A molar view of goal direction and habit.

When we treat behavior within an evolutionary framework and as temporally extended flow, two fundamental questions arise: (a) What is an organism? and (b) What is behavior? An organism is a process that stays intact by constantly exchanging energy with the environment. It takes in resources and puts out waste. The behavior of an organism consists of those process parts of the organism process that make up the exchange. These activities serve the function of reproducing, which generally depends on surviving. Surviving and reproducing depend on responding to phylogenetically important events (PIEs). A PIE induces activities that enhance or mitigate the PIE. Organisms respond not only to a PIE but also to events that covary with the PIE. Both activities and environmental features may covary with a PIE. When either type of covariance is introduced to an organism, behavior adapts over time. The early stages of adaptation constitute what researchers call "goal direction," and the later stages constitute what researchers call "habit." Behavior and environment constitute a dynamic system, and manipulations of the covariances and environmental features of the system allow many experimental interventions. This molar approach allows experiments on goal direction and habit to be understood without appeal to everyday mentalistic terms.

Choice and rate-amount independence in a titration procedure.

The generalized matching law or Law of Allocation proposed by Baum (2018a, 2018b) potentially provides a broad conceptual framework within which to understand the allocation of time among activities. In its simplest form, the law incorporates power-function induction of activities by variables such as rate and amount of delivered inducers. Whether these variables affect allocation independently of one another is a central issue, because independence of the variables would allow simple multiplication of power functions and would make quantitative prediction simple too. The present experiment used a titration procedure to test the independence of rate and amount of food in determining pigeons' allocation of pecking between two keys. Amount ratio was varied within sessions to engender different peck ratios. Rate ratio was varied across two series of conditions. The results conformed to the predictions of the simple version of the Law of Allocation by strongly supporting independence of rate and amount. The Law of Allocation may have broad application for understanding activities in natural settings and everyday life.

Rate matching, probability matching, and optimization in concurrent ratio schedules.

Much research has documented rate matching in concurrent variable-interval schedules, but comparatively little research has examined performance in concurrent variable-ratio schedules, except in discrete-trials procedures that sometimes produce probability matching. One should expect that the two types of schedule would result in different performances, because ratio schedules cannot improve with time the way interval schedules do; ratio schedules lack the temporal dynamics of interval schedules. The present experiment exposed rats to concurrent variable-ratio schedules. Seven unsignaled components were presented in random order within each daily session, with probability ratios ranging from 1:8 to 8:1. Three conditions were studied that varied the overall probability of food while leaving probability ratios the same. Choice appeared to conform to probability matching, because sensitivities in the rate-matching relation were close to 0.5, whereas sensitivities to probability ratio were close to 1.0. The sensitivities alone, however, could not confirm probability matching, because undermatching to rate occurs often. Analyses at smaller time scales supported the interpretation of probability matching. In particular, control by food deliveries was highly local in concurrent variable-ratio schedules, in contrast with concurrent variable-interval schedules, in which control is extended. Activity continued to switch between alternatives throughout components, contradicting optimal sampling theory.

Covariance, feedback, and discounting in ratio schedules.

Action-outcome relations in the everyday world are often looser than the relations programmed in the laboratory. Reinforcement theory addressed this looseness by theorizing about "delay of reinforcement." Response-reinforcer interval affects response rate and choice, but need not be thought of as delay. Experiments on temporal discounting are often said to reveal effects of delay also, but the intervals chosen differ from delays, because an outcome after a week, month, or year occurs after life has continued and many events have intervened. They are better called lag intervals. The present experiment studied the effects of varying lag interval on pigeons' responding on ratio-like schedules of food. In a continuous procedure, a lag interval separated pecking that earned food from delivery of the food earned. In one series of conditions, the relation was fixed equivalent to variable-ratio (VR) 40 while the lag interval varied. In a second series, the equivalent VR varied while the lag interval was fixed. The results accorded with matching theory, but with a factor for temporal discounting or loose feedback. The results suggest that both temporal discounting and probability discounting may be characterized within matching theory, but that temporal discounting requires a separate treatment from probability discounting.

Behavioral ephemera, difficult discriminations, and behavioral stability.

