bearing striking resemblance to human behavior. The ape's facial expressions, manual gestures, and howls of displeasure as it wraps its head in its arms remind the viewer of our own efforts to make our feelings understood by others around us. These similar behavioral patterns have led many human observers to speculate that the chimpanzees' behavior and their own share important common factors.
Young children also convey feelings to parents and caregivers in nonverbal ways, for example, with tears or sobs. Yet, unlike chimpanzees and other nonhuman species, most children somehow seem to learn to identify their feelings and to report them to others through natural socialization processes. The present target article is about the nature of this process and the mechanisms responsible for creating individual differences in communicative behavior involving one type of private stimulation, emotional or feeling states. We will propose a nonhuman laboratory animal model for this process.
Figure 1 illustrates our animal model: A pair of birds was trained to exchange arbitrary cues, "letters," based on drug-induced state variations in the internal environment of one of them. The drug-cue bird received cocaine, pentobarbitol, or saline and was trained to communicate with another bird by pecking response keys corresponding to these drug-induced state variations. As far as we know, this interaction represents the first laboratory demonstration in a nonhuman species of the exchange of arbitrary communicative cues based on the internal ("affective") state of one of the participants. We believe this model can help us understand how humans exchange information about private stimulation, even when it is novel. It may also shed light on human individual differences in the skill and tendency to engage in this kind of communication. ----------------------------------------------------------------- --------------------------------------- Insert Figure 1 about here ----------------------------------------------------------------- ---------------------------------------
Ours will be an animal model of one way humans may acquire the ability to report on their private stimulation. To better understand it, we will have to review relevant variables from a number of areas that were drawn upon and combined to generate the performances illustrated in Figure 1, including animal models of human psychological phenomena, both natural and artificial animal communication, the status of private events in psychology, and mechanisms of interoceptive stimulus control (with particular emphasis on pharmacological agents). Our second objective is to show that the forms of behavior used to generate our model are also applicable to much more sophisticated behaviors in later evolved species. Our model was based on the more limited response capabilities of birds, subsequent animal models of communication based on private stimulation will undoubtedly display more complex features.
1. Models of Human Psychological Phenomena
The notion that inanimate and animate systems may serve as models for human phenomena is very old (Gunderson, 1985; Keehn, 1986). Leonardo da Vinci (Keele, 1983) drew mechanical devices he believed could model human actions if they were actually built and Rene Descartes (1662) hypothesized that the automata he designed captured key components of human behavior. Both models were inanimate, however. By the middle of the 19th century, Darwin's (1859) theory of evolutionary continuity became a basic assumption of the rapidly developing field of biology. Darwin's work suggested that to gain biological insight bearing into human beings it may be more illuminating to study nonhuman animate systems rather than the inanimate models of da Vinci and Descartes. Claude Bernard (1885), the founder of experimental medicine, used dogs in laboratory preparations as suitable models of human physiology, assuming basic continuity in physiological functions across species. Both Darwin and Bernard argued that anatomy, physiology, and behavior not only look similar in different animals but often share common evolutionary origins and current regulatory mechanisms.
In modeling human psychological phenomena using nonhuman species, one accordingly assumes functional continuity in behavior. Thorndike's (1911) model of learning in chickens and cats, Pavlov's (1927) description of associative conditioning in dogs, and Tolman's early studies on "insight" in rats (Tolman & Honzik, 1930) were all thought to have important implications for human conduct. In the middle of the 20th century, however, biological continuity was questioned within linguistics and some areas of psychology (cf. Miller, 1990). Human behavior (especially communicative behavior) was thought by many to differ qualitatively from the behavior of other animals. Even among some behaviorists, the human tendency to communicate thoughts and feelings was frequently mentioned as beyond the realm of cross-species generalization and laboratory experimentation.
H. S. Terrace (1985, p. 1026), for example, wrote: "The ability to name is also relevant to a basic aspect of human consciousness. As part of our socialization, we learn to refer to various inner states: our feelings, emotions, thoughts, and so on (cf. Skinner, 1945). If one applied to internal states the same distinction one makes between perceiving an external event and naming that event, one is left with a very interesting difference between animal and human consciousness. Human beings are able to name their inner states, animals are not (e.g., Gallup, 1977; Griffin, 1976)."
This target article will explore the possibility that arbitrary, nonspecies-specific communication between organisms based on private events may extend beyond Homo Sapiens and does not require language. We will present evidence that, there are differences within and across species in the disposition to respond discriminatively to and communicate about stimuli that are accessible only to the subject, such as events beneath the skin. This extension of biological continuity to communication based on private events may not only facilitate our understanding of human development but it may also elucidate other psychological phenomena. We will discuss animal models of communicative behavior sharing features of the relationships between "speaker" and "listener." We believe these communicative interactions, which were formed in the laboratory using the experimental tactic of behavioral synthesis (Catania, 1983; Epstein, 1984), accurately model some aspects of human interpersonal communication. 1.1 Behavioral synthesis is the methodology underlying our model. Behavioral constituents are combined into novel combinations that have some of the functional properties of complex human behavior. Laboratory animals are trained to display several independent repertoires that are hypothesized antecedents to a complex human performance. Once established, cues for each are presented in the proper temporal arrangement (e.g., Findley, 1962; Thompson & Lubinski, 1986) to determine whether an integrated pattern resembling human behavior emerges. The final product of a behavioral synthesis is a plausible reconstruction of some aspect of human behavior. Such behavioral interpretations (Day, 1976; Schnaitter, 1978; Skinner, 1969) parallel evidence provided by simulations in other scientific domains, providing a "plausibility proof" (cf. Epstein, 1984; Kordig, 19781) of the variables underlying complex behavioral phenomena especially those that are not directly accessable or cannot be investigated for ethical reasons.
Antecedents to this methodology are found in Kohler's (1925) classic experiments on "insight" learning in chimpanzees, which motivated one of the first contemporary demonstrations of the utility of behavioral synthesis. Kohler suspended a banana out of a chimpanzee's reach and placed a large wooden box in his cage that was sturdy enough to stand on. The chimpanzee solved the problem by pushing the box below the banana and standing on the box to reach it. At the moment when the chimpanzees detected a new instrumentally useful rearrangement of preexisting behaviors (viz., pushing large objects + reaching for food), they were said to have experienced an "insight" analogous to insight learning in humans. In the present context, insight may be interpreted as a behavioral synthesis of two previously independent behaviors that are now components of a more complex behavioral form.
Kohler's finding led Epstein et al. (1981a) to reason that apes could solve the problem without specific training because they had previously learned all the constituent behaviors necessary for this novel performance (reaching for edibles, pushing or moving large objects). The problem of how to reach the fruit provides the opportunity for both constituents to come into close temporal conjunction. Thus, a new form of instrumental behavior emerges as a result of the simultaneous presentation of cues relevant to both repertoires. Indeed, many instrumentally effective behaviors that appear totally novel, or wholly creative, actually stem from recombinations of existing behaviors (cf. Lubinski & Dawis, 1992; Lubinski & Thompson, 1986; Skinner, 1957). Epstein et al. set out to determine whether pigeons, equipped with relevant constituent repertoires could solve the "insight" problem too. They showed that the complex primate behavior observed by Kohler (1925) could be synthesized in the laboratory with pigeons.
Using grain delivery as reinforcement, Epstein et al. (1981a) trained pigeons to peck a small box in their experimental chamber. Pecks that moved the box to various areas of the chamber were reinforced. The box was then removed from the apparatus and in its place a small plastic banana was suspended from the ceiling of the chamber. Subjects were then trained to peck this small plastic fruit, again with grain delivery as reinforcement. After both component repertoires were trained, birds were placed in their chamber with both the box and the banana. The box was located off to the side, while the banana was suspended near the top of the ceiling of the apparatus (out of pecking range, but not if subjects were standing on the box). Eventually, the birds began pecking at the box until they had moved it directly below the banana, hopped up on the box, pecked the banana, and earned a reinforcer.2 As in the case of Kohler's apes, the constituents of the pigeons' behavior were synthesized in a novel way to gain access to food. Outside the laboratory, of course, feral species frequently confront unique problem-solving situations with familiar components that set the occasion for nonexperimental forms of behavioral synthesis. Like the behavioral syntheses observed in the laboratory, these are interpreted as novel responses to unique stimulus configurations.
Epstein and his colleagues have also used pigeons to synthesize performances they describe as "symbolic communication" (Epstein, Lanza, & Skinner, 1980), "tool use" (Epstein & Medalie, 1983), "self-awareness" (Epstein, Lanza, & Skinner, 1981b), and the spontaneous use of "memoranda" (Epstein & Skinner, 1981), as found in humans and other primates. Until recently, however, laboratory simulations of covert behavior (i.e., thoughts, feelings, or images) have received limited attention, because of experimental difficulties in controlling the relevant variables. To set the stage for our synthesis, we must first turn to the study of animal communication in general.
2. The Study of Animal Communication
To understanding communication is difficult; communication based on private stimulation is even more perplexing. With regard to the arbitrariness of the medium of communication, communicative behavior falls along a continuum. Behavior that is elicited by pheromone release or by the visable presence of a conspecific and that subsequently evokes a behavioral change in another member of the species lies near one end (Salzinger, 1973). Near the opposite end are human vocal utterances which by virtue of their symbolic relationships and organization, produce unique and characteristic responses from other members of that community (e.g., I love you). In this target article, we are concerned with nonhuman communicative behavior that shares properties of human behavior, but is unlike either end of this communicative continuum. The communicative behaviors we will be discussing are arbitrary (nonspecies specific) behavior patterns that are taught to animals to enable them to communicate in nonspecies-specific ways. Such learned communicative behavior shares features with complex human communicative behavior and represents a form of verbal behavior according to Skinner (1957), but it is not language.
Traditional accounts of animal communication have typically referred to species-typical exchanges of inherited signal systems. Numerous definitions exist, for example, "the term 'animal communication' has often been used to refer to the kinds of signals which pass to and fro between social animals and help to mold each others' behavior towards some goal which is to their mutual advantage" (Cullen, 1972, p. 101). By "kinds of signals," Cullen meant species-specific response-produced-stimuli controlled by the presence of food or a predator. Dawkins and Krebs (1978) argued that communicative interactions evolved by natural selection; the actor (their term for the sender of a signal) is selected to change the behavior of a reactor (typically a conspecific) to the advantage of the former. Other definitions proposed by comparative psychologists and ethologists are similar.
Recent developments in the laboratory synthesis of interanimal communication, however, reveal that such descriptions are unduly restrictive. Many species have been trained to communicate in nonspecies-specific ways and have continued to interact in this manner without unconditioned reinforcement, aversive stimulation, or deprivation of a primary reinforcer (described further below). These conspecific interactions are based on experimentally imparted (and subsequently synthesized) communicative repertoires that differ from the species-typical biologically shared signal system of the participants.
In these experiments, communicative behavior can be viewed as an exchange of arbitrary discriminative stimuli that subsequently results in the opportunity for reinforcement (conditioned or unconditioned) for at least one of the participants. (A discriminative stimulus, notated "SD," is defined as an environmental event that sets the occasion for reinforced responding, cf. Skinner, 1974). This definition of communicative behavior incorporates important elements of traditional ethological theory (enhanced instrumental effectiveness as a function of interacting with another species), as well as recent developments in laboratory synthesis (viz., the arbitrariness of the cues exchanged). ----------------------------------------------------------------- --------------------------------------- Insert Figure 2 about here ----------------------------------------------------------------- ---------------------------------------
Figure 2, based on interactions between two of Savage-Rumbaugh's (Savage-Rumbaugh, 1984a: Savage-Rumbaugh et al., 1978) chimpanzees, Austin and Sherman, illustrates how communicative exchanges may be analyzed. Communication can be described as a series of interanimal exchanges of response-produced stimuli. In Verbal Behavior, Skinner (1957) referred to the outline of Figure 2 as an interlocking paradigm. The interlocking paradigm highlights the multiple functions of response-produced-stimuli exchanged between organisms during communication. A response can serve as a conditioned reinforcer (Sr) for preceding behavior as well as a discriminative stimulus (SD) for a subsequent response, hence (the notation "Sr D"). These response-generated stimuli are linked socially by communicative responses (Rc).
