Skip to main content

Same/different concept learning by primates and birds


Same/different abstract-concept learning experiments were conducted with two primate species and three avian species by progressively increasing the size of the training stimulus set of distinctly different pictures from eight to 1,024 pictures. These same/different learning experiments were trained with two pictures presented simultaneously. Transfer tests of same and different learning employed interspersed trials of novel pictures to assess the level of correct performance on the very first time of subjects had seen those pictures. All of the species eventually performed these tests with high accuracy, contradicting the long-accepted notion that nonhuman animals are unable to learn the concept of same/different. Capuchin and rhesus monkeys learned the concept more readily than did pigeons. Clark’s nutcrackers and black-billed magpies learned as readily as monkeys, and even showed a slight advantage with the smallest training stimulus sets. Those tests of same/different learning were followed by delay procedures, such that a delay was introduced after the subjects responded to the sample picture and before the test picture. In the sequential same/different task, accuracy was shown to diminish when the stimulus on a previous trial matched the test picture previously shown on a different trial. This effect is known as proactive interference. The pigeons’ proactive interference was greater at 10-s delays than 1-s delays, revealing time-based interference. By contrast, time delays had little or no effect on rhesus monkeys’ proactive interference, suggesting that rhesus monkeys have better explicit memory of where and when they saw the potential interfering picture, revealing better event-based memory.

Same/different concept learning

The ability to learn whether two (or more) stimuli are the same or different (an abstract concept) was proclaimed (130 years ago) to be the “very keel and backbone of our thinking” (William James 1890/1950, p. 459). David Premack stated that (nonhuman) animals (at least those without any so-called language training) did not have the ability to learn a same/different abstract concept (Premack, 1978, 1983; Premack & Premack, 1983). Premack’s challenge inspired researchers, and today abstract-concept learning is firmly established in a variety of diverse species (for some reviews of the issues see Katz & Wright 2021; Katz, Wright, & Bodily, 2007; Wright & Katz, 2006; Wright, Kelly, & Katz, 2018). Variables and processes important for abstract-concept learning include domain restriction, interference, observing response, perceptual symmetry, response outcomes, stimulus generalization, stimulus modality, and training set size. In this article, we emphasize the role of set size and interference in our two-item same/different task with a focus on functional relationships (Wright, 2013).

In a typical same/different task a subject is presented with two items that are different (e.g., orange ball and blue key) or the same (e.g., orange cube and orange cube). The subjects must make a response to indicate whether the two items are the same or different. After learning the task (if the subject is capable), the subject must successfully transfer to novel stimuli in order to demonstrate the task was solved using an abstract same/different concept. Many of the original studies that failed to find concept learning used small training sets. The issue with small training sets is they create proactive interference (PI). Specifically, small numbers of training stimuli result in confusion when a previous stimulus happened to be presented again as a test stimulus. Nevertheless, during trials where the test stimulus differed from the initial stimulus, a different response would be correct, but instead this situation tended to elicit a same response, which was incorrect. These particular types of errors are caused by proactive interference, because the experience of the stimulus presentation earlier in the session interfered with the subsequent presentation. Even humans succumb to proactive interference when a small number of repeating stimuli are used (as clearly shown by Keppel & Underwood, 1962). These confusion errors may reduce accuracy during the task, leading to the incorrect conclusion that animals may not be capable of abstract-concept learning. For example, high levels of PI may force subjects to separate trials in time in order to acquire the task, and this may lead to memorizing the stimuli. Such memorization has the potential to block abstract-concept learning (see Nakamura, Wright, Katz, Bodily, & Sturz, 2009). There is a catch though. While animals can learn a task with small stimulus sets, they may fail transfer tests, but if animals are trained with a large stimulus set, they may pass transfer tests but take a long time to learn the task and apply the abstract concept (see Wright, Cook, Rivera, Sands, & Delius, 1988). Given this “catch-22,” originally the approach was taken to start with a small training set and expand to larger training sets, while testing for the abstract concept through the expansion process. How animals learn to combat PI in this expansion procedure may vary depending on whether explicit (event based) or implicit (time based) memory is used to solve the task (Devkar & Wright, 2016; Wright, Katz, & Ma, 2012). In the following sections we present our set size expansion results for five different species using a simultaneous two-item same/different task followed by proactive interference tests using a sequential two-item same/different task to determine the locus of PI and how the animals combat PI.