Every species possesses abilities for successfully interacting with its environment. These result from phylogeny. In the laboratory, one may arrange artificial conditions that thwart an organism's abilities. The result may be a "phenomenon." With sufficient training, however, the phenomenon may prove to be ephemeral, as the organism's basic abilities reassert themselves. Pigeons respond extremely well to differences and nondifferences in rate of obtaining food. This ability may be thwarted in a variety of ways, but the results tend to be ephemeral. An example is an experiment that pitted pigeons' preference for unimpeded responding against their ability to respond to food rate. In a concurrent-chains procedure, the terminal links were identical variable-interval schedules, but in one terminal link, every response produced a timeout. The duration of the timeout varied, and preference varied with it, but the relation vanished with training, in keeping with the equality of food rate across the 2 terminal links. Some other examples of "phenomena" that tend to disappear with sufficient training and sufficient variation in experimental parameters are behavioral contrast, conditioned reinforcement, resistance to extinction, and suboptimal choice. These "phenomena" depend on pigeons' failing to make difficult discriminations. They appear to be behavioral ephemera.

Matching, induction, and covariance with mixed response-contingent food and noncontingent food.

The multiscale molar view of behavior is based on three basic laws of behavior: the Law of Allocation, the Law of Induction, and the Law of Covariance. Experiments that mix response-contingent food with noncontingent food shed light on these three laws. Food, like other phylogenetically important events, induces various activities that compete in allocation. Quantitative accounts represent induction with power functions. These power functions define activities' competitive weights, and relative time allocation among activities matches relative competitive weight. Behavior-food covariance determines which activities are induced. Phylogenetic (behavior-fitness) covariance determines which adjunctive activities are induced. Ontogenetic covariance may be represented in feedback functions. Feedback functions for variable-interval schedules may be observed even when overlaid by noncontingent food deliveries. Equations derived from the three laws describe responding in experiments with mixed response-contingent and noncontingent food. Equations derived here accounted for responding in three data sets: (a) from Rachlin and Baum (1972); (b) a new data set in which overall food rate was fixed while the proportion of response-contingent and noncontingent food varied; and (c) a new data set in which food occurred according to various variable-interval schedules. The same pigeons served throughout. All results were accommodated by the derived equations.

Resistance to extinction versus extinction as discrimination.

The hypothesis that response strength might be measured by persistence of responding in the face of extinction was discredited in the 1960s because experiments showed that responding persists longer following intermittent reinforcers than following continuous reinforcers. Instead, researchers proposed that the longer persistence following intermittent reinforcers arises because intermittent reinforcement more closely resembles extinction-a discrimination theory. Attention to resistance to extinction revived because one observation seemed to support the persistence hypothesis: Following training on a multiple schedule with unequal components, responding usually persisted longer in the formerly richer component than in the formerly lean component. This observation represents an anomaly, however, because results with single schedules and concurrent schedules contradict it. We suggest that the difference in results arises because the multiple-schedule procedure, while including extensive training on stimulus discrimination, includes no training on discrimination between food available and food unavailable, whereas comparable single- and concurrent-schedule procedures include such training with repeated extinction. In Experiment 1, we replicated the original result, and in Experiment 2 showed that when the multiple-schedule procedure includes training on food/no-food discrimination, extinction following multiple schedules contradicts behavioral momentum theory and agrees with the discrimination theory and research with single and concurrent schedules.

Behavior, process, and scale: Comments on Shimp (2020), "Molecular (moment-to-moment) and molar (aggregate) analyses of behavior".

If we study the behavior of organisms, we must understand the ontological status of both "organism" and "behavior." A living organism maintains itself alive by constantly interacting with the environment, taking in energy and discarding waste. Ontologically, an organism is a process. Its interactions with the environment, which constitute its behavior, are processes also, because the parts of any process are themselves processes. Processes serve functions, and the function of a process must be part of its identity. A process, by definition, extends in time. Time is the fundamental and universal measure of behavior. All processes have the property of scale. Activities of an organism have parts that are themselves activities on a smaller time scale. Scale varies continuously, and behavior may be studied on as large or as small a time scale as seems necessary. When researchers refer to the "structure" of behavior, they refer to smaller-scale activities. Attaching a switch to a lever or key is convenient, but one should never confuse operation of a switch with a unit of behavior. Shimp's (2020) "molecular" measures are small-scale measures. The molecular view based on discrete events has outlived its usefulness and should be replaced by a multiscale molar paradigm.

