The legacy of Guttman and Kalish (1956): Twenty-five years of research on stimulus generalization

This paper is a selective review of the methods, problems, and findings in the area of operant stimulus generalization over the 25 years since the publication of the original paper by Guttman and Kalish (1956) on discriminability and spectral generalization in the pigeon. The paper falls into five main sections, which encompass the main themes and problems stemming from the Guttman and Kalish work and its immediate successors. The first section addresses the relationship between stimulus generalization and stimulus control, as well as the variety of testing procedures and dependent variables used to measure generalization. The next section reviews the limited literature on the effects of early rearing on the generalization gradient. The relationship between discriminability among test stimuli and the slope of the spectral gradient is discussed in the third section, with emphasis upon recent reassessments of the pigeon's hue discriminability function. The fourth section reviews the topic of inhibitory stimulus control, one which developed with the discovery of the peak shift following intradimensional discrimination training. Problems of definition and measurement are discussed in conjunction with the gradient forms used to index inhibitory control. The last section is devoted to attentional effects and the two principal theories postulated to account for them. A survey of different attentional paradigms is provided and the possible role of constant irrelevant stimuli as a source of control is examined. A brief conclusion summarizes the contribution of the generalization technique toward an understanding of the nature and acquisition of stimulus control.     

Sample-stimulus discriminability and sensitivity to reinforcement in delayed matching to sample

Five pigeons were trained in a delayed matching-to-sample task with red and green stimuli. The retention interval between sample-stimulus presentation and the availability of the choice stimuli was varied between 0.01 s and 12 s within each session. The probability of food produced by correct-red and correct-green responses was varied across conditions. Sample-stimulus discriminability and response bias were measured at four different retention intervals. The results of these analyses showed an interaction between the discriminability of the sample stimuli and the control exerted by differential reinforcement. At longer retention intervals, sample discriminability decreased and sensitivity of choice behavior to changes in the red/green reinforcer ratio increased. An analogous relation has been reported in conditional discriminations in which the physical disparity of stimuli has been varied. This correspondence suggests that increasing the delay between presentation of one of two stimuli and an opportunity to respond discriminatively to it may be functionally similar to increasing the physical similarity of the two stimuli.     

Psychophysics of Remembering: The Discrimination Hypothesis

In a psychophysical approach to remembering, the events to be remembered are discriminated from other possibilities at the time of remembering, and not at the time of encoding or learning. The discrimination is specific to the retention interval at which remembering occurs, as shown by experiments demonstrating that discriminability and response bias are delay–specific. This article discusses a discrimination model for remembering that emphasizes the individual's history of learning about reward payoffs in similar experiences in the past. This model predicts the two characteristics of forgetting functions, initial discriminability and rate of forgetting.     

Divided stimulus control: a replication and a quantitative model

Four pigeons were trained on a conditional discrimination. The conditional stimuli were compounds of pairs of stimuli from two different dimensions, fast versus slow cycles of red or green stimuli, and short- versus long-duration presentations of these cycles. Across conditions, the probability of reinforcers for correctly responding to each dimension was varied from 0 to 1. Discriminability, measured by log d, for stimuli on a dimension increased as the relative frequency of reinforcers for that dimension increased, replicating the results of Shahan and Podlesnik (2006). Two further conditions showed that discriminability between stimuli on each dimension was unaffected by whether the stimuli on the other dimension varied or were constant. Finally, maximal discriminability was unchanged in a redundant-relevant cues condition in which either of the stimuli comprising a compound signaled the same correct response. Davison and Nevin's (1999) model provided an excellent quantitative account of the effect of relative reinforcer frequency on discriminability, and thus of the way in which divided stimulus control is itself controlled by relative reinforcement.     

On Stimulus Generalization and Stimulus Classes

Stimulus generalization has been defined as the spread of effect of reinforcement for responses emitted in the presence of one stimulus to different stimuli presented under extinction conditions. As a result of stimulus generalization, novel stimuli come to exert stimulus control over members of the response class. Studies in the applied behavior analysis literature, however, have reported experimental preparations that included prompting and reinforcement procedures during what were claimed to be stimulus generalization conditions. These studies violated the procedural requirement that stimulus generalization be tested under extinction conditions. Responses that come under the control of a class of stimuli may do so by direct training or by stimulus generalization. It is desirable for organisms to respond in the presence of members of an appropriately constructed stimulus class, but we should understand the mechanism of entry into the class by its members. If inaccurate claims of stimulus generalization are made when training procedures are used in the ostensible generalization conditions, the robustness of the original training procedures will be over estimated. By adhering to the operational requirements of behavioral definitions, we could better understand the power and limits of our educational and training procedures.     