Same/different abstract-concept learning by monkeys

Due to proactive interference resulting from using small numbers of repeating stimuli, we manipulated the number of training stimuli to determine how large a training set-size would need to be to show same/different abstract-concept learning and whether this criterion would differ across species. We began with a small number of training stimuli (eight-item training set). This small training set was then successively doubled (logarithmic scale) to a sufficiently large set size that produced full-concept learning, defined as novel-stimulus transfer accuracy being equivalent to (or above) accuracy with training trials. For large training set-sizes to be effective, the stimuli needed to be as distinctly different as possible. We used a large number (up to 1,024) of distinctly different travel-slide pictures (e.g., scenes, objects, animals, buildings, people; see Wright & Katz, 2006, for training stimuli of each set size, stimulus pairs used for testing transfer and concept learning, and control conditions used to evaluate alternative explanations). An advantage of using a substantial number of expanding training sets, via systematically varying set size, is that the set-size expansion procedures allow for a more powerful comparison across species than any individual training set-size or much fewer set sizes. That is, by determining performance at each set size one is able to construct a set-size function. These functions reveal the functional relationship between set size and performance. By constructing such functions across species tested in the same procedure one can then determine quantitative and qualitative differences and/or similarities. Hence, such functions revealed through set-size expansion have the advantage over testing at a single training set or fewer training set sizes that do not produce full concept learning (Katz & Wright, 2021; Wright, 2013).

Our first test with these procedures was conducted using a computerized task for presenting visual stimuli to three (experimentally naïve) capuchin (Cebus apella) monkeys (Wright, Rivera, Katz, & Bachevalier, 2003). Trials began with the simultaneous presentation of two pictures pseudorandomly selected from a small training set of eight pictures. The pictures were positioned one above the other. A white rectangle was also present in the lower right-hand corner (see Fig. 1, Same Trial and Different Trial). If the two pictures were the same, then a touch (choice) to the lower picture was correct; if the pictures were different, then a touch to the white rectangle was correct. Correct responses were reinforced with a banana pellet, followed by a 15-s inter-trial interval. Incorrect responses were not reinforced, and early in training incorrect responses were followed by a 15-s timeout with the incorrect trial repeated (i.e., correction procedure). Accuracy was based only on first trial performance. Each daily session contained 100 trials (50 same trials and 50 different trials). The capuchin monkeys learned this task in about 35 sessions (mean = 3,533 trials) and to a criterion of 80% correct or better on three consecutive sessions.

Fig. 1
figure 1

Examples of same and different trials. Capuchin monkeys were trained with both pictures and the white rectangle simultaneously presented as shown for the same and different trial. Rhesus monkeys touched the sample picture ten times, after which the lower picture and white rectangle were simultaneously presented. Pigeons, Clark’s nutcrackers, and black-billed magpies pecked the sample picture 20 times, after which the lower picture and white rectangle were simultaneously presented. For all species, a response to the bottom picture was correct on same trials, whereas a response to the white rectangle was correct on different trials. The other six pictures of the initial eight-item training set are shown at the bottom

Three rhesus monkeys (Macaca mulatta) were also presented with this same/different task, using identical procedures (Katz, Wright, & Bachevalier, 2002). Compared to the capuchin monkeys, only one rhesus monkey learned the task but required more training sessions (200 sessions). The other two rhesus monkeys showed little or no learning even after 250 sessions. Therefore, three additional experimentally naïve rhesus monkeys were trained to touch the sample ten times before being presented simultaneously with the upper (sample) item, the lower (test) item, and the white rectangle (Fig. 1), at which point they could make their choice response. With the ten sample-response requirement to the sample stimulus, all of these rhesus monkeys learned the eight-item set same/different task at a rate (mean = 3,766 trials) similar to capuchin monkeys (Katz et al., 2002).