Response-reinforcer contiguity versus response-rate-reinforcer-rate covariance in rats' lever pressing: Support for a multiscale view.

The multiscale molar view sees behavior as a flow, like a river, extended in time. Matching theory expresses the way activities compete for time. Relative time taken by any activity depends on relative induction. The present experiment tested matching theory applied to concurrent contingent and noncontingent food. As adjunctive activities that compete with operant activity, we recorded hopper head entries and presses on a lever near the food hopper that had no programmed consequences. Eight naïve rats were first exposed to a variable-time 60 s schedule, which across conditions was gradually transformed into a variable-interval 60 s schedule by increasing the proportion of food that was delivered contingent on pressing a lever far from the hopper. Another group of 4 rats that had been trained to press a lever near a food hopper were introduced in the second condition, in which one food delivery was contingent on far-lever pressing. We found induction following a power function to describe pressing on the far lever (operant activity). Matching theory combined with power-function induction also accounted for adjunctive activity. Results with single contingent food deliveries provided little support for the molecular view that behavior consists of discrete responses "strengthened" by immediately following reinforcers.

Matching theory and induction explain operant performance.

Matching theory is a general framework for understanding allocation of behavior among activities. It applies to choice in concurrent schedules and was extended to single schedules by assuming that other unrecorded behavior competes with operant behavior. Baum and Davison (2014) found that the competing activities apparently are induced by the "reinforcers" (phylogenetically important events, e.g., food) according to power functions. Combined with power-function induction, matching theory provides new equations with greater explanatory power. Four pigeons were exposed to conditions in which 7 different schedules of food delivery were presented within each experimental session. We replicated earlier results with variable-interval schedules: (a) a negatively accelerated increase of peck rate as food rate increased in the low range of food rates; (b) an upturn in pecking at higher rates; and (c) a downturn in pecking at extremely high food rates. When the contingency between pecking and food was removed, the food continued to induce pecking, even after 20 sessions with no contingency. A ratio schedule inserted in place of 1 variable-interval schedule maintained peck rates comparable to peck rates maintained by short interval schedules. We explained the results by fitting equations that combined matching theory, competition, and induction.

Relativity in Hearing and Stimulus Discrimination.

What does it mean to hear a sound? What does it mean to perceive anything? Sound has no objective reality, such as "vibration." Of two people together, one may hear a sound and one may not. We know only that their actions-their judgments-differ. Such comparison underlies all discriminations. In experiments on concept learning, for example, pigeons peck when they are shown a slide containing human beings and do not peck when the slide contains no humans. The experimenters judge beforehand whether the slides contain human beings or not, and the pigeons' concept of human being is determined by the comparison between the experimenters' judgments and the pigeons' pecking or not. Similarly, to tell which of two people is hearing or deaf, an observer that can hear must judge whether their behavior corresponds with the observer's judgments. In an experiment by Lubinski and Thompson (Behavioral and Brain Sciences, 16, 627-680, 1993), in which pigeons pecked at different keys depending on which of two different drugs they had received beforehand, the experimenters judged which drug had been injected, and the pigeons' pecking corresponded to the experimenters' judgments. If two persons' judgments differ, they can only resolve the difference by deciding that one of them is mistaken. If no one is there to hear a tree fall in the forest, from the point of view of a science of behavior, it made no sound.

Multiscale behavior analysis and molar behaviorism: An overview.

In the context of evolutionary theory, behavior is the interaction between the organism and its environment. Two implications follow: (a) behavior takes time; and (b) behavior is defined by its function. That behavior takes time implies that behavioral units are temporally extended patterns or activities. An activity functions as an integrated whole composed of parts that are themselves smaller-scale activities. That behavior is defined by its function implies that behavior functions to change the environment in ways that promote reproductive success. Phylogenetically important events (PIEs) are enhanced or mitigated by activities they induce as a result of natural selection. Induction explains all the phenomena that have traditionally been explained by reinforcement. This multiscale view replaces discrete responses and contiguity with multiscale activities and covariance. A PIE induces operant activity as a result of covariance in the form of a feedback relation between the activity and the PIE. A signal (conditional inducer) induces PIE-induced activities as a result of covariance between the PIE and the signal. In an ontological perspective, behavior is a process, and an activity is a process individual. For example, ontological considerations clarify the status of delay and probability discounting. A true natural science of behavior is possible.