Skinner (1957) extended the "three-term-contingency," SD->Rv->Sr, to verbal behavior (where Rv = verbal response as opposed to communicative response Rc) and he stipulated that reinforcement for emitting a Rv is provided by another organism. It is clear that the interlocking paradigm (shown in Figure 2) amounts to the interspecies connection of three-term-contingencies (such that reinforcements for each participants' behaviors are inter-dependent or socially linked). This approach has been used in a number of primate studies, as well as more recent experimentation using pigeons, for establishing experimental syntheses of interanimal communication (Epstein et al., 1980; Lubinski & MacCorquodale, 1984; Lubinski & Thompson, 1987). These studies provided the foundation for the communicative features of our synthesis.
Before we apply the interlocking paradigm to our model, we will briefly review experimentally imparted communicative repertoires to illustrate how sophisticated some of these behaviors have become, and how novel the emerging behaviors have been despite their not having been specifically trained.
3. Experimentally Synthesized Communicative Behavior in Nonhuman Organisms
Roger Brown (1970) wrote: "It is lonely being the only language-using species in the universe. We want a chimp to talk to so we can say: 'Hello, out there? What's it like, being a chimpanzee?'" (p. 211). Investigators have been trying to impart arbitrary communicative behaviors to other species for a very long time. Over a century ago, the first scientific article on this topic appeared in Nature. In his, "Teaching animals to converse," Lubbock (1884) described a method for interacting with dogs via pieces of cardboard containing printed words such as "water," "out," "bone," and "food." Other attempts followed; they even involved training apes to vocalize (Furness, 1916; Kellogg & Kellogg, 1933; Hays, 1951). Furness (1916), for example, trained an orangutan to vocalize the words "cup" and "Papa" (in the presence of an actual cup or himself, respectively). The training of this two-response verbal repertoire was arduous, however. The ape's vocal productions were cumbersome, frail, and, quite frankly, unimpressive.
These attempts achieved only marginal success because the apes were trained to engage in vocal behavior (Furness, 1916; Hays, 1951) rather than more appropriate nonvocal communication. In the early 1960's, Beatrix and R. Allen Gardner were the first to point this out, hypothesizing that, because chimps are not well equipped anatomically to engage in spoken verbal behavior, a different modality might enable them to communicate more effectively. They abandoned the vocal medium and adopted American Sign Language (ASL), a communication system used by people with profound hearing impairment. ASL is a manual language based on hand and finger movements; the great apes could execute these relatively easily. The Gardners taught their female chimpanzee, Washoe, over 130 different word signs (Gardner, Gardner, & VanCantfort, 1989). Washoe combined signs into phrases, which generalized to novel situations spontaneously. In addition, through closed circuit television, she was occasionally observed signing upon seeing pictures of objects in books even when no one else was present.
3.1 Experimentally synthesized symbolic systems in primates
Other nonvocal modalities have been useed with comparable success in attempts to address other kinds of questions (cf. Premack, 1970; Rumbaugh and Gill, 1976; Rumbaugh, 1977; Terrace, 1979). One of Premack's chimpanzees (Sara), for example, using an experimentally trained symbolic communication system, displayed behavior consistent with Piaget's concept of conservation for both liquids and solids (Woodruff, Premack, & Kennel, 1978). In subsequent studies, chimpanzees were taught to perform elementary numerical operations using symbols (Boysen & Berntson, 1989; Woodruff & Premack, 1981; see Davis & Perusse, 1988, for a general discussion) and to use symbols to form higher-order conceptual classes (Gardner & Gardner, 1984; Savage-Rumbaugh, Rumbaugh, Smith & Lawson, 1980). Savage-Rumbaugh et al. (1980), for example, first trained chimpanzees to ask for various tools and food items symbolically. Then she taught her apes to further classify these items, as well as some new ones that had not been introduced before, into food and tool categories. Gardner and Gardner (1984) found that their chimpanzees could perform a similar task by classifying pictures (again both novel and familiar ones) into various classes. Thus, we can conclude from these studies that apes are capable not only of using nonvocal communicative modalities but also of extending their repertoire of signs and lexicons to form higher-order conceptual classes.
Gardner and Gardner (1984) reported that once three stimulus equivalence relationships were established, that is, the ability to exchange "objects" for "signs of objects" and for "photographs of objects," their chimpanzees would naturally extend their repertoire to new photographs of familiar objects, signing "orange," for example, in response to a previously unseen photograph of the fruit. In addition to generalizing to novel photographs, the chimps spontaneously displayed "creative" stimulus -> response relationships without being specifically taught. Great apes equipped with these communicative repertoires use symbols in uninstructed ways, combining strings of symbols in meaningful combinations to communicate integrated concepts (e.g., "sweet" and "water" to report on the eating of watermelon, Gardner & Gardner, 1984; Premack, 1976). These findings oppose the idea that "Man is the only animal to have combinatory productive language...a species-specific form of behavior" (Miller, 1967, p. 83).
Conspecific tutoring has recently been observed in the Gardners' laboratory (cf. Gardner et al., 1989). Over a five-year period, the Gardners raised an infant chimpanzee ("Loulis"). Loulis lived with Washoe and other chimpanzees who were proficient with ASL signs. During these five years, human experimenters refrained from signing in the presence of Loulis, ensuring that the chimpanzees were the only ones who had the opportunity to communicate with signs in Loulis's presence. The Gardners wanted to ascertain whether the chimpanzees would spontaneously try to teach Loulis to sign. Washoe, indeed, appeared to instruct Loulis actively in using signs; during the early stages of instruction, she even moulded Loulis's hands to form the shape of certain signs. At the end of the observation period, Loulis had not only acquired 51 signs, but had spontaneously learned to combine them in novel ways (cf. Gardner et al., 1989).
3.2 Communication based on private states
Chimpanzees also appear to have the ability to communicate about emotional states using their acquired symbolic repertoires. For example, one of Savage-Rumbaugh's chimpanzees, Austin, upon seeing two men in white coats carrying an anesthetized chimp, signed "scared." Along with this behavior, Austin displayed pilo-erection, which was interpreted as a sign of anxiety in the chimpanzee (Savage-Rumbaugh, 1984a). The manual sign "scared" had been taught as a request to play a game in which the teacher dressed up in a costume and pretended to scare the chimps (Savage-Rumbaugh, 1984a, p. 242). One of Terrace's chimps (Nim) signed "angry" or "bite" rather than physically attacking (Savage-Rumbaugh, Pate, Lawson, Smith, & Rosenbaum, 1983, p. 458). These anecdotes, and others, suggest that it may be possible to train primates and other species to report to one another on their internal emotional states in nonspecies-specific and perhaps even very sophisticated ways (cf. Mackintosh, 1987).
The accumulation of this impressive evidence led Griffin (1976, 1982, 1984) to state: "the possibility of animal introspection is more than a will-o'-the wisp; it is a potential method which has already been employed to a very limited degree by the Gardners, Fouts, and other students of chimpanzees, and one that is ready for development and exploitation with other species to roughly the degree that they employ flexible communication systems" (Griffin, 1976, p. 534). Terrace and Bever (1976) have speculated along similar lines: "monkeys can discriminate between the expression of different emotional states, as shown on video displays....Thus, learning to label emotional displays may not prove difficult for chimpanzees....While we have no bases at present for asserting that a chimpanzee could engage in the kind of introspection that is entailed in a description of its own feelings, or emotions, we find this possibility intriguing" (p. 581).
4. Laboratory Synthesized Symbolic and Communicative Exchanges Between Nonhuman Organisms
"It would be astounding to discover insects or fish, birds or monkeys, are able to talk to one another....[because] Man is the only animal that can talk...that can use symbols" (Black, 1969, p. 3). The first convincing demonstration that Black's assertion is incorrect was provided by Savage-Rumbaugh et al. (1978)--the interanimal exchange illustrated in Figure 2. They successfully taught chimpanzees to use symbols to report to other chimpanzees the presence or absence of the objects in a nearby room that only one animal could see. This ground breaking study demonstrated that organisms other than humans can learn to interact communicatively by exchanging arbitrary symbols with one another. Nonhuman animal/animal exchanges of symbols, as well as experimenter/nonhuman exchanges turned out to be possible.
Savage-Rumbaugh et al. (1978) taught a pair of chimpanzees (Sherman & Austin) a long sequence of chained behaviors: They were first taught to select 11 different geometric forms (e.g., "square," "circle," "triangle," etc.) from an experimental panel. These forms correspond to 11 distinctive food and drink items ("bean cake," "orange juice," "banana," etc.). Subsequently, Sherman and Austin learned to select photographs of those 11 consumables, for which they were given the corresponding food items (see Figure 2). This completed the first phase of the experiment. In the second phase, the animals were taught to communicate with one another. One chimpanzee would be taken to a nearby room and shown one of the 11 consumables while the other waited by an experimental panel. Sherman and Austin were then reunited. The chimp who had seen the item informed the other by pressing the corresponding geometric form on the response panel. The second chimpanzee, upon observing this response, was taught to point to the photograph of the hidden item from among 11 pictures of the items located nearby. The symbolic exchange was thereby completed. If both subjects performed the symbolic exchange correctly, they were rewarded with social praise or food or drink, though not with the specific consumable item that had been presented in the adjacent room.
Following Savage-Rumbaugh et al. (1978), Epstein et al. (1980) set out to determine whether interanimal symbolic exchanges could be learned by pairs of pigeons if they were provided with the proper training and experimental medium. Their goal was to teach two pigeons to reliably exchange discriminative stimuli that had been matched to an aspect of their external environment. Only three symbolic relationships were taught to the pigeons, however, in contrast to the 11 taught to the chimpanzee. Their experimental apparatus (Figure 3) consisted of two contiguous chambers separated by a transparent divider. Therein, two pigeons ("Jack" and "Jill") were trained to exchange information regarding colors ("Red," "Green," and "Yellow") using letters ("R," "G," and "Y," respectively). First, the pigeons were independently trained with traditional fading and shaping techniques (Catania, 1992) to match arbitrary discriminative stimuli: the 3 colors matched the 3 letters. These colors were recessed behind a curtain in one chamber and hence available only to one of the pigeons. "Jill" was taught to match colors to lettered discriminative stimuli. "Jack" was trained to request a color by pecking the "What Color?" key and then the "Thank You" key after having received this information. Both birds were food deprived and received mixed grain for correct responses.
Once each bird had learned the string of components necessary for the targeted synthesis, they were placed in their experimental chamber simultaneously and the apparatus was programmed so that reinforcement for each component was contingent on the other bird's behavior. That is, the constituent components of the birds' independent behaviors were experimentally programmed to be interdependent, comparable to the interlocking paradigm illustrated in Figure 2 for the chimpanzees. The following experimental synthesis was then observed. ----------------------------------------------------------------- --------------------------------------- Insert Figure 3 about here ----------------------------------------------------------------- ---------------------------------------
Jack started the communicative exchange by pecking a rectangular key labeled "What color?" (see Figure 3). This response illuminated one of the 3 colors behind Jill's curtain; Jill would then thrust its head through the curtain to observe what color was illuminated and then peck one of the 3 lettered keys corresponding to the illuminated color. This response automatically illuminated the letter pecked by Jill. Next, Jack pecked a key labeled "Thank you," thus rewarding Jill with mixed grain. Jack then observed the illuminated letter on Jill's response panel and pecked the appropriate corresponding color on its response panel. If Jack pecked the correct key, it received mixed grain; the cycle would then continue with different colors appearing behind Jill's curtain (following the "What color?" response) quasirandomly. Hence many of the important social components of the Savage-Rumbaugh interanimal communicative exchange were structurally preserved in the Epstein et al. study. Indeed, some intriguing nonprogrammed communicative behaviors emerged over time and appeared to contribute to the integrity and efficiency of the birds' exchanges. The birds seemed to attend closely to the discriminative stimulus changes in each other's chamber, after each had completed a response. If either bird was at all sluggish after completing the next component of the interaction, the other would peck at the Plexiglas, which appeared to "hurry the other along."