Following eight-item set learning, both monkey species were tested for transfer to novel stimuli. Transfer sessions consisted of 90 training (baseline) trials and ten transfer trials (five same trials and five different trials). The pictorial stimuli shown during transfer trials were novel. There were six transfer test sessions, conducted on six consecutive daily sessions. Correct responses on transfer trials were reinforced identically to that of correct responses on baseline (training) trials. Incorrect responses were followed by the inter-trial interval and then the next trial. Transfer trials were pseudorandomly placed after trial 7 amongst baseline trials and transfer trials, and could not occur on consecutive trials. Trial sequences varied over transfer sessions. Following the initial transfer testing, the eight-item training set was doubled followed by training and doubling the training set three additional times. Transfer testing was conducted following training with 32-, 64-, and 128-item set sizes when performance accuracy was 85% correct (or greater) on a single training session (Fig. 2). Despite the response-requirement difference between monkey species, once the task was learned, the rhesus monkeys performed as accurately as capuchin monkeys that were not required to make additional sample-stimulus responses.

Fig. 2
figure 2

Mean percent correct performance of three rhesus monkeys (red) and three capuchin monkeys (orange) on training (baseline) trials (broken lines, filled symbols) and novel (transfer) trials (unbroken lines, unfilled symbols) with successively expanded training sets from eight to 128 picture stimuli. The dotted line at 50% correct is chance performance. Error bars are + one standard error of the mean

Neither of the monkey species showed novel-picture transfer significantly different from chance (50% correct) performance following training with the initial eight-item training set (Fig. 2). However, as the set-size was progressively increased to 32, 64, and 128 stimuli, transfer performance increased monotonically. Importantly, novel-stimulus transfer performance following training with the 128-stimulus set was not statistically different compared to baseline performance accuracy revealing that both monkey species had learned the same/different abstract concept. Control tests (not shown) with other rhesus monkeys were also conducted where the set size was not expanded, but the numbers of training sessions and transfer-testing sessions were otherwise matched to the experimental group (Katz et al., 2002). Transfer by this control group did not improve despite the same amount of training and transfer testing, thereby demonstrating that the manipulation of progressively expanding the training set and exponentially expanding the number of exemplars was the key variable for these rhesus monkeys to learn the same/different abstract concept.

Same/different abstract-concept learning by pigeons

Four pigeons (Columba livia) were trained and tested with procedures similar to those previously described for monkeys (Katz & Wright, 2006). Since pigeons peck more rapidly than monkeys touch, we required 20 pecks from the pigeons before they were presented with the test stimulus and white rectangle. Otherwise, trials and sessions were very similar to those used to train monkeys: 100 trial sessions (50 same, 50 different), the same training accuracy criteria, same transfer procedures with each of six transfer sessions containing 90 baseline (training) trials and ten novel (five same, five different) transfer trials, the same set-size expansions with training for three or more sessions with at least 85% correct. Training and testing were repeated six times for 8, 16, 32, 64, 128, 256, 512, and 1,024 item training sets, with the last three training sets being larger to maximize chances of obtaining full abstract-concept learning (Fig. 3). Pigeons (and Clark’s nutcrackers and black-billed magpies described later) did not receive transfer tests at 16 items to be consistent with the procedures used for capuchin and rhesus monkeys.