Concurrent variable-interval variable-ratio schedules in a dynamic choice environment.

Most studies of operant choice have focused on presenting subjects with a fixed pair of schedules across many experimental sessions. Using these methods, studies of concurrent variable- interval variable-ratio schedules helped to evaluate theories of choice. More recently, a growing literature has focused on dynamic choice behavior. Those dynamic choice studies have analyzed behavior on a number of different time scales using concurrent variable-interval schedules. Following the dynamic choice approach, the present experiment examined performance on concurrent variable-interval variable-ratio schedules in a rapidly changing environment. Our objectives were to compare performance on concurrent variable-interval variable-ratio schedules with extant data on concurrent variable-interval variable-interval schedules using a dynamic choice procedure and to extend earlier work on concurrent variable-interval variable-ratio schedules. We analyzed performances at different time scales, finding strong similarities between concurrent variable-interval variable-interval and concurrent variable-interval variable- ratio performance within dynamic choice procedures. Time-based measures revealed almost identical performance in the two procedures compared with response-based measures, supporting the view that choice is best understood as time allocation. Performance at the smaller time scale of visits accorded with the tendency seen in earlier research toward developing a pattern of strong preference for and long visits to the richer alternative paired with brief "samples" at the leaner alternative ("fix and sample").

Selection by consequences, behavioral evolution, and the price equation.

Price's equation describes evolution across time in simple mathematical terms. Although it is not a theory, but a derived identity, it is useful as an analytical tool. It affords lucid descriptions of genetic evolution, cultural evolution, and behavioral evolution (often called "selection by consequences") at different levels (e.g., individual vs. group) and at different time scales (local and extended). The importance of the Price equation for behavior analysis lies in its ability to precisely restate selection by consequences, thereby restating, or even replacing, the law of effect. Beyond this, the equation may be useful whenever one regards ontogenetic behavioral change as evolutionary change, because it describes evolutionary change in abstract, general terms. As an analytical tool, the behavioral Price equation is an excellent aid in understanding how behavior changes within organisms' lifetimes. For example, it illuminates evolution of response rate, analyses of choice in concurrent schedules, negative contingencies, and dilemmas of self-control.

Allocation of speech in conversation.

In a replication and extension of Conger and Killeen's (1974) widely cited demonstration of matching in conversations, we evaluated nine participants' allocation of speech and gaze to two conversational partners. German speakers participated in two 90-min sessions in which confederates uttered approval on independent variable-interval schedules. In one of the sessions, confederates uttered approval contingent upon and contiguous with eye contact whereas in the other session approval was uttered independent of the participant's gaze. Several measures of participants' verbal behavior were taken, including relative duration and rate of speech and gaze. These were compared to confederates' relative rate of approval and relative duration and rate of talk. The generalized matching equation was fitted to the various relations between participants' behavior and confederates' behavior. Conger and Killeen's results were not replicated; participants' response allocation did not show a systematic relation to the confederates' relative rate of approval. The strongest relations were to overall talk, rather than approval. In both conditions, the participant talked more to the confederate who talked less-inverse or antimatching. Participants' gaze showed the same inverse relation to the confederates' talk. Requiring gaze to be directed toward a confederate for delivery of approval made no difference in the results. The absence of a difference combined with prior research suggests that matching or antimatching in conversations is more likely due to induction than to reinforcement.

Limits to preference and the sensitivity of choice to rate and amount of food.

Studies of choice holding food-amount ratio constant while varying food-rate ratio within sessions showed that local changes in preference depend on relative amount of food. The present study investigated whether sensitivity of choice to food-rate ratio and sensitivity to food-amount ratio are independent of one another when food-rate ratios are varied across sessions and food-amount ratios are varied within sessions. Food deliveries for rats' presses on the left and right levers were scheduled according to three different food-rate ratios of 1:1, 9:1, and 1:9; each food-rate ratio lasted for 106 sessions and was arranged independently of seven food-amount ratios (7:1, 6:2, 5:3, 4:4, 3:5, 2:6, and 1:7 food pellets) occurring within sessions in random sequence. Each amount ratio lasted for 10 food deliveries and was separated from another by a 60-s blackout. Sensitivity to rate ratio was high (1.0) across food deliveries. Sensitivity to amount ratio was low when food rates were equal across alternatives, but was high when rate ratio and amount ratio opposed one another. When rate ratio and amount ratio went in the same direction, choice ratio reached an elevenfold limit which reduced sensitivity to approximately zero. We conclude that three factors affect sensitivity to amount: (1) the limit to preference, (2) the equal effect on preference of amounts greater than four pellets, and (3) the absence of differential effects of switches in amount in the equal-rates (1:1) condition. Taken together, these findings indicate that rate and amount only sometimes combine independently as additive variables to determine preference when amount ratios vary frequently within sessions.