It should be noted, however, that the stimulus interrelationships were more complex for the chimpanzees than for the pigeons. Only a subset of the criteria for symbolic meaning (with respect to the letters exchanged between subjects) was met by the birds, the letters and colors had only a unidirectional functional significance, rather than the bi-directional/transitive symbolic relationships discussed earlier in the primate studies (Gardner & Gardner, 1984; Savage-Rumbaugh, 1984a & b; Savage-Rumbaugh et al., 1980). This critical difference between the chimpanzee and the pigeon preparations does not affect the structure of the social interaction, however, only the nature of the stimulus equivalence relationships between elements of the exchange. The stimulus (cue) relationships were more complex for the apes than for the pigeons.
Following the Epstein et al. (1980) demonstration, Lubinski and MacCorquodale (1984) reported a second two-pigeon communicative exchange, with birds trained to interact in the absence of food deprivation and without material rewards corresponding to such states. One of the objectives of this study was to answer the objection of several commentators who had noted that, unlike most human transactions, Epstein et al.'s tasks (like many primate tasks) were tied to specific biological drive states and material rewards (cf. Mounin, 1976; Savage-Rumbaugh, 1984b; Walker, 1983). Lubinski and MacCorquodale wanted to determine whether a comparable performance could be maintained through conspecific social stimulation (i.e., without specific drive states and material rewards).
Lubinski and MacCorquodale constructed an experimental apparatus similar to Epstein et al.'s: two adjoining experimental chambers, separated by a Plexiglas divider, each supplied with an individual response panel (see Figure 4). Two pigeons served as subjects, the "tacter" (trained in the right chamber) and the "mander" (trained in the left chamber).3 Three colors ("red," "white," and "yellow") served as discriminative stimuli; these were arbitrarily matched to the three letters "R," "W," and "Y," respectively (see Figure 4). ----------------------------------------------------------------- --------------------------------------- Insert Figure 4 about here ----------------------------------------------------------------- ---------------------------------------
The experimental design was similar to Epstein et al.'s, except that when the mander pecked "Thank you," the tacter received an intermittently flashing light, which had previously been paired contingently with both food and water (multiple commodities making it a more powerful consequence, a generalized conditioned reinforcer, notated SGcr in Figure 4). When the flashing light appeared, the tacter could receive food or water by pecking the corresponding food or water keys. (The mander was food-deprived to 85% of its free-feeding weight throughout the entire experiment.) Initially, the tacter performed this interaction while deprived of either food or water. Subsequently, however, its performance was observed when it was satiated for food and water and unconditioned reinforcement was absent (the food and water keys were inoperative); under these conditions, the presentation of the flashing light, contingent on a correct response, operated as the sole programmed consequence for completing the inter-animal exchange. Some intriguing communicative responses emerged in this study as well, and these appeared to contribute to the integrity and efficiency of the exchange in, if anything, more ways than the emergent behaviors observed by Epstein et al.
When satiated, the tacter's performance was noticeably more sluggish than when it was deprived of either food or water. On those days, following the mander's response "What Color?," the tacter would frequently not respond immediately to the discriminative stimulus change in its chamber. At this point, the mander would peck the Plexiglas repeatedly and exhibit species characteristic aggressive displays apparently directed toward the tacter. Often, the tacter would then approach the keyboard and match the projected color to the appropriate letter. Because of the dynamic characteristics of this interaction, Lubinski and MacCorquodale (1984) decided to assess the significance of the mander's social facilitation in a second experimental phase.
For this phase of the synthesis, all contingencies regulating the interanimal exchange were maintained except the following. The birds were prevented from seeing one another by covering the Plexiglas barrier with an opaque divider. The tacter's deprivation and satiation conditions were alternated (ABAB) and the birds' communicative performance was observed. When satiated and without visual access to the mander, the tacter stopped matching letters to colors and this condition was eventually terminated. Subsequently, however, the opaque barrier was removed for a final condition (with the tacter deprived and satiated in alternate ABAB sessions, as before). With visual access to the mander reinstated, the tacter's matching behavior reappeared, following the mander's request, even though the tacter was food and water satiated. That is, when satiated and unable to see the mander, the tacter stopped matching colors to letters; however, when satiated but with visual access reinstated, the tacter continued to communicate with the mander by exchanging letter discriminative stimuli matched to colors. This matching behavior appeared to depend on the mander's display behavior (i.e., nonprogrammed emergent communicative phenomena), which was now readily available to the tacter through the Plexiglas.
Savage-Rumbaugh (1986) and Terrace (1985) have argued that pigeon preparations model communicative behavior only superficially -- that subjects appeared to influence each other but, on closer scrutiny, their behavior was simply a function of the programming equipment. The Lubinski and MacCorquodale (1984) study demonstrates that the presence of a conspecific is necessary for social facilitation under certain circumstances (i.e., when the behavior of one of the participants of the dyadic exchange has low probability--because of satiation and the absence of material rewards). Under these circumstances, emergent communicative phenomena independent of the programming equipment appeared to be critical for maintaining communication. We next review the nature and status of private events in psychology in order to develop the rationale for this component of our synthesis.
5. The Causal Significance of Private Events and the Accuracy of Verbal Reports of Private Events
Private events have been defined as "[t]hose phenomena of psychological interest, taking place "inside the skin," at a covert level, observable beyond the first person through indirect means, if at all. 'Feelings,' 'thoughts,' and 'perceptions' are terms covering the majority of these phenomena" (Schnaitter, 1978, p. 1). A causal role of some sort of stimulation, arising from within the skin, has been assumed in much of psychological and psychiatric theory. That is, many theorists, as well as most lay people, assume that the primary stimuli for human action arise from inside the body. Moreover, the putative causal role of internal stimulation has been explored at length (cf. Boring, 1950; Lyons, 1986), with major theories of psychological adjustment and maladaptivity assuming that thoughts and feelings prompt actions. Indeed, the psychological significance of private phenomena is noted in nearly all comprehensive views on human behavior, from James (1890), Wundt (1894), and Titchener (1899) through the humanistic movement (Maslow, 1962; Rogers, 1961), psychoanalysis (Freud, 1895; Jung, 1959), phenomenological psychology (Allport, 1937), and learning theory (Hull, 1943; Skinner, 1957; Spence, 1956; Tolman, 1932). Interest in covert behaviors has also emerged in the applied areas of psychology, where counselors and therapists [using, for example, Beck's (1976) Cognitive-Behavioral Therapy or Ellis' (1970) Rational-Emotive Therapy] have attempted to moderate clients' affect by restructuring their thoughts (cf. Kendall & Hollon, 1981). In spite of psychology's ubiquitous interest in covert events, however, it is only recently that nonhuman models of communication based on such phenomena have been attempted, although a possible paradigm for imparting primitive vocabularies to species has been available for several years.
Over 20 years ago, Kenneth MacCorquodale (1969) wrote: "It is somewhat curious that Skinner, the most thoroughgoing behaviorist, is the only one who has been willing to discuss private stimuli, which he has done with characteristic consistency since 1945" (p. 837). MacCorquodale was referring to Skinner's treatment of verbal reports of private events, which first appeared in Skinner (1945). This original account has been augmented by Skinner and various writers over the last 50 years (Catania, 1972, 1980, 1992; Day, 1976; MacCorquodale, 1969, 1970; Moore, 1980, 1984; Schnaitter, 1978; Segal, 1977; Skinner, 1953, 1957, 1974; Winokur, 1976; and Zuriff, 1979, 1985). These writers have argued that we come to know our environments (i.e., both our external and internal environments) through two types of sensory modalities--exteroceptive and interoceptive receptors.
Stimuli emanating from our external (public) environment are sensed through the exteroceptive receptors, which include the common sensory modalities (e.g., vision, audition, gustation and olfaction). In contrast, internal (private) stimulus events (i.e., stimulus events happening underneath the skin) are sensed through interoceptors. It is commonly accepted that interoceptive sensory events emanating largely from smooth muscles and glands are involved in ("phenomenological") stimulation accompanying verbal reports of "anxiety," "fear," "joy," "depression," and "excitement." Although not responsible for all subjective components of emotional states, the circulatory, digestive, proprioceptive, and respiratory systems are involved in providing interoceptive stimulation on which statements about affects are based; and covariations in the properties of these physiological events as a function of reinforcing stimuli, emotionally charged punishers and reinforcers, are well documented (Buck, 1987; Carlson, 1986; Tuma & Masur, 1985).
5.1 CNS interoceptive transducers
Discussions of discriminative performance based on exteroceptive cues typically proceed without reference to their transducers (e.g., the photoreceptors of the retina, or the hair cells of the Basilar membrane), because these mechanisms, well known for nearly a half century, have been taken for granted. By contrast, transduction mechanisms involved in brain mediated interoceptive discriminations were largely unknown until recently, and had been the subject of a good deal of speculation. Advances over the past two decades, in neurochemical receptor assays and in behavioral and neuropharmacology research methods have begun to clarify some of the transduction mechanisms in such internal discriminations (e.g., Balster & Colpaert, 1988), including those involved in familiar affective states. Since this literature is not as well known as the transduction literature on external discriminative performance, it is worth noting that research in this area has uncovered a number of intriguing findings..
For example, the mu, kappa, and sigma opioid recepters of the CNS are associated with analgesia and euphoria produced by opiates (cf. Woods et al., 1988); and the GABA recepter complex plays a critical role in anxiety (Gray, 1982). Recent research with neuropeptide Y suggests that it acts on brain chemical receptors controlling aspects of hunger sensations, though these interoceptive stimuli are not isomorphic with those associated with hypoglycemia (Jewett et al., 1991; Schuh et al., 1991). Interneurons are not typically thought of as afferent transducers, but their responsiveness to chemical stimulation appears to be not unlike those in taste or olfaction; the difference is that the chemical compounds of the former are carried to their source of effect by the circulatory system. This may provide yet another way of illustrating Natsoulas's (1983, 1985) point that certain internal states ("sensations") may occur without stimulation being carried to the CNS by peripheral channels.
5.2 Establishing exteroceptive and interoceptive discriminations
Skinner (1945, 1957) argued that the three term contingency used to explain exteroceptive stimulus control and inter-animal communicative exchanges thereof (namely, exteroceptive discriminative stimulus -> verbal response -> conditioned reinforcer), notated SD-> Rv -> Sr, could be extended to the functional control of behavior by interoceptive stimulation. This extrapolation, however, exchanges SD = public exteroceptive stimulus events with SD = private interoceptive stimulus events. The same paradigm used to explain discriminations based on (public) exteroceptive stimuli, along with properties of discrimination/generalization, is thereby extended to (private) interoceptive stimulation. A problem arises, however, regarding the way in which interoceptive stimulus control is learned, compared to exteroceptive stimulus control.
Exteroceptive stimulus control is relatively easy to explain; discriminations of public objects (e.g., cats and dogs) can be taught to a child by reinforcing appropriate verbal responses based on conventional correspondences (e.g., responding "it's a cat" when a cat enters the room and reinforcing a child for doing so). Extinguishing inappropriate responses further facilitates learning the discriminations. Thus, if parents and teachers implement the following four contingencies:
SD(Dog)-> Rv(Dog)-> sr("Yes") and SD(Cat)-> Rv(Cat)-> sr("Yes") versus SD(Dog)-> Rv(Cat)-> sr("No") and SD(Dog)-> Rv(Cat)-> sr("No")
most children will readily learn to discriminate these two classes of animals. How children learn to discriminate and report on distinct private events (affective states like anxiety versus excitement versus joy--private events emanating from interoceptive stimulation) requires a more complex explanation, however. Given that parents and teachers never have direct access to these events, how can they teach children to discriminate them?