Fig. 3
figure 3

Mean percent correct performance of four pigeons on training (baseline) trials (broken green line, filled symbols) and novel (transfer) trials (unbroken green line, unfilled symbols) with successively expanded training sets from eight to 1,024 picture stimuli in the same/different task. The dotted line at 50% correct is chance performance. Error bars are + one standard error of the mean

Pigeons learned the initial eight-item task in roughly 30 sessions, similar to the rhesus (40 sessions) and the capuchin (35 sessions) monkeys. Initially, the pigeons’ novel-stimulus transfer was similar (51.3%) to that of the monkeys, and also did not differ from chance (50%) performance. The pigeons’ transfer performance increased as the training set size increased and became statistically equivalent to the training baseline levels for set sizes 256, 512, and 1,024, resulting in compelling evidence that pigeons can learn the same/different abstract concept. Other pigeons in a control group (not shown) were trained without expansion of the eight-item training set-size, but were yoked to individual experimental pigeons in terms of the number of training and transfer sessions. There was no increase in transfer performance for that pigeon control group, demonstrating that training with set-size expansion was the critical manipulation producing high accuracies and same/different concept learning, similar in many respects to the primates studied.

Although the pigeons’ same/different learning was quite remarkable in itself, the set size required by pigeons (256-item set) was twice the size of that required by monkeys (128-item set). The monkeys’ set-size transfer functions were nearly entirely above the pigeons. The steeper set functions for monkeys show that the abstract-concept learning occurred more rapidly and required fewer exemplars of the same/different rule than pigeons. Nevertheless, pigeons like monkeys can and did attain the same/different abstract concept, which supports the necessity for using large training sets and systematically increasing the training set. By systematically increasing the training set size, we were able to directly compare how many exemplars were necessary for the species to show concept learning. This highlights the importance of comparing across set-size functions where transfer performance ranges from chance to full concept learning. The functions reveal how the functional relationship between set size and transfer performance can vary across species and is particularly important for cognitive functions like same/different abstract-concept learning that can be compared across different species (Wright, 2013). In summary, these data clearly show a quantitative difference in abstract concept formation with a qualitative similarity in that monkeys and pigeons fully learn the abstract same/different concept.

Same/different abstract-concept learning by Clark’s nutcrackers and black-billed magpies

Pigeons required a substantially larger set-size compared to capuchin and rhesus monkeys. Although this finding may highlight differences among primates and birds, drawing such a conclusion would be premature given the limited sampling of avian species. Although pigeons are traditionally used for studies of avian visual cognition, examining other species provide opportunities to better understand the evolution of abstract-concept learning. Many corvids (birds from the family Corvidae, which include ravens, crows, jays, and nutcrackers) are known for their impressive cognitive abilities (see review by Emery & Clayton, 2004). For instance, the food-storing Clark’s nutcracker has been shown to use human gestures (Clary & Kelly, 2013; Tornick, Gibson, Kispert, & Wilkinson, 2011), mirror self-recognition (Clary & Kelly, 2016), and engage in behaviors to protect caches from thieving conspecifics (Clary & Kelly, 2011). Clark’s nutcrackers also have highly developed spatial abilities, argued to be adaptive specializations for its strong reliance on food-stores. An individual nutcracker will place 20,000–30,000 pine nuts in thousands of hidden caches in autumn, and return to these caches several months later when food is scarce (Tomback, 1998; Vander Wall, 1982; Vander Wall & Balda, 1977). To find the cache, nutcrackers need to rely on their spatial memory for the location as the cache is typically obscured by a thick layer of snow. Such highly developed spatial memory depends upon precise relational memory for encoding the location of hundreds of cache sites relative to local landmarks (see Gould, Kelly, & Kamil, 2010). Black-billed magpies, like nutcrackers, are also members of the corvid family and belong to the same clade as nutcrackers. They have also been shown to have remarkable cognitive abilities, such as mirror self-recognition (Prior, Schwarz, & Güntürkün, 2008). Unlike Clark’s nutcrackers, magpies inhabit more temperate altitudes and are more omnivorous, thus they only cache small amounts of food for short periods of time (Trost, 1999). Thus, magpies are intermediate between nutcrackers and the non-storing pigeon (although also not a member of the corvid family) in terms of caching and retrieval of food. Examining these three species may help to resolve whether the mechanisms supporting same/different abstract-concept learning are similar across bird species, or are related to the relative reliance on food stores.