A clarification of continuous repertoire development

The key-peck response of five pigeons was reinforced on a schedule whenever the interval between pecks at two response keys was between 1.0 and 2.33 seconds in the presence of a 2,500-Hertz tone or between 4.66 and 6.0 seconds in the presence of a 1250-Hertz tone. There was no tendency for responses of intermediate duration to occur when test tones of intermediate frequency were presented. This result clarifies a previous finding using a similar procedure but with a visual intensity stimulus dimension.     

Wavelength generalization and discrimination in the pigeon

This three-part study describes wavelength generalization gradients around a series of training wavelengths ranging from 480 to 645 om. Luminance was controlled for the pigeon’s spectral sensitivity. The response measure was probability of keypecking during a 2-sec stimulus presentation. Both an extinction procedure, where stimulus wavelengths occurred in 15-nm steps, and a maintained discrimination procedure, where step size was 2 to 4 nm, were used to obtain gradients. During a portion of the maintained discrimination procedure, new luminances were introduced, so that the effect of luminance level on gradient slope could be examined. Comparison of the resulting functions across training wavelengths revealed: (1) consistent differences in gradient slope in different spectral regions and (2) an increase in slope with luminance increase. The findings are related to recent electroretinographic wavelength contrast data and to psychophysical measures of wavelength discriminability.     

General attentiveness effects of discriminative training

Using a design that permitted the simultaneous assessment of intra-, inter-, and extradimensional effects of discriminative training, the generality of discriminative effects that have been said to reflect increases in “general attentiveness” was assessed. Pigeons received either discriminative training with two stimuli correlated with reinforcement and one stimulus correlated with nonreinforcement, or nondifferential reinforcement (control) training. One positive stimulus was part of an intradimensional task and the other was not. After training, generalization tests were conducted to assess stimulus control along several dimensions. Discriminative training resulted in increased control along dimensions of the positive stimulus involved in the intradimensional task, but not along any dimensions of the other positive stimulus. The results suggested that discriminative training leads to increases in attention that are neither so general as suggested by the “general attentiveness” view nor so specific as to be revealed solely by intradimensional effects.     

Pharmacological Shock Therapy as a Psycho-Biological Problem

Five mothers of 50 newborn rat pups were maintained on either standard lab chow or a diet composed of the same chow with onion powder added. Taste preference tests administered during weaning indicated that the pups whose mother was fed the onion chow showed a preference for onion diet, but that such preference dissipated across the five day testing period. A control group of pups who were exposed to any onion diet cues available through channels other than the mother's milk did not show a preference for onion diet. It was concluded that exposure to onion taste in mother's milk was responsible for an initial preference for onion diet.     

A discrimination analysis of training-structure effects on stimulus equivalence outcomes.

Experiments designed to establish stimulus equivalence classes frequently produce differential outcomes that may be attributable to training structure, defined as the order and arrangement of baseline conditional discrimination training trials. Several possible explanations for these differences have been suggested. Here we develop a hypothesis based on an analysis of the simple simultaneous and successive discriminations embedded in conditional discrimination training and testing within each of the training structures that are typically used in stimulus equivalence experiments. Our analysis shows that only the comparison-as-node (many-to-one) structure presents all the simple discriminations in training that are subsequently required for consistently positive outcomes on all tests for the properties of equivalence. The sample-as-node (one-to-many) training structure does not present all the simple discriminations required for positive outcomes on either the symmetry or combined transitivity and symmetry (equivalence) tests. The linear-series training structure presents all the simple discriminations required for consistently positive outcomes on tests for symmetry, but not for symmetry and transitivity combined (equivalence) or transitivity alone. Further, the difference in the number of simple discriminations presented in comparison-as-node training versus the other training structures is larger when the intended class size is greater than three or the number of classes is larger than two. We discuss the relevance of this analysis to interpretations of stimulus equivalence research, as well as some methodological and theoretical implications.     