Skinner hypothesized that people (teachers) infer private stimulation based on collateral behavior or attendant contextual factors. For example, the presentation or removal of punishers and reinforcers provide cues for inferring how people feel. Following a positive reinforcer (e.g., when a person receives praise or an expression of love), we infer feelings of satisfaction or joy on the part of the recipient. On the other hand, removal of powerful positive reinforcers (e.g., death of a loved one, being dismissed from a valued job) is associated with feelings of depression or discouragement. Anxiety and fear, moreover, are related to the presentation of conditioned and unconditioned aversive stimuli, respectively, whereas the removal of these negative reinforcers is associated with experiencing relief. Thus, observing what punishers and reinforcers are presented or removed yields useful cues for inferring an individual's emotional state. In addition to these cues, people often exhibit characteristic responses associated with attendant emotional states (e.g., laughing when happy, crying when sad). Darwin (1872) discussed these cues in detail, often observing cross-cultural consistency between specific facial expressions and inferences of distinct emotional states made by observers.
According to this line of reasoning, people infer and differentially reinforce verbal reports that discriminate between contrasting forms of internal stimulation by observing the events responsible for generating the states directly or by observing the behavioral correlates of such states, or both. Contextual cues (e.g., the social setting: academic settings, cocktail parties, & family settings) can also contribute to accurate inferential assessments of private states and can be important indicators of relevant changes in internal conditions.
Although verbal reports of private events involve discriminations of real physical events generated by interoceptive stimulation (or possibly direct interneuronal activation) as well as observing one's own behavior, the accuracy of such reports must be interpreted with caution. People who differentially reinforce verbal responses in the presence of public cues (correlated with distinct internal happenings) never have direct access to the attendant private states or their intensity. People must rely on fallible public exteroceptive stimuli associated with private events and they must differentially reinforce correspondences between such events and appropriate verbal responses (the Rv's in Skinner's three-term contingency). This is far removed from the precision found in teaching exteroceptive discriminations (like our earlier example of discriminating cats and dogs). Although collateral behaviors and contextual factors may be correlated with distinct forms of private stimulation, their association with such stimulation is often less than consistent.
If, however, we could somehow achieve reliable access to private stimulation (i.e., technically, if we could gain access to the functional relation SD -> Rv , private event -> verbal response), as we believe we can, and differentially reinforce it, as in the establishment of an exteroceptive discrimination, there is no reason to suppose that interoceptive stimuli are incapable of generating interoceptive discriminations as precise as exteroceptive stimuli are in establishing exteroceptive discriminations (Overton, 1971). Our next section provides empirical support for this position.
6. Experimental Demonstrations of Interoceptive Stimulus Control
Just as the classical (Pavlov, 1927) and instrumental (Thorndike, 1911) conditioning paradigms lead organisms to successful discrimination of exteroceptive stimuli (Catania, 1992; Honig & Staddon, 1977; Mackintosh, 1974), they can lead to the same capability when interoceptive stimuli are used. A substantial literature has demonstrated how private stimulus events can control the behavior of laboratory animals. Bykov and Ivanova (cf. Bykov, 1928), in the first study of this kind, revealed that interoceptors responded to classical conditioning in the same manner as exteroceptors. In this pioneering work, Bykov and Ivanova used infusions of saline in the stomach and diuresis (or urine formation) as the unconditional stimulus (US) and unconditional responses (UR), respectively. They demonstrated, through Pavlovian pairings, that the diuretic response could be brought under the control of irrigation-injections of saline solution injected into the stomach. Saline-irrigation-injections acquired the role of a conditional stimulus (CS). Thus, autonomic behavior not under direct instrumental control can be classically conditioned. Several subsequent reports revealed that instrumental responses could also be controlled by interoceptive stimuli (Hull, 1933; Leeper, 1935; Kendler, 1946). Amsel (1949), for example, using a T-maze, conditioned rats to escape shock by going to one arm of the maze when they were hungry and the other when thirsty. Thus, the differential stimuli associated with food versus water deprivation acquired discriminative stimulus properties for differential responding that was negatively reinforced.
Laboratory animals can also be trained to differentiate between the internal state produced by a behaviorally active drug and that associated with a vehicle (usually saline) injection or placebo (Schuster & Brady, 1964; see Thompson & Pickens, 1971; & Colpaert & Balster, 1988, for summaries). In these methods, a food deprived animal is injected with a training drug (e.g., morphine) and given the opportunity to respond by pressing one of two levers or pecking one of two keys, depending on the species, and this leads to food reinforcement. The appropriate response following drug administration is the drug cue lever or key; responses to the other alternative (when in a drug state) produce no reinforcer. On alternate days, however, the animal receives a vehicle injection. On those days, the opposite response is defined as correct: presses on the drug lever go unreinforced, whereas presses on the vehicle level are reinforced. With this procedure, animals rapidly learn to respond only to the drug cue lever on days the training drug has been administered and to the alternative lever (vehicle cue lever) on days when the vehicle has been administered. Differential responding with this procedure has been demonstrated using a wide array of behaviorally active drugs (Colpaert, 1978; Goldberg & Stolerman, 1986; Griffiths et al., 1985; Holtzman, 1985; Overton, 1971). These agents also engender stimulus generalization gradients similar to those of exteroceptive stimuli as a function of modifications of their chemical composition. So unfamiliar or novel drugs can be "classified" by laboratory animals provided they have learned to discriminate a compound with similar effects. Moreover, drugs classified by animals as members of the same class (through interoceptive stimulus generalization studies) generate similar verbal responses on conventional mood questionnaires designed to assess distinct emotional states in humans.
6.1 Discriminative Stimulus Properties of Drugs
Many of the psychophysical properties of exteroceptive stimuli have been replicated using drugs as interoceptive stimuli (Colpaert & Balster, 1988). Across a variety of species (pigeon, rat, rhesus monkey, squirrel monkey, and primates), specific agents in a given pharmacological class, for example, amphetamine and cocaine ("stimulant"), alcohol and pentobarbital ("depressant"), and morphine and heroin ("opiate analgesic") generalize to one another in the two-choice (drug versus saline) discrimination tasks described earlier (Goldberg & Stolerman, 1986; Griffiths et al., 1985; Seiden & Balster, 1985; Thompson & Unna, 1977). Drugs generalizing from one to another in nonhuman laboratory studies typically create similar subjective effects in humans (see below). Moreover, the drugs with the greatest reinforcing efficacy in laboratory animals generate, in humans, the highest subjective ratings of "euphoria" or "liking." Thus, the aptness of pharmacological agents for our animal model is confirmed by a variety of cross-connections bridging human and nonhuman behavioral pharmacology.
In human studies, for example, subjective effects of drugs have been assessed via verbal reports in response to standardized questionnaires (Beecher, 1959; Schuster & Johanson, 1988). The content of these inventories is very similar (often identical) to that of conventional mood inventories designed to assess common affective or emotional states (cf. Nowlis, 1953, 1970; Nowlis & Nowlis, 1956; McNair, Lorr, & Droppleman, 1971; Tellegen, 1985; Watson & Pennebaker, 1989; Zevon & Tellegen, 1982; Zuckerman & Lubin, 1965). In these drug studies, subjects are asked to report on their mood or the likelihood of certain behaviors in ways that can be objectively scored (such as, true/false): "I feel like going for a walk" or "My stomach feels funny." Different subsets of items that correlate with one another (for internal-consistency reliability) and with common feeling or mood states (for external validity) were clustered and allowed to function as scales. These scales are then used to assess the subjective strength of various classes of behaviors and mood while subjects are experiencing influences of different drugs (Fischman, 1977; Fischman, Schuster, Resnekov, Shick, Krasnegor, Fennell, & Freedman, 1976; Griffiths & Balster, 1979; Johanson & Uhlenhuth, 1980a, 1980b, 1981, 1982; Martin, Sloan, Sapira, & Jasinski, 1971; Schuster & Johanson, 1988; Uhlenhuth, Johanson, Kilgore, & Kobasa, 1981).
One of the more widely used instruments of this type is the Addiction Research Center Inventory (ARCI) (Hill, Haertzen, Wolbach, & Miner, 1963). This inventory consists of 550 (true-false) items combined into seven scales that measure reports of mood fluctuations and internal conditions associated with contrasting drug states. Drug states that can be assessed with this instrument appear in quotes (which represent prototypes of the drug class measured by the scale) and are followed by (true/false) item exemplars: "Morphine" (I have a pleasant feeling in my stomach; My nose itches), "Amphetamine" (My thoughts come more easily than usual; I feel as if I would be more popular with people today); "Alcohol" (I feel like joking with someone; My appetite is increased).
Not only do verbal responses on the ARCI vary as a function of the common drug-states (and familiar emotional states), but drugs with which a person has had no experience but that share pharmacological properties with familiar drugs generate similar patterns of verbal responses. Thus, administering barbiturates to human subjects generates patterns of verbal responses on the ARCI similar to those of benzodiazepines (Griffiths, Roache, Ator, Lamb & Lukas, 1985); cocaine generates a response pattern comparable to that of amphetamine (Schuster et al. (1981). Results like these lead Schuster et al. (1981) to the following generalization: ... "it is possible to determine whether an unknown drug belongs to the opiate, psychomotor stimulant, sedative-hypnotic, or hallucinogenic drug class on the bases of its subjective effects" (p. 116) ... They concluded that... "[T]he drug classes based upon discriminative effects in animals and upon subjective effects in humans are in striking concordance" (p. 121).
Just as the precision offered by acoustical and optical engineers for controlling exteroceptive stimuli has enhanced our understanding of corresponding exteroceptive systems, the precision offered by pharmacology in controlling interoceptive stimuli may help both to sort out individual and species differences in the nature of interoceptive systems and to for build an animal model of interpersonal communication based on private stimulation. Indeed, it is generally agreed that some pharmacological stimuli generate in laboratory animals interoceptive states that share components with affective states in humans; there is also evidence that comparable subcortical structures are involved in the mediation of these phylogenically shared states (Gray, 1982; Schuster & Johanson, 1988; Tuma & Maser, 1985).
Given that certain drugs are capable of controlling one class of private events (feelings) with a high degree of reliability, they seem ideally suited for the present experimental study of interanimal communication. Through behavioral- and neuropharmacology, compounds have been identified that selectively bind to specific interoceptors and may be useful in resolving the problem of the inaccessibility of private events (and inaccuracies regarding verbal reports of such events). Behavioral pharmacology offers a technology for establishing interoceptive discriminations that serve as a stimulus for communicative exchanges in which one animal reports to another animal how it feels. We now have a way to gain precise access to the SD -> Rv (private event -> verbal response) relation.
7. An Animal Model of the Interpersonal Communication of Interoceptive ("Private") States
We now return to an explication of the model illustrated in Figure 1. Our idea was to train pigeons to discriminate three contrasting states in their internal milieu (Lubinski & Thompson, 1987). The experimental apparatus was a modified version of the one used by Lubinski and MacCorquodale (1984) (see Figure 5). The experimental chambers were preserved, but the colored lights and letters corresponding to specific colors were replaced by symbols representing specific drugs and drug classes. The drug class symbols corresponded to "stimulant," "depressant," and "no drug," while symbolic names for specific drugs corresponded to "cocaine," "pentobarbital," and "saline," respectively. (As in all the preceding pigeon models, these specific response-key names were chosen for clarity in experimental exposition rather than for imparting symbolic meaning.) The experiment involved two groups of pigeons: drug-cue birds (three) and decoders (two). ----------------------------------------------------------------- --------------------------------------- Insert Figure 5 about here ----------------------------------------------------------------- ---------------------------------------
The individual performances of the decoders were acquired rather quickly: pecking the "How do you feel key?" followed by the "Thank You" key and then matching a letter to a color (they learned this sequence at greater than 90% accuracy over the course of approximately three-months). The drug-cue birds were required to learn a complex interoceptive discrimination. Although in the literature at the time there was one example of a three-key interoceptive discrimination using pigeons (France & Woods, 1985), the subjects' performance was established and maintained under only one drive state. We wanted to impart a three-key interoceptive discrimination across two drive states, thirst and hunger, hence, 2(drive states) X 3(drug states) = 6 conditions. The training procedure for the drug-cue birds was the following.