Nine wild-caught Clark’s nutcrackers (Nucifraga columbiana) and ten wild-caught black-billed magpies (Pica hudsonia) were tested for their same/different abstract-concept learning with procedures very similar to those used to test pigeons including the same: stimuli, stimulus pairs, sequences of stimulus pairs used in training and transfer testing, display size, 15-s inter-trial intervals, and required 20 pecks to the sample stimuli (Magnotti, Katz, Wright, & Kelly, 2015; Magnotti, Wright, Leonard, Katz, & Kelly, 2017; Wright, Magnotti, Katz, Leonard, & Kelly, 2016; Wright, Magnotti, Katz, Leonard, Vernouillet, & Kelly, 2017). Nutcrackers and magpies made their pecks from a perch in front of the stimulus display. Similar to pigeons and monkeys, a response to the comparison picture was correct when it matched the sample picture; a response to the white rectangle to the right of the comparison picture was correct when the comparison picture did not match the sample picture. Correct choice responses were reinforced with mealworms delivered below the monitor via a rotating dispensing wheel. Like pigeons and monkeys, these birds were trained on 100-trial sessions (50 same, 50 different trials). Following acquisition (85% correct or better), abstract-concept learning was assessed in six consecutive transfer sessions, where each session contained 90 baseline (training) trials and ten novel (five same, five different) transfer trials. Correct responses on transfer trials were reinforced identically to baseline trials. The cycle of set-size expansion, training for a minimum of three sessions, obtaining 85% correct or better, and novel-stimulus transfer testing was (as for pigeons) repeated six times for training sets of 32, 64, 128, 256, 512, and 1,024 picture items (see Wright & Katz, 2006, for training and testing stimuli.)

Abstract-concept learning, as measured by transfer to novel picture pairs was 65% and 67% correct (chance 50% correct) following learning with the initial eight-item training set for nutcrackers and magpies, respectively (Fig. 4). Transfer by both corvid species increased regularly and monotonically with the training set-size expansions until transfer performance was indistinguishable from baseline. Magpies and nutcrackers were statistically equivalent in their substantial transfer following training with the eight-item set. Both of these corvid species clearly outperformed pigeons across the rising portion of the set-size functions. Nutcrackers and magpies also outperformed the monkeys in their initial transfer (partial concept learning) following training with the eight-item set. Similar to monkeys, the nutcrackers and magpies attained full same/different abstract-concept learning with the same 128-item training set size.

Fig. 4
figure 4

Mean percent correct performance of nine Clark’s nutcrackers (blue) and ten black-billed magpies (purple) on training (baseline) trials (broken lines and filled symbols) and novel (transfer) trials (solid lines, unfilled symbols) with successively expanded training sets from eight to 1,024 picture stimuli in the same/different task. The dotted line at 50% correct is chance performance. Error bars are + one standard error of the mean

Comparing learning rates on the initial acquisition, nutcrackers (mean = 3,000 trials) and magpies (mean = 3,540 trials) were very similar to the other species (rhesus 3,766, capuchins 3,533, and pigeons 2,875 trials). Table 1 presents the mean trials to criterion for each species at each of the trained set sizes. Learning rates generally declined as the training set size was expanded for every species, demonstrating the benefit of progressively better transfer and partial concept learning (see also Wright & Katz, 2007).

Table 1 Mean number of trials to criterion at each of the training set sizes for five species

Conclusions from same/different concept learning

Training and testing these different species with the same stimuli and procedures facilitates direct species comparisons of how novel-stimulus transfer develops, and systematically changes, with training set-size including: initial transfer with a small eight-item training set, the training set size where transfer is equivalent to baseline training performance, and the overall differences and similarities among the set-size transfer functions for these different species.