Relational learning in a context of transposition: a review

In a typical transposition task, an animal is presented with a single pair of stimuli (for example, S3+ S4−, where plus and minus denote reward and nonreward and digits denote stimulus location on a sensory dimension such as size). Subsequently, an animal is presented with a testing pair that contains a previously reinforced or nonreinforced stimulus and a novel stimulus (for example, S2–S3 and S4–S5). Does the choice of a novel S2 instead of previously reinforced S3 in a testing pair S2–S3 indicate that the animal has learned a relation (i.e., “select smaller”)? This review of empirical evidence and theoretical accounts shows that an organism''s behavior in a transposition task is undoubtedly influenced by prior reinforcement history of the training stimuli (Spence, 1937). However, it is also affected by two other factors that are relational in nature—a similarity of two testing stimuli to each other and an overall similarity of the testing pair as a whole to the training pair as a whole. The influence of the two latter factors is especially evident in studies that use multiple pairs of training stimuli and a wide range of testing pairs comprising nonadjacent stimuli (Lazareva, Miner, Young, & Wasserman, 2008; Lazareva, Wasserman, & Young, 2005). In sum, the evidence suggests that both prior reinforcement history and relational information affect an animal''s behavior in a typical transposition task.     

Shifts in postdiscrimination gradients within a stimulus dimension based on bilateral facial symmetry

A shift in generalization gradients away from S+ and towards stimuli on the opposite end of the stimulus dimension from S- is a well established phenomenon in the laboratory, occurring with humans and nonhumans and with a wide range of stimuli. The phenomenon of gradient shifts has also been observed to have an analogous relationship to a variety of apparent biases in preference observed in the natural environment. One way to examine the validity of such analogies is by examining whether gradient shifts can be observed with complex and naturalistic stimuli. In the present experiment, undergraduates were trained to discriminate between faces that varied in terms of relative bilateral facial symmetry (a stimulus dimension correlated with health and attractiveness). Comparisons were made within subjects, using two sets of images. For both sets, the faces varied from naturally asymmetrical to symmetrical, and S+ was a face equidistant to the two extremes. With one set, S- was the naturally asymmetrical face, and with the other, S- was the symmetrical face. A peak shift was obtained in both conditions, although the effect was clearer in the aggregate than on the level of the individual. Overall, the results are consistent with the view that the processes responsible for gradient shifts in the lab are relevant to judgments made in the natural environment.     

Failure to find evidence of stimulus generalization within pictorial categories in pigeons

Pigeons' key pecks were reinforced in the presence of pictures from one of two categories, cats or cars. A single picture associated with reinforcement was used in Experiment 1, and 20 pictures from the same category were associated with reinforcement in Experiment 2. Pigeons then were presented with novel test pictures from the training category and from the other, previously unseen, category. During Session 1 of testing, pigeons pecked no more often at pictures from the reinforced category than at pictures from the previously unseen category. When pigeons were trained with pictures associated with reinforcement or its absence from different categories in Experiment 3, differential responding to novel pictures from different categories appeared during Session 1. These findings argue against a process of automatic stimulus generalization within natural categories and in favor of the position that category distinctions are not made until members of at least two categories are compared with one another.     

Discrimination of restrictions in sequentially-encoded auditory displays: Block designs

Sequential constraints, in the form represented by equal frequencies within blocks, were imposed upon finite-state statistical generators. The sequences were encoded in the form of interval-coded pulse trains and were transduced to sound by earphones. In discrimination tests, interval thresholds between the finite states provided a measure of the relative discriminability between different sequential constraints. These thresholds are shown to be quantitatively related to the difference in uncertainty in specification of the sequences. To a first approximation: Equal interstate interval thresholds are associated with equal differences in uncertainty.     

Direct remembering,mediated remembering,and atypical forgetting functions

Atypical forgetting functions have been demonstrated in several recent studies of delayed matching to sample, in which experimental conditions are altered partway through the retention interval. The forgetting functions are atypical in that accuracy or discriminability is not always a negatively accelerated monotonic function of increasing retention interval duration, but may increase at later times in the retention interval. Atypical forgetting functions reflect changes in levels of discrimination. A switch from a lower level to a higher level of discrimination, or vice versa, can occur at any time in the retention interval. The behavioral theories of remembering proposed by Nevin, Davison, Odum, and Shahan (2007), and White and Brown (2014), offer quantitative predictions of forgetting functions that differ in intercept or slope. Both theories are able to account for atypical forgetting functions, by assuming time‐independent changes in the mediating effect of attending to sample and comparison stimuli (in Nevin et al.'s model) or in the direct effect of the context of reinforcement of the conditional discrimination (in White & Brown's model). Despite differences in their main assumptions, the theories have an edge over any theory that assumes that forgetting is time‐dependent.     

The control of responding by auditory stimuli: interactions between different dimensions of the stimuli.