While deprived of either food or water, drug-cue birds received an I.M. injection of cocaine (3.0 mg/kg), pentobarbital (8.0 mg/kg), or normal saline and were then placed in their darkened experimental chamber. After a 20-minute interval (to allow the drug to become absorbed and distributed), the chamber lights were illuminated and, simultaneously, the three response keys in the drug-cue birds chamber were transilluminated (see Figure 5). Pecks matching the birds' interoceptive state (viz., saline injection = "No drug," pentobarbitol = "Depressant," & cocaine = "Stimulant") were followed by the presentation of the flashing blue light (indicating that food and water pecks would be reinforced with corresponding reinforcers), just as correct exteroceptive matching behavior was reinforced by Lubinski and MacCorquodale (1984). If the drug-cue birds made an incorrect response, a mild punisher ensued: their chamber lights were dimmed for 4 sec. and the trial would start over. This interoceptive discrimination took approximately 10 months to establish (with training sessions conducted six days per week). Nonetheless, 90% accuracy was eventually achieved across all six conditions.
Finally, the experimental contingencies for the decoders and drug-cue birds were programmed to be interdependent, as shown by the interlocking paradigm in Figure 6. The interanimal exchange was synthesized to resemble all relevant features of the Lubinski and MacCorquodale (1984) study, but instead of matching letters to colors, the drug-cue birds matched arbitrary visual symbols to pharmacologically manipulated interoceptive states. Decoders were food deprived throughout the experiment, whereas drug-cue birds were (initially) deprived of either food or water. The drug-cue birds received an injection of one of the aforementioned agents and were then placed in their darkened experimental chamber while, at the same time, a decoder was placed into the adjacent chamber. Following the twenty minute pretreatment interval, the overhead lights in both chambers were illuminated along with the decoder's "How do you feel?" key. The ensuing performance followed. ----------------------------------------------------------------- --------------------------------------- Insert Figure 6 about here ----------------------------------------------------------------- ---------------------------------------
7.1 The synthesis
7.1.1. Phase 1. The synthesized communicative sequence began when the decoder pecked its "How do you feel?" key. This response illuminated the drug-class names, one on each of the drug-cue bird's three response keys (i.e., "Depressant," "No Drug," and "Stimulant"). The drug-cue bird then pecked the response key corresponding to the drug it had received. This response in turn illuminated the "Thank You" key in the decoder's chamber. When the decoder pecked the "Thank You" key, two events ensued concurrently. The drug-cue bird's blue light began to flash and the drug class name previously pecked by the drug-cue bird appeared on the decoder's sample key. From this point on, the rest of the response sequences of the two birds were independent of each other. With the blue light flashing, the drug-cue bird could receive food or water by pecking the appropriate key, while the decoder could receive food by correctly matching the specific drug (among its comparison response keys) to the drug class (on its sample key). Overall, subjects performed this interaction with a high degree of accuracy. For all groups of birds, correct correspondence on the first trial of each experimental session (i.e., a correct discrimination by both birds) across all six states, 2(deprivation) X 3(drug), ranged from 70% to 100%.
At first, the birds performed in a mechanical fashion: they appeared at times to be disinterested in each others behavior. Over the course of a number of days, however, each birds' behavior came more and more under the control of its counterpart. After completing a component of the interlocking sequence, each bird gradually began to orient toward the stimulus change in the adjacent chamber, which was in close proximity to the other bird. After consuming food or water, for example, the drug-cue bird approached the area near the decoder's "How do you feel?" key. If the decoders were at all sluggish in pecking the key when this light became illuminated, the drug-cue bird would rapidly peck the Plexiglas directly above the key while orienting toward the decoder. At this point, the decoders would typically approach and peck the "How do you feel?" key and, then move toward the area by the "Thank You" key, standing in position until the drug-cue bird finished matching its state to the associated letters. If the drug-cue birds were at all sluggish in performing the interoceptive discrimination, the decoders would peck the Plexiglas and emit species characteristic aggressive displays directed toward the drug-cue bird.
We have not found other examples in the literature of interanimal exchange of discriminative stimuli based on the interoceptive state of one of the participants. The drug-cue birds' performances in this interanimal exchange involved discriminating interoceptive stimuli. Their key-pecking performance was controlled by events inside their skin, which did not covary with the particulars of their state of food or water deprivation or aversive stimulation, nor was it reinforced with a characteristic unconditioned reinforcer (e.g., only food or only water) corresponding to a particular state of deprivation or aversive stimulation.
7.1.2 Phase 2. A second objective was to determine whether the discriminative performances established in Phase 1 would generalize to similar states (i.e., private events induced by pharmacological agents that subjects had not previously experienced but were of the same class as those used during their training). This phase of the synthesis was predicated on the idea that the reason humans can describe novel stimulation (e.g., unique "mixing of emotions") is because these states often share components with familiar states they have learned to report. ) and ) as novel agents for the generalization probe, drug-cue birds reliably generalized from cocaine by responding on the stimulant key after amphetamine administration; they generalized from pentobarbitol by responding on the depressant key after chlordiazepoxide administration. One can attribute meaning to these behaviors of the pigeons by suggesting that what they were reporting was that the interoceptive cues from the amphetamine were more similar to cocaine than to pentobarbitol or saline and, conversely, that the cues produced by chlordiazepoxide were more similar to pentobarbitol than to cocaine or saline. (The dyadic accuracy ranged from 84% to 92% correct correspondence during this phase.) Technically, what we observed in this phase of the experiment was stimulus generalization. It is through a similar process, according to Skinner (1945), that humans have the capacity to verbally describe novel, exteroceptive and interoceptive, stimulation.
7.1.3. Phase 3. In Mind, Self and Society (1934), George H. Mead wrote: "It is quite impossible to assume that animals do undertake to express their emotions. They certainly do not undertake to express them for the benefit of other animals" (p. 16). To ascertain whether the type of communicative behavior displayed by the drug-cue birds in Phase 2 would continue when the animals were food and water satiated and recieved no deprivation-relevant reinforcement (with only the flashing light contingent on correct responding), the following experimental probe was conducted.
The same procedure as described for Phase 2 was maintained; however, on several intermittent days, the drug-cue birds were placed in their experimental chambers after receiving twenty-four hour free access to both food and water but subsequent to receiving an injection of one of the pharmacological agents. Their food and water keys were also inoperative during these sessions. When they were food- and water-satiated and had no consumable rewards but only the flashing light contingent on correct responding, the drug-cue birds continued correctly responding to decoder's requests by accurately reporting on their internal states 83-100% of the time. The emergent communicative behavior reported in the earlier phase of this experiment also appeared in this final phase and seemed to maintain the integrety of the exchange.
7.2 Limitations of our model. Linguists and philosophers single out the intent of the communicative exchange as an essential element in human communication. It is in this regard that there are important differences between the pigeons' communicative exchanges in Lubinski and Thompson's (1987) study and reports from one person to another about their feelings. When one person tells another that they are "feeling depressed" or "feeling anxious," we have good reason to believe that the listener's expression of concern and interest motivates the person to disclose feelings. Much of the interview process in psychotherapy, for example, is based on an assumption that clients intend to reveal their feelings, and the listener's interest in those reports maintains the exchange. We have no reason to believe that the pigeons' discriminative responding is primarily controlled by its impact on the other pigeon's behavior. Savage-Rumbaugh (1986) and Terrace (1985), for example, have asserted that the exchanges reported by Epstein et al. (1980), which did not involve interoceptive stimuli, could have proceeded purely mechanically, without the second bird being present. Yet this was clearly not the case in the Lubinski and MacCorquodale (1984) study, in which it was found that when the other bird was not present and motivation for communicating was low, reporting the color of the hidden light ceased. Subsequently, however, when the paired bird was reintroduced into the situation, the bird doing the reporting resumed the interanimal interaction. So critical components of the interaction between the pigeons appeared to come under the control of each other's behavior; these were not programmed by the experimenter, yet they emerged, and were critical for maintaining the exchange. The same emergent behaviors were also observed by Lubinski and Thompson (1987).
The controlling relation, nonetheless, remains different for the pigeon and human under most circumstances. For the human, the listener presumably provides social reinforcement, which is contingent on reports of internal stimulation. For the pigeon, the "listener" behaved aggressively in order to maintain the communicative behavior. Specifically, it flapped its wings and pecked against the transparent partition separating the two compartments until the satiated bird reported the color of the hidden light. Whereas this is very unlike the typical human exchange based on private stimulation, it resembles exchanges seen in certain human situations. Examples would come from confrontational forms of psychological therapy, involving clients who are unwilling or unable to report feelings. Even though such clients may have good reason to feel angry, they are unable to report that feeling. In this situation, the therapist, or possibly other members of a therapy group, gesticulate, raise their voices, or in other ways present an aversive setting to client-listeners until they blurt out their feelings. Functionally, this is the pigeon acting aggressively until its conspecific counterpart responds. Just as pigeons may "have no interest in reporting feelings" to fellow pigeons, humans may also either not recognize their feelings or be reluctant to report them. A difference remains, but the difference is perhaps less than one would think.
8. Implications and Human Parallels
The pigeon experimental situation shares some features of the interactions between parents and children with autism or other developmental disabilities (Keogh, Whitman, Beeman & Halligan, 1987; Reichle, Lindamood, & Sigafoos, 1986). Autistic children are often minimally verbal and unable to report their private experiences to others. Indeed, a common diagnostic feature of autism is the failure to use gestures or other symbolic communicative devices to indicate needs or wants. Thus, to create the desired behavior or outcome in their parents, autistic children may pull or push their parents to the refrigerator and then begin to scream or to hit them until they open the door and retrieve the orange juice. Yet, just as in pigeons, extensive training can often be used to teach autistic children to discriminate familiar objects correctly (e.g., food items, articles of clothing, household objects, places). Children with autism have been taught to use words or signs to report on states of hunger or emotional arousal (e.g., "angry") but they seldom do so spontaneously. Their performance often retains a mechanical quality that requires, at least intermittently, an extrinsic reinforcer to maintain symbolic responding. Communicating their internal feelings to another person does not seem to be a significant factor for maintaining the exchange (apart from the primary reinforcer provided by the listener). The same seems to hold for the pigeon. A final parallel between the present model and children with such handicaps is that the latter seldom develop the capacity to report refined descriptions of the components of their internal experiences (e.g., novel mixings of emotions) whereas most well-socialized adults readily generalize their internal stimulation. The interoceptive discriminative repertoires of some humans, like those of our pigeons tend to be restricted to crude categorizations.
Individual differences in the ability to report on interoceptive stimulation may emanate from endogenous (biological) as well as exogenous (experiential-learning) sources of influences. For example, both human and laboratory animal studies indicate that specific neurochemical receptors are associated with familiar affective states (e.g., the GABA/benzodiazepine receptor and anxiety). Thus, differential numbers of such receptors, selective affinity, differences in intrinsic activity, or differential release of endogenous ligands binding to such receptors could all account for individual differences in the tendency to report feelings. These would modulate the intensity of inner feelings and thereby the likelihood that they were reported. These sources of variability are akin to how differences in retinal cell function is related to visual perception.
There are also basic individual differences in conditionability. People with personality attributes characterized as Personality Disorder or those who score very high on the Psychopathic Deviate (4) scale of the Minnesota Multiphasic Personality Inventory seldom report feeling guilty or anxious. In 1957, Lykken demonstrated that people with such characteristics are not well suited to developing classically conditioned responses to conditioned stimuli, which precede unconditioned aversive stimuli (e.g., skin shock). Lykken and others (Lykken, 1968, 1982, 1984; Tellegen et al., 1988) have argued that these differences stem from individual differences in genetic constitution and may hence help explain (1) such individuals' apparent inability to learn from repeated aversive stimulation or punishment as well as (2) certain individual differences in the readiness to "introspect." Some people have internal economies that are simply more emotionally active than others (Watson & Pennebaker, 1989).