The nutcrackers’ and magpies’ set-size functions for same/different abstract-concept learning were virtually equivalent including initial transfer (partial concept learning) and more complete concept learning, and therefore do not point to the caching and recovery skills of nutcrackers being an advantage in same/different abstract-concept learning. Similarly, sociality is unable to explain the similar set-size functions (e.g., social brain hypothesis; Dunbar, 1998), as although black-billed magpies are relatively social, Clark’s nutcrackers are not as social as magpies. It appears that same/different abstract-concept learning is facilitated by evolved neural architecture of nutcrackers and magpies. Moreover, the results from these two corvid species point to the possibility that corvids generally (not just Clark’s nutcrackers and black-billed magpies) might be able to fully learn a higher-order abstract concept following exposure to a similar number of concept exemplars (128-picture set training) in comparison to old-world (rhesus monkeys) and new-world (capuchin monkeys) nonhuman primates.

Proactive interference in same/different tasks

In the previous same/different abstract concept-learning studies discussed above with small training sets there were frequent stimulus repetitions from previous trials. Some of these stimulus repetitions likely produced proactive interference, even with simultaneous presentations of the pictures shown in Fig. 1, that would have diminished accuracy, hindered learning, and prevented better same/different concept learning. Nevertheless, proactive interference is much more prevalent with delayed presentations (in a sequential same/different task) when a previously seen sample is re-presented as a test picture on a later different trial, as shown in Fig. 5. Having seen that test picture before (e.g., the sample in the immediately previous trial) creates confusion and increases the likelihood that a same response (an incorrect response in this case) will be made.

Fig. 5
figure 5

An example of trials used to test for proactive interference in the sequential same/different task. On interference test trials (Trial n, in the example) the test stimulus does not match the sample stimulus on that trial, but it does match the sample stimulus on a previous trial (e.g., Trial n-1). This match to a sample stimulus from a previous trial tends to create confusion, and can result in making an incorrect same response to the test picture

Tests of proactive interference are best made by inserting occasional proactive-interference test trials within a much larger session of non-interference trials (i.e., trial unique stimuli). By using only occasional proactive interference test trials, overall performance on non-interference trials will be maintained at the “baseline” level for non-interference trials. Hence, by using trial unique sessions, except for the repeated stimuli at the time of test, in the sequential same/different task allowed us to precisely test the locus of proactive interference. During these tests of proactive interference, “interfering” stimuli were presented one to 16 trials prior to the test trial, resulting in a function of proactive interference due to the proximity of the interference trial to the test trial.

Pigeons: Time-based proactive interference

Such tests of proactive interference were conducted with four pigeons (Wright et al., 2012). A large stimulus set of 1,024 images (the same 1,024 picture set used to train same/different abstract-concept learning) was used so that consecutive sessions could be constructed to be trial unique. Proactive interference was tested by placing potentially interfering stimuli one, two, four, eight, or 16 trials prior to test trials. Because stimuli for baseline trials were trial-unique and not repeated for more than 2 weeks of testing, we were able to determine the locus of interference across many trials. Pigeons pecked the sample stimulus 20 times, followed by a delay (1 s or 10 s, in a block design), then a test stimulus and white rectangle was presented. The pigeons then made a choice response, followed by a 15-s inter-trial interval. Each daily session consisted of 64 trials with five interference tests (one test each at the one-, two-, four-, eight-, and 16-trial separations). There were 32 same and 32 different trials per session, 24 sessions per block, and four blocks alternating between 1-s and 10-s delays.

The shorter 1-s delay (red function), produced a moderate amount of proactive interference that was gradually alleviated as trial separation approached the No-PI condition shown in Fig. 6. With the longer 10-s delay (blue function), there was a considerably larger 47% proactive-interference effect when the interfering stimulus was presented in the immediately preceding trial (n-1). This greater proactive interference also dissipated with increasing trial separation, but there was still a residual interference effect of 11% for interfering stimuli presented 16 trials prior (i.e., more than 6.5-min prior), and greater interference throughout the function compared to the 1-s delay condition with interfering stimuli presented 16 trials prior.