Sounds have position in addition to other dimensions, such as intensity and frequency. Position rapidly gains control of spatially contiguous responses and this may interfere with control of responding by other acoustic dimensions. These experiments investigated interference of a tone-noise discrimination by the discrimination of acoustic position. Squirrel monkeys were studied when responding was differentially reinforced in the presence of both spectral content (tone-noise) and positional differences between the stimuli, and when responding was differentially reinforced only in the presence of spectral differences. Under the first condition, responding rapidly came under the control of the position of the noise in the two monkeys tested. The position of the tone controlled responding in one monkey; in the second monkey, responding came under the control of the spectral content of the tone. Under the second condition, responding was initially under the control of the noise in all three monkeys tested. This persisted for the duration of the condition for two of the monkeys; in one monkey, responding came under the control of the spectral content of the noise. Under the second condition, responding was also initially under the control of the position of the tone for all monkeys, but control by spectral content of the tone relatively rapidly developed in two of three monkeys.     

Generalization of a tactile stimulus in horses.

Using horses, we investigated the control of operant behavior by a tactile stimulus (the training stimulus) and the generalization of behavior to six other similar test stimuli. In a stall, the experimenters mounted a response panel in the doorway. Located on this panel were a response lever and a grain dispenser. The experimenters secured a tactile-stimulus belt to the horse's back. The stimulus belt was constructed by mounting seven solenoids along a piece of burlap in a manner that allowed each to provide the delivery of a tactile stimulus, a repetitive light tapping, at different locations (spaced 10.0 cm apart) along the horse's back. Two preliminary steps were necessary before generalization testing: training a measurable response (lip pressing) and training on several reinforcement schedules in the presence of a training stimulus (tapping by one of the solenoids). We then gave each horse two generalization test sessions. Results indicated that the horses' behavior was effectively controlled by the training stimulus. Horses made the greatest number of responses to the training stimulus, and the tendency to respond to the other test stimuli diminished as the stimuli became farther away from the training stimulus. These findings are discussed in the context of behavioral principles and their relevance to the training of horses.     

Stagewise multidimensional visual discrimination by pigeons

We trained six pigeons in a stagewise Multiple Necessary Cues (MNC) go/no‐go task to document the dynamics of discrimination learning involving increasingly complex visual stimuli. The compound stimuli were composed from four dimensions, each of which could assume either of two extreme values or their intermediate value: Shape, Size, Line Orientation, and Brightness. Starting with a stimulus composed entirely from intermediate values, we replaced those values with each of the two extreme dimensional values in four successive stages, thereby increasing the stimulus set from 2 in Stage 1 to 16 in Stage 4. In each stage, only one combination of values signaled food (S⁺), whereas the remaining combinations did not (S^?s). We calculated the rate of pecking during the first 15 s of each stimulus presentation and, in any given stage, training continued until the rate of responding to all of the S^?s was less than 20% of the rate of responding to the S⁺. All pigeons successfully acquired the final discrimination, suggesting that they attended to all of the dimensions relevant for the discrimination. We also replicated the key results of prior MNC studies: (1) the number of extreme dimensional values in each stage was positively related to the amount of training required for pigeons to acquire the discrimination; (2) attentional tradeoffs were most often observed when three or four dimensions were being trained; and (3) throughout training, the number of dimensional values in the S^?s that differed from the S⁺ was positively related to their discriminability from S⁺.     

质地密度逐渐变化对触觉识别的作用

研究关于输入、本体感受和输出对表面触觉认知的影响,旨在探索运用触觉识别质地密度逐渐变化的表面时,理论关于表面粗糙度恒定的情况下,主动手指运动产生的信息不影响触觉认知的解释是否仍然成立。实验采用表面质地密度呈正态曲线式逐渐变化,组成正负两组变化方向。正方向变化为表面质点密度向中逐渐增大,与手指触摸运动的加减速方式一致;负方向变化为表面质点密度向中逐渐减小,与手指触摸运动的加减速方式相反;另外包括一个表面密度不变的平面。让被试辨别表面质地密度变化的方向,结果发现被试能够很好地判别表面质地密度变化的方向,但在判别正向变化的表面时显得更准确;当表面密度没有变化时,被试倾向于判别为正向变化。这一发现没有支持表面密度恒定的触觉认知解释模型。与常规直觉相反,研究还发现即时反馈并没有改进触觉识别,反而降低了辨别的正确率。质地密度逐渐变化的触觉识别在变化的大小、准确和自信心方面都显示了不同,但没有发现认知学习作用该触觉任务辨别的过程。基于这些发现,文章讨论了触觉质地感知和有关存在的理论解释模式。