There are other likely sources of variability too. Children with mental retardation often find learning visual or auditory discriminations, especially those involving multiple stimulus elements, very difficult (Zeaman & House, 1979). Even if the child's retinal or cochlear functions are within the normal range, a child with an IQ of 50 will often take far longer to learn a visual or auditory discrimination than an age mate of normal IQ (Baumeister, 1967). Some children with moderate to severe intellectual disabilities may never learn conditional discriminations involving several elements (e.g., when the red light is on, a vertical line is correct; but when the green light is on, a horizontal line is correct). Although we are not aware that it has ever been tested, it seems likely that similar difficulties in the learning of discriminations based on interoceptive stimuli would confront the individual with a significant intellectual deficit, especially given the lack of consistency between private stimulation and the associated public covariates present throughout the learning process. Conditional discriminations involving interoceptive cues should be even more difficult to learn (e.g., "When I have just won a prize and I feel this way, it's called 'excited'", but when I have been knocked down by another child and I feel this way, it's called 'angry'").
Autistic people tend to respond selectively to one element in a complex stimulus array in which all elements are associated with reinforcement. This phenomenon has been called "stimulus overselectivity" (Schreibman & Lovaas, 1973). Autistic children often attend to an element of the stimulus situation that has been irregularly correlated with correct responding but may actually be irrelevant (e.g., the tone of the teacher's voice or the angle of their head as they present the educational task). Presented with several stimuli, one of which is interoceptive and the rest exteroceptive, the autistic child may selectively attend to the interoceptive stimulus. The parent, teacher or therapist would have no way of knowing that the child's behavior was under the control of the interoceptive stimulus, since children with autism have extremely limited verbal ability to report what they are attending to. An autistic childs persistent responding on the bases of an interoceptive discriminative stimulus could present an insolvable puzzle to an adult searching for the external cue to which the child is responding.
When one teaches children to respond discriminatively to interoceptive stimuli and then to report on such stimulation one assumes that the people in the children's environment attend to cues usually correlated with affective states and that they then use those cues to teach children the appropriate label for their feeling. The child cries when hurt, has a gleeful expression when receiving a gift, or wears an unhappy expression when an anticipated enjoyable event is postponed. Some parents may be less effective than others in attending to such cues or in taking the time to teach children to selectively respond to their internal states in the presents of associated cues. The child with the disappointed expression upon learning that a trip to the circus has been delayed would find it difficult to learn the name of the feeling being experienced unless an adult takes advantage of the opportunity to teach it. Parents who are totally preoccupied with other matters (e.g., the chemically dependent person seeking drugs) or who devote much of the time responding to their own private stimulation (e.g., the person with schizophrenia who is embroiled in ruminations and delusional thinking) would probably be ineffective in teaching children to discriminate and accurately report their own private stimulation.
In other instances, the children's reports may be punished by a parent who rejects them (e.g., the parent who believes that feeling discouraged is a sign of weakness or that feeling angry when frustrated is unacceptable). Under such circumstances the ability to accurately report feelings would be attenuated. This could lead to difficulties for which a counselor may be sought.
Finally, just as people have moderated their public experiences with the aid of external prostheses (e.g., eyeglasses & hearing aids), they have used informal pharmacological prostheses to modulate their private experiences. Controlled interoceptive prostheses directed at attenuating, increasing or sharpening discriminative control of interoceptive stimuli have emerged over the past 35 years with the advent of modern psychiatric drugs. Neuroleptic medications have been used to control thought disorders in schizophrenics, and benzodiazepines have been used to reduce pre-aversive interoceptive stimuli among people with serious anxiety problems. Patients with panic attack disorder respond discriminatively to their own interoceptive anxiety stimuli, which are attenuated by administering tricyclic antidepressant medications. People with obsessive compulsive disorder respond discriminatively, not only to the contaminants that lead them to wash their hands, but also to their own feelings of anxiety when they fail to do so. Since washing their hands only reduces the intensity of the interoceptive anxiety stimuli for a short period, the avoidance behavior resumes at frequent intervals until the source of the interoceptive stimulation is attenuated pharmacologically. Individual biological differences contributing to such internal events are the foundation upon which variations in private stimulation are built. So even when parents, teachers, and therapists are consistent in their attempts to impart interoceptive discriminations, individual differences in underlying mechanisms contribute to individual differences in private states and their reports.
Perhaps Catania (1990) was well justified in hypothesizing that language functions primarily as a means of changing others' behavior rather than for communicating information about internal feelings. In the nonhuman world, overt behavioral manifestations may be sufficiently correlated with internal cues experienced by the organism in a state of emotional arousal to render supplementary communicative information based on these private stimuli superfluous. This does not mean that chimpanzees (or other organisms) cannot learn such discriminative repertoires, only that there is typically little adaptive utility for them to do so. As Pinker and Bloom (1990) noted in BBS, private affective stimuli are not well suited as bases for linguistic (grammatical) communication: what T. S. Eliot (1943, p. 16) referred to as ..."the general mess of imprecision of feeling." The human tendency to conceal emotional stimulation for social advantage makes it understandable that such language based repertoires would be more common in humans than in other organisms. In the overt behavior of adult humans, the lack of consistent correlates of affective stimuli means that some other correlated behavior (language) is needed for this useful information. Individual differences in biological substrates (Tellegen et al, 1988) and learning histories (Skinner, 1945) give some people an adaptive advantage in this respect over others.
Figure 1. A two-pigeon communicative exchange based on the internal state of one of the birds (Decoder, left; Drug-cue bird, right). (A) The Decoder pecks the "How do you feel?" key. (B) The Drug-cue bird pecks the drug-class letter corresponding to its internal state. (C) The Decoder pecks the "Thank You" key, which presents the flashing blue light to the Drug-cue bird. This response also presents to the Decoder the drug-class letter previously pecked by the Drug-cue bird. (D) The Decoder matches the drug-class letter (projected on its sample key) by pecking it and then pecking the letter representing the specific drug that the Drug-cue bird is currently experiencing. The Drug-cue bird attends to the flashing blue light. (E) The Decoder receives food upon correctly completing the communicative exchange and the Drug-cue bird receives water (adapted from Lubinski & Thompson, 1987).
Figure 2. An interlocking paradigm of a communicative exchange between two of Savage-Rumbaugh's (1984a) chimpanzees ("Austin" & "Sherman"). The notational system follows: SD = discriminative stimulus, Sr D = a conditioned reinforcer which also functions as a discriminative stimulus, Sr R = a conditioned reinforcer plus an unconditioned reinforcer, Rc = a communicative response.
Figure 3. Epstein, Lanza, and Skinner's (1980) experimental arrangement: Jack was trained on the left, Jill on the right. Jack needs information about the color behind the curtain in the upper right-hand corner of Jill's keyboard. The R, G, and Y letters on Jill's keyboard correspond to the colored lights projected behind the curtain. The three keys below the WHAT COLOR? key on Jack's keyboard are yellow, red, and green, from left to right. The birds are separated from one another by a Plexiglas divider.
Figure 4. Lubinski and MacCorquodale's (1984) experimental arrangement for training an exteroceptive communicative exchange, ultimately without deprivation and material rewards. The top two panels are the front wall of each chamber. The third panel is the top view of the floor of the left chamber. Adjoining keyboards for the two birds are separated by a Plexiglas barrier.
The mander's keyboard is on the left; the tacter's is on the right.
Figure 5. The details of the response operandi, stimulus configurations, and reward mechanisms structuring the experimental synthesis in Figure 1 (adapted from Lubinski & Thompson, 1987). Figure 6. Interlocking communicative paradigm illustrating the technical features of the stimulus events exchanged between subjects. The notational system follows: SD = discriminative stimulus, S- = S - delta (i.e., nonavailability of reinforcement), Sr D = a conditioned reinforcer which also functions as a discriminative stimulus,
Sr R = a conditioned reinforcer plus an unconditioned reinforcer, Rc = a communicative response, SR = unconditioned reinforcer, R = response, SGcr = generalized conditioned reinforcer.
Allport, G. W. (1937). Personality: A psychological interpretation. New York: Holt.
Amsel, A. (1949). Selective association and the anticipatory goal response mechanism as explanatory concepts in learning theory. Journal of Experimental Psychology, 59, 785-799.
Baumeister, A. A. (1967). Problems in comparative studies of mental retardates and normals. American Journal of Mental Deficiency, 71, 869-875.
Beck, A. T. (1976). Cognitive therapy and the emotional disorders. New York: International Universities Press.
Beecher, H. K. (1959). Measurement of subjective responses: Quantitative effects of drugs. New York: Oxford University Press.
Bernard, C. (1885). An introduction to the study of experimental medicine. Trans. by Henry C. Green (New York: Henry Schuman, 1949; reprinted Dover, 1957).
Black, M. (1969). The labyrinth of language. New York: Praeger.
Boring, E. (1950). A history of introspection. Psychological Bulletin, 50, 169-189.
Boysen, S. T., & Berntson, G. G. (1989). Numerical competence in a chimpanzee (pan troglodytes). Journal of Comparative Psychology, 103, 23-31.
Brown, R. (1970). The first sentences of child and chimpanzee. In R. Brown (Ed.), Psycholinguistic papers (pp. 208-234). New York: Free Press.
Buck, R. (1987). Human motivation and emotion. (second edition). New York: Wiley.
Bykov, K. M. (1928). The cerebral cortex and the internal organs. Translated by W. H. Gannt (1957). New York: Chemical Publishing Co., Inc.
Carlson, N. R. (1986). Physiological psychology (second edition). New York: Allen & Bacon, Inc.
Catania, A. C., (1972). Concurrent performances: Synthesizing rate constancies by manipulating contingencies for a single response. Journal of the Experimental Analysis of Behavior, 17, 139-145.
Catania, A. C. (1980). Autoclitic processes and the structure of behavior. Behaviorism, 8, 175-186.
Catania, A. C. (1983). Behavior analysis and behavior synthesis in the extrapolation from animal to human behavior. In G. Davey (Ed.), Animal models of human behavior (pp. 51-69). Chichester: Wiley
Catania, A. C. (1990). What good is five percent of a language competence? Behavioral and Brain Sciences, 13, 729-731.
Catania, A. C. (1992). Learning (3rd. ed.). Englewood Cliffs, NJ: Prentice Hall.
Colpaert, F. C. (1978). Discriminative stimulus properties of narcotic drugs. Pharmacology, Biochemistry, and Behavior, 9, 863-887.
Colpaert, F. C., & Balster, R. L. (1988). Transduction mechanisms of drug stimuli. Berlin: Springer-Verlag.
Cullen, J. M. (1972). Some principles of animal communication. In R. A. Hinde (Ed.), Non-verbal communication (pp. 101-122). Cambridge: Cambridge University Press.
Darwin, C. (1859). On the origin of the species. London: John Murray.
Darwin, C. (1965). The expression of the emotions in man and animals. Chicago: University of Chicago Press. (Originally published in 1872).
Davis, H., & Perusse, R. (1988). Numerical competence in animals: Definitional issues, current evidence, and new research agenda. Behavioral and Brain Sciences, 11, 561-615.
Dawkins, R., & Krebs, J. R. (1978). Animal signals. In J. R. Krebs, & N. B. Davies (Eds.), Behavioral ecology. Oxford: Blackwell.
Day, W. F. (1976). Contemporary behaviorism and the concept of intention. In J. K. Cole & W. J. Arnold (Ed.), Nebraska Symposium on Motivation (pp. 65-131). Lincoln: University of Nebraska press.
Descartes, R. (1962) Rene Descartes (1596-1650) On mechanisms of human action, 1962. Translated by Tamar March from Victor Cousin, ed. Oeuvres de Decartes, IV. (Paris 1824) pp. 349-363. Also in R. J. Herrnstein & E. G. Boring (Eds.) (1965) A source book in the history of psychology. (pp. 266-272). Cambridge, MA: Harvard University Press.