The role of induction in operant schedule performance.

Baum and Davison (2014b) showed that Baum's (2012) recasting of reinforcement as induction may be quantified by assuming that induction follows a power function of reinforcer rate. This power-function induction is readily integrated with theory based on the matching law. Herrnstein (1970) originally assumed background activities (BO) and their associated reinforcers ro to be constant, but ro should vary with BO. Further, power-function induction implies that BO should vary with reinforcer rate. Baum (1993) reported performance on a wide range of variable-ratio (VR) and variable-interval (VI) schedules. Pigeons' VR peck rate followed an inverted U-shaped relation, but VI peck rate separated into three ranges of food rate: low-to-moderate, moderate-to-high, and extremely high. As food rate increases, the concave downward relation in the low range reaches an inflection point and gives way to a concave upward relation in the higher range. At the extremes of food rate, VI peck rate decreases. A model based on competition between induced pecking and BO accounted for VI peck rate in the moderate to extreme range of food rates. Further research will account for all three ranges, either by integrating power-function induction with matching theory or with a model based on competition between induced activities.

Background activities, induction, and behavioral allocation in operant performance.

In experiments on operant behavior, other activities, called "background" activities, compete with the operant activities. Herrnstein's (1970) formulation of the matching law included background reinforcers in the form of a parameter rO, but remained vague about the activities (BO) that produce rO. To gain more understanding, we analyzed data from three studies of performance with pairs of variable-interval schedules that changed frequently in the relative rate at which they produced food: Baum and Davison (2014), Belke and Heyman (1994), and Soto, McDowell, and Dallery (2005). Results sometimes deviated from the matching law, suggesting variation in rO. When rO was calculated from the matching equation, two results emerged: (a) rO is directly proportional to BO, as in a ratio schedule; and (b) rO and BO depend on the food rate, which is to say that BO consists of activities induced by food, as a phylogenetically important event. Other activities unrelated to food (BN ) correspond to Herrnstein's original conception of rO and may be included in the matching equation. A model based on Baum's (Baum, 2012) concepts of allocation, induction, and contingency explained the deviations from the matching law. In the model, operant activity B, BO, and BN competed unequally in the time allocation: B and BO both replaced BN , BO replaced lever pressing (Soto et al.), and key pecking replaced BO (Baum & Davison). Although the dependence of rO and BO on food rate changes Herrnstein's (1970) formulation, the model preserved the generalized matching law for operant activities by incorporating power-function induction.

Choice with frequently changing food rates and food ratios.

In studies of operant choice, when one schedule of a concurrent pair is varied while the other is held constant, the constancy of the constant schedule may exert discriminative control over performance. In our earlier experiments, schedules varied reciprocally across components within sessions, so that while food ratio varied food rate remained constant. In the present experiment, we held one variable-interval (VI) schedule constant while varying the concurrent VI schedule within sessions. We studied five conditions, each with a different constant left VI schedule. On the right key, seven different VI schedules were presented in seven different unsignaled components. We analyzed performances at several different time scales. At the longest time scale, across conditions, behavior ratios varied with food ratios as would be expected from the generalized matching law. At shorter time scales, effects due to holding the left VI constant became more and more apparent, the shorter the time scale. In choice relations across components, preference for the left key leveled off as the right key became leaner. Interfood choice approximated strict matching for the varied right key, whereas interfood choice hardly varied at all for the constant left key. At the shortest time scale, visit patterns differed for the left and right keys. Much evidence indicated the development of a fix-and-sample pattern. In sum, the procedural difference made a large difference to performance, except for choice at the longest time scale and the fix-and-sample pattern at the shortest time scale.