Eliot, T. S. (1943). Four Quartets, (p. 16). New York: Harcourt, Brace and Company.
Ellis, A. A. (1970). The essence of rational psychotherapy: A comprehensive approach to treatment. New York: Institute for Rational Living.
Epstein, R. (1984). Simulation research in the analysis of behavior. Behaviorism, 12, 41-59.
Epstein, R., Kirschnit, C., Lanza, R. P., & Rubin, L. (1981a). "Insight" in the pigeon: Antecedents and determinants of an intelligent performance. Nature, 308, 61-62.
Epstein, R., Lanza, R. P., & Skinner, B. F. (1980). Symbolic communication between two pigeons (Columba livia domestica). Science, 207, 543-545.
Epstein, R., Lanza, R. P., & Skinner, B. F. (1981b). "Self-awareness" in the pigeon. Science, 212, 695-696.
Epstein, R., & Medalie, S. (1983). The spontaneous use of a tool by a pigeon. Behavior Analysis Letter, 3, 241-247.
Epstein, R., & Skinner, B. F. (1981). The spontaneous use of memoranda by pigeons. Behavior Analysis Letters, 1, 241-246.
Findley, J. D. (1962). An experimental outline for building and exploring multi-operant repertoires. Journal of the Experimental Analysis of Behavior, 5, 113-166.
Fischman, M. W. (1977). Evaluating the abuse potential of psychotropic drugs in man. In T. Thompson & K. R. Unna (Eds.), Predicting dependence liability of stimulant and depressive drugs (pp. 261-284). Baltimore: University Park Press.
Fischman, M. W., Schuster, C. R., Resnekov, L., Shick, J. F. E., Krasnegor, N. A., Fennell, W., & Freedman, D. X. (1976). Cardiovascular and subjective effects of intravenous cocaine administration in humans. Archives of General Psychiatry, 33, 983-989.
France, C. P., & Woods, J. H. (1985). Opiate agonist-antagonist interactions: Application of a three-key drug discrimination procedure. Journal of Pharmacology and Experimental Therapeutics, 234, 81-89.
Freud, S. (1895). Project for a scientific psychology. In Standard edition of the complete works of Sigmund Freud: Vol 1. London: Hogarth Press, 1966.
Furness, W. H. (1916). Observations of the mentality of chimpanzees and orangutans. Proceeding of the American Philosophical Society, 55, 281-290.
Gallup, G. G. (1977). Self-recognition in primates. American Psychologist, 32, 329-338.
Gardner, R. A., & Gardner, B. T. (1984). A vocabulary test for chimpanzees (pan troglodytes). Journal of Comparative Psychology, 98, 381-404.
Gardner, R. A., Gardner, B. T., & VanCantfort, T. E. (1989). Teaching sign language to chimpanzees. Albany: State University of New York Press.
Goldberg, S. R., & Stolerman, I. P. (1986). Behavioral analysis of drug dependence. Orlando: Academic Press.
Gray, J. A. (1982). Precis of The neuropsychology of anxiety: An enquiry into the functions of the septo-hippocampal system. The Behavioral and Brain Sciences, 5, 469-534.
Griffin, D. R. (1976). The question of animal awareness: Evolutionary continuity of mental experience. New York: Rockefeller University Press.
Griffin, D. R. (1982). Animal mind-Human mind. Berlin: Dahlem Konferenzen.
Griffin, D. R. (1984). Animal thinking. Cambridge, MA: Harvard University Press.
Griffiths, R. R., & Balster, R. L. (1979). Opioids: Similarity between evaluations of subjective effects and animal self-administration results. Clinical Pharmacology and Therapeutics, 25, 611-617.
Griffiths, R. R., Bigelow, G. E., & Henningfield, J. E. (1985). Similarities in animal and human drug taking behavior. In N. K. Mello (Ed.), Advances in substance abuse: Behavioral and biological research. Greenwich, CT: JAI Press.
Griffiths, R. R., & Roache, J. D. (1984). Abuse liability of benzodiazepines: A review of human studies evaluating subjective and/or reinforcing effects. In D. E. Smith, & D. R. Wesson (Eds.), Benzodiazepines: Standards of use in clinical practice. Lancaster, England: MTP Press.
Griffiths, R. R., Roache, J. D., Ator, N. A., Lamb, R. J., & Lukas, S. E. (1985). Similarities in reinforcing and discriminative effects of diazepam, triazolam, and pentobarbital in animals and humans. In L. S. Seiden & R. L. Balster (Eds.) Behavioral pharmacology: The current status (pp. 419-432).New York: Alan R. Liss, Inc.
Gunderson, K. (1985). Mentality and machines (second edition). Minneapolis: University of Minnesota Press.
Hayes, C. (1951). The ape in our house. New York: Harper.
Hill, H. E., Haertzen, C. A., Wolbach, A. B., & Miner, E. J. (1963). The Addiction Research Center Inventory: Standardization of scales which evaluate subjective effects and morphine, amphetamine, pentobarbital, alcohol, LSD-25, parahexyl and chlorpromazine. Psychopharmacologia 4, 167-183.
Holtzman, S. G. (1982). Discriminative properties of opioids in the rat and squirrel monkey. In Colpaert, F. C. and Slanger, J. L. (eds) Drug discrimination: Applications in CNS pharmacology. Amsterdam: Elsevier.
Honig, W. R., & Staddon, J.E.R. (1977). Handbook of operant behavior. Englewood Cliffs, NJ: Prentice Hall.
Hull, C. L. (1933). Differential habituation to internal stimuli in the albino rat. Journal of Comparative Psychology, 16, 255-273.
Hull, C. L. (1943). Principles of behavior. New York: Appleton-Century-Crofts.
James, W. (1890). The principles of psychology. New York: Holt, Rinehart, & Winston.
Jewett, D., Schaal, D., Cleary, J., Thompson, T., & Levine, A. (1991). Discriminative effects of neuropeptide Y. Brain Research, 561, 166-168.
Johanson, C. E., & Uhlenhuth, E. H. (1980a). Drug preference and mood in humans: Diazepam. Psychopharmacology, 71, 269-273.
Johanson, C. E., & Uhlenhuth, E. H. (1980b). Drug preference and mood in humans: Amphetamine. Psychopharmacology, 71, 275-279.
Johanson, C. E., & Uhlenhuth, E. H. (1981). Drug preference and mood in humans: Repeated assessment of amphetamine. Pharmacology, Biochemistry and Behavior, 14, 159-163.
Johanson, C. E., & Uhlenhuth, E. H. (1982). Drug preference and mood in humans. Federation Proceedings, 41, 228-233.
Jung, C. G. (1959). The archetypes and the collective unconscious. In H. Read, M. Fordham & G. Adler (Eds.), Collected works. Princeton, NJ: Princeton University.
Keehn, J. D. (1986). Animal models for psychiatry. London: Routledge & Kegan Paul.
Keele, K. D. (1983). Leonardo da Vinci's Elements of the science of man. New York: Academic Press.
Kellogg, W. N., & Kellogg, L. A. (1933). The ape and the child. New York: McGraw Hill.
Kendall, P. C., & Hollon, S. D. (1981). Assessment strategies for cognitive behavioral interventions. New York: Academaic Press.
Kendler, H. H. (1946). The influence of simultaneous hunger and thirst drives upon the learning of two opposed spatial responses of the white rat. Journal of Experimental Psychology, 36, 212-220.
Keogh, D., Whitman, T., Beeman, D., & Halligan, K. (1987). Teaching interactive signing in a dialog situation to mentally retarded individuals. Research in Developmental Disabilities, 8, 39-53.
Kohler, W. (1925). The mentality of apes. London: Routledge and Kegan Paul.
Kordig, C. R. (1978). Discovery and justificatiton. Philosophy of Science, 45, 110-117.
Leeper, R (1935). The role of motivation in learning: A study of the phenomenon of differential motivational control of the utilization of habits. Journal of Genetic Psychology, 46, 3-40.
Lubbock, J. (1884). Teaching animals to converse. Nature, 2, 547-548.
Lubinski, D., & Dawis, R. V. (1992). Aptitudes, skills and proficiencies. In M. D. Dunnette & L. M. Hough (Eds.) Handbook of industrial & organizational psychology (pp. 1-59) (Second Edition). Palo Alto: Consulting Psychologists Press.
Lubinski, D., & MacCorquodale, K. (1984). "Symbolic communication" between two pigeons (Columba livia) without unconditioned reinforcement. Journal of Comparative Psychology, 98, 372-380.
Lubinski, D., & Thompson, T. (1986). Functional units of human behavior and their integration: A dispositional analysis. In T.
Thompson & M. D. Zeiler (Eds.), Analysis and integration of behavioral units (pp. 275-314). Hillsdale, NJ: Earlbaum.
Lubinski, D., & Thompson, T. (1987). An animal model of the interpersonal communication of interoceptive (private) states. Journal of the Experimental Analysis of Behavior, 48, 1-15.
Lykken, D. T. (1957). A study of anxiety in the sociopathic personality. Journal of Abnormal and Social Psychology, 55, 6-10.
Lykken, D. T. (1968). Statistical significance in psychological research. Psychological Bulletin, 70, 151-159.
Lykken, D. T. (1982). Research with twins: The concept of emergenesis. Psychophysiology, 19, 361-373.
Lykken, D. T. (1984). Detecting deception in 1984. Minnesota Psychologist, Summer, 9-13.
Lyons, W. (1986). The disappearance of introspection. Cambridge, MA: MIT Press.
MacCorquodale, K. (1969). B. F. Skinner's Verbal Behavior: A retrospective appreciation. Journal of the Experimental Analysis of Behavior, 12, 831-841.
MacCorquodale, K. (1970). On Chomsky's review of Skinner's Verbal Behavior. Journal of the Experimental Analysis of Behavior, 13, 83-99.
Mackintosh, N. J. (1987). Animal minds. In C. Blackemore & S. Greenfield (Eds.). Mindwaves: Thoughts on intelligence, identity, and consciousness (pp. 111-120). Oxford: Basil Blackwell.
Mackintosh, N. J. (1974). The psychology of animal learning. New York: Academic Press.
Martin, W. R., Sloan, J. W., Sapira, J. D., & Jasinski, D. R. (1971). Physiologic, subjective and behavioral effects of amphetamine, metamphetamine, ephedrine, phenmetrazine, and methylphenidate in man. Clinical Pharmacology and Therapeutics, 12, 245-258.
Maslow, A. H. (1962). Toward a psychology of being. Princeton, NJ: Van Nostrand.
McNair, D., Lorr, M., & Droppleman, L. (1971). Manual: Profile of mood states. San Diego, CA: Educational and Industrial Testing Service.
Mead, G. H. (1934). Mind, self and society. Chicago: University of Chicago Press.
Miller, G. A. (1967). The psychology of communication. New York: Basic Books.
Miller, G. A. (1990). The place of language in scientific psychology. Psychological Science, 1, 7-14.
Miller, S. L. (1953). A production of amino acids under possible primitive earth conditions. Science, 117, 528-529.
Miller, S. L., & Orgel, L. E. (1973). The origins of life on earth. Englewood Cliffs, NJ: Prentice-Hall.
Moore, J. (1980). On behaviorism and private events. Psychological Record, 30, 459-475.
Moore, J. (1984). On privacy, causes and contingencies. The Behavior Analyst, 7, 3-16.
Mounin, G. (1976). Language, communication and language. Current Anthropology, 17, 1-7.
Natsoulas, T. (1983). Perhaps the most difficult problem faced by behaviorism. Behaviorism, 11, 1-26.
Natsoulas, T. (1985). The treatment of conscious content: Disorder at the heart of radical behaviorism. Methodology and Science, 18, 81-103.
Nowlis, V. (1953). The development and modification of motivational systems in personality. In M. R. Jones (Ed.), Current theory and research in motivation. Lincoln: University of Nebraska Press.
Nowlis, V. (1970). Mood: Behavior and experience. In M. B. Arnold (Ed.), Feelings and emotions (pp. 261-277). New York: Academic Press.
Nowlis, V., & Nowlis, H. H. (1956). The description and analysis of mood. Annals of the New York Academy of Sciences, 65, 345-355.
Overton, D. A. (1971). Discriminative stimulus functions of drugs. In T. Thompson & R. Pickens (Eds.), Stimulus properties of drugs (pp. 87- 110). New York: Appleton-Century-Crofts.
Pavlov, I. P. (1927). Conditioned reflexes. Oxford, UK: Oxford University Press.
Pinker, S., & Bloom, P. (1990). Natural language and natural selection. Behavioral and Brain Sciences, 13, 707-784.
Premack, D. (1970). A functional analysis of language. Journal of the Experimental Analysis of Behavior, 14, 107-125.
Premack, D. (1976). Intelligence in ape and man. Hillsdale, NJ: Erlbaum.
Reichle, J., Lindamood, L., & Sigafoos, J. (1986). The match between reinforcer class and response class: Its influence on communication intervention strategies. Journal of the Association for Persons with Severe Handicaps, 11, 131-135.
Rogers, C. R. (1961). On becoming a person. Boston: Houghton Mifflin.
Rumbaugh, D. M. (1977). Language learning by a chimpanzee: The Lana Project. New York: Academic Press.
Rumbaugh, D. M., & Gill, T. (1976). The mastery of language-type skills by the chimpanzee (Pan). In S. R. Harnad, H. O. Steklis, & J. Lancaster (Eds.), Origins and evolution of language and speech (Vol 280). New York: New York Academy of Sciences.
Salzinger, K. (1973). Animal communication. In D. A. Dewsbury & D. A. Rethlingshafer (Eds.), Comparative psychology: A modern survey (pp. 161-193). New York : McGraw-Hill.
Savage-Rumbaugh, E. S. (1984a). Verbal behavior at a procedural level in the chimpanzee. Journal of the Experimental Analysis of Behavior, 41, 223-250.
Savage-Rumbaugh, E. S. (1984b). Acquisition of functional symbol usage in apes and children. In H. L. Roitblat, T. G. Bever, & H.S. Terrace (Eds.), Animal cognition (pp. 291-310). Hillsdale, NJ: Lawrence Erlbaum Associates.
Savage-Rumbaugh, E. S. (1986). Ape language: From conditioned response to symbol. New York: Columbia University Press.
Savage-Rumbaugh, E. S., & Rumbaugh, D. M. (1980). Requisites of symbolic communication - or, are words for birds? The Psychological Record, 30, 305-318.
Savage-Rumbaugh, E. S., Pate, J. L., Lawson, J., Smith, S. T., & Rosenbaum, S. (1983). Can a chimpanzee make a statement? Journal of Experimental Psychology: General, 112, 457-492.
Savage-Rumbaugh, E. S., Rumbaugh, D. M., & Boysen, S. (1978). Intentional symbolic communication in chimpanzee (Pan troglodytes). Science, 201, 641-644.
Savage-Rumbaugh, E.S., Rumbaugh, D. M., Smith, S. T., & Lawson, J. (1980). Reference: The linguistic essential. Science, 210, 922-925.
Schnaitter, R. (1978). Private causes. Behaviorism, 6, 1-12.
Schreibman, L., & Lovaas, O. I. (1973). Overselective response to social stimuli by autistic children. Journal of Abnormal Child Psychology, 1, 152-168.
Schuh, K. J., Schaal, D. W., Cleary, J., Thompson, T., Levine, A. s. (1991). Discrimination and generalization of the stimulus effects of feeding-induced operations. Society for Neuroscience Abstracts, 17, Part 1, Abstract no. 196.19, p. 496.
Schuster, C. R., & Brady, J. V. (1964). The discriminative control of a food-reinforced operant by interoceptive stimulation. Pavlov Journal of Higher Nervous Activity, 14, 448-458.
Schuster, C. R., & Johanson, C. E. (1988). Relationship between the discriminative stimulus properties and subject effects of drugs. In F.C. Colpaert & R. L. Balster (Eds.), Transduction mechanisms of drug stimuli (pp. 161-175). Berlin: Springer-Verlag.
Schuster, C. R., Fischman, M. W., & Johanson, C. E. (1981). Internal stimulus control and subjective effects of drugs. In T. Thompson & C. E. Johanson (Eds.), Behavioral pharmacology of drug dependence: NIDA Research Monograph, 37, 116-129.
Segal, E. F. (1977). Toward a coherent psychology of language. In W. K. Honig & J. E. R. Staddon (Eds.), Handbook of operant behavior (pp. 628-653). Englewood Cliffs, NJ: Prentice-Hall.
Seiden, L. S., & Balster, R. L. (1985). Behavioral pharmacology: The current status. New York: Alan R. Liss, Inc.
Skinner, B. F. (1945). The operational analysis of psychological terms. Psychological Review, 52, 270-277; 291-294.
Skinner, B. F. (1953). Science and human behavior. New York: Macmillan.
Skinner, B. F. (1957). Verbal behavior. New York: Appleton-Century- Crofts.
Skinner, B. F. (1969). Contingencies of of reinforcement: A theoretical analysis. New York: Appleton-Century-Crofts.
Skinner, B. F. (1974). About behaviorism. New York: Knopf.
Spence, K. W. (1956). Behavior theory and conditioning. New Haven, CT: Yale University Press.
Tellegen, A. (1985). Structures of mood and personality and their relevance to assessing anxiety, with emphasis on self-report. In A. H. Tuma & J. D. Maser (Eds.), Anxiety and the anxiety disorders (pp. 681-706). Hillsdale, NJ: Lawrence Earlbaum Associates.
Tellegen, A., Lykken, D. T., Bouchard, T. J., Wilcox, K.J., Segal, N.L., & Rich, S. (1988). Personality similarity in twins reared apart and together. Journal of Personality and Social Psychology, 54, 1031-1039.
Terrace, H. S. (1979). Nim. New York: Knopf.
Terrace, H. S. (1985). In the beginning was the "name." American Psychologist, 40, 1011-1028.
Terrace, H. S., & Bever, T. G. (1976). What might be learned from studying language in the chimpanzee? The importance of symbolizing oneself. Annals of the New York Academy of Sciences, 280, 579-588.
Thompson, T., & Lubinski, D. (1986). Units of analysis and kinetic structure of behavioral repertoires. Journal of the Experimental Analysis of Behavior, 46, 219-242.
Thompson, T., & Pickens, R. (1971). Stimulus properties of drugs. New York: Appleton.
Thompson, T., & Unna, K. R. (1977). Predicting dependence liability of stimulant and depressive drugs. Baltimore: University Park Press.
Thorndike, E. L. (1911). Animal intelligence: Experimental studies. New York: Macmillan.
Titchener, E. B. (1899). An outline for psychology. New York: Macmillan.
Tolman, E. C. (1932). Purposive behavior in animals and men. New York: Century.
Tolman, E., & Honzik, C. H. (1930). Introduction and removal of reward, and maze performance in rats. University of California Publications in Psychology, 4, 257-275.
Tuma, A. H., & Masur, J. D. (1985). Anxiety and the anxiety disorders. Hillsdale, NJ: Lawrence Erlbaum Associates.
Uhlenhuth, E. H., Johanson, C. E., Kilgore, & Kobasa, S. C. (1981). Drug preference and mood in humans: Preference for amphetamine and subject characteristics. Psychopharmacology, 74, 191-194.
Urey, H. C. (1952). The possible concentrations of organic compounds in the primitive oceans. Proceedings of the National Academy of Science, 38, 351-362.
Walker, S. (1983). Animal thought. London: Routledge and Kegan Paul.
Watson, D., & Pennebaker, J. W. (1989). Health complaints, stress, and distress: Exploring the central role of negative affectively. Psychological Review, 96, 234-254.
Winokur, S. (1976). A primer of verbal behavior: Englewood Cliffs, NJ: Prentice-Hall.
Woodruff, G. Premack, D., & Kennel, K., (1978). Conservation of liquid and solid quantity by the chimpanzee. Science, 202, 991-994.
Woodruff, G., & Premack, D. (1981). Primitive mathematical concepts in the chimpanzee: Proportionality and numerosity. Nature, 293, 568-570.
Woods, J. H., Bertamio, A. M., Young, A., Essman, W. D., & Winger, G. (1988). In F. C. Colpaert & R. L. Balster (Eds) Transduction mechanisms of drug stimuli. (pp. 95-106) Berlin: Springer-Verlag.
Wundt, W. (1894). Human and animal psychology. New York: Macmillan.
Zeaman, D., & House, B. J. (1979). A review of attention theory. In N. R. Ellis (Ed.), Handbook of Mental Deficiency (2nd. Ed.) (pp. 63- 120). Hillsdale NJ: Lawrence Erlbaum Associates.
Zevon, M. A., & Tellegen, A. (1982). The structure of mood change: An ideographic/nomethetic analysis. Journal of Personality and Social Psychology, 43, 111-122.
Zuckerman, M., & Lubin, B. (1965). Normative data for the multiple adjective checklist. Psychological Reports, 16, 438.
Zuriff, G. E. (1979). Ten inner causes. Behaviorism, 7, 1-8.
Zuriff, G. E. (1985). Behaviorism: A conceptual reconstruction. New York: Columbia University Press.
We have profited greatly from conversations with and comments on the topics contained in this paper from: Robert L. Balster, Camilla P. Benbow, Keith Gunderson, Mark Egli, Chris Ellyn Johanson, M. Jackson Marr, Paul E. Meehl, Trevor W. Robins, and Ian Stolerman. This work was supported in part by a grant form the National Institute of Child Health and Human Development to the Institute for Disabilities Studies of the University of Minnesota (P30HD25710) and the National Institute of Mental Health, National Research Award N14257 to the Department of Psychology of the University of Illinois, Dr. Lawrence E. Jones, Program Director.
1. Kordig (1978) has argued that the distinction between discovery and justification is often ambiguous and that acquiring scientific knowledge involves three tiers of credibility: initial thinking, plausibility, and acceptability. Initial thinking occurs prior to plausibility and acceptance (i.e., in the context of discovery). Plausibility and acceptance require unassailable logic and good evidence; both concepts are analyzed in the context of justification. Plausibility is necessary but not sufficient for achieving acceptability. Concepts achieving the status of the latter must satisfy more stringent criteria, but the requirements for both levels of justification are of the same logical character. Plausibility proofs provide good reasons for attaching scientific merit to a posited entity or a particular interpretation of the data; when adequately conducted, they achieve the first level of Kordig's two- stage conceptualization of justification.
As a case in point, the "hot soup" theory of organic evolution achieved plausibility in the 1950's (Miller, 1953; Urey, 1952). Using geological information about the early inorganic properties of earth and conjectures about electrical storm activity in the earth's prebiotic atmosphere, Miller and Urey simulated conditions hypothesized to give rise to early life forms (see Miller & Orgel, 1973) producing amino and hydroxy acids, which are important antecedents of life. Although this did not confirm their hypothesis, it enhanced its plausibility.
2. There is actually an intriguing film of this performance available to educators: "Cognition, creativity, and behavior" (1982). Producers, N. Baxley. Champaign, IL: Research Press. .
3. Mands and tacts are neologisms used to describe two classes of verbal behavior in Skinner's (1957) analysis. Mands are verbal operants controlled by a state of deprivation or aversive stimulation and specify their reinforcer. (They are the most primitive form of verbal behavior in Skinner's analysis.) A tact is defined as "a verbal operant in which a response of a given form is evoked (or at least strengthened) by a particular object or event or property of an object or event" (Skinner, 1957, pp. 81-82). These terms were chosen because, technically, the "mander" was trained to emit mands, whereas the "tacter" was trained to emit tacts (for details see Lubinski & MacCorquodale, 1984).