Fig. 6
figure 6

Percent correct performance by pigeons for 1- and 10-s delays, with model fits (1 sigma bands) based on a signal-detection-theory model of the ratio of log time to the test-trial sample divided by the log time to the interfering sample (see Wright et al., 2012)

Greater interference at the longer 10-s delay relative to the shorter 1-s delay is somewhat counterintuitive. For example, if a stimulus occurs say 15 s back in time versus 2 s back in time from the moment of a test, then the stimulus closer in time (the 2-s stimulus) should produce more interference. Hence, counterintuitively, as shown in Fig. 6 the opposite happens with the 10-s delay condition producing greater proactive interference than the 1-s delay condition. To elaborate, interfering stimuli encountered more distantly in the past should, according to models of decay or limited capacity, translate to more forgetting and therefore less interference. But just the opposite occurred. This counterintuitive finding was shown to obey a Weber-Fechner time ratio of time discriminations from the test to the current-trial sample stimulus divided by time to the interfering stimulus (see Wright et al., 2012). The model was fit simultaneously to both PI functions (colored shaded bands) using the same parameters (bias and maximum accuracy) and accounted for 95% of the variance, including the no-PI condition.

Monkeys: Event-based proactive interference

Three rhesus monkeys were tested in the sequential same/different task for PI from prior trials, using similar procedures to those for testing pigeons (Devkar & Wright, 2016). Most of the conditions were the same for monkeys as they were for pigeons, including: trial-unique pictures selected from the same 1,024 picture set without replacement on baseline trials, interference tests separated by one, two, four, eight, or 16 trials prior, 15-s inter-trial intervals, numbers of trials per sessions, sessions per block, and repetitions of blocks (also described above). Monkeys were tested with three delays (1 s, 10 s, and 20 s; Fig. 7). The longer 20-s delay was used to evaluate whether there was any time-based interference by monkeys, as was shown by pigeons.

Fig. 7
figure 7

The three monkeys’ mean percent-correct performance on trial blocks with 1-, 10-, or 20-s delays between the off-set of sample presentation and the onset of the test presentation (n-1 was the immediately preceding trial) compared with baseline performance (No-PI trials). Chance performance was 50% correct. Error bars are + one standard error of the mean (Devkar & Wright, 2016)

There were, however, no statistically significant differences or interactions among the three proactive-interference functions, and thus no evidence for time-based proactive interference for rhesus monkeys, like that found for pigeons. If PI was time based like in pigeons then there would have been similar decreases in relation to the delay at n-2, 4, 8, and 16 to that seen for n-1. Indeed, even a different time manipulation (short 5-s inter-trial interval or a longer 15-s inter-trial interval) had no effect on these monkeys’ proactive-interference functions.

The monkeys’ functions revealed considerable proactive interference at n-1 with little or no substantial effect of time over the five n-back trials tested, as was evident with pigeons. This insensitivity to time points to the primates’ proactive interference being ‘event’ based, not ‘time’ based. Good event memory is the hallmark of explicit memory, as opposed to time memory (by pigeons), which is based on familiarity (implicit) memory (Devkar & Wright, 2016). The findings indicate both a qualitative and quantitative difference between the rhesus monkeys and pigeons in how they combat proactive interference in the sequential same/different task.


Pigeons tended to rely on familiarity (i.e., implicit memory) for their same/different judgments, but they do learn the concept of same/different. Monkeys learned the same/different concept more readily than pigeons, perhaps due to their reliance on memory of events (i.e., explicit memory). Corvid species (Clark’s nutcrackers and black-billed magpies) also learned the same/different concept more readily than the pigeons. Indeed, the corvids initially even out-performed the monkeys at the beginning of testing as the set-sizes were being expanded. The corvids may process the items more similarly to the primates (possibly using event-based memory) than to the pigeons (using time-based memory), which would be an interesting avenue for future research. Likewise, whether some bird families, such as corvids or psittacines, which have been shown capable of remarkable cognitive feats (such as same/different discrimination, Pepperberg, 1987) are also better able to learn abstract concepts when provided with fewer exemplars, warrants further investigation.

In conclusion, we have revealed two types of functional relationships that allow scientists to better understand same/different concept learning. First, the functional relationship that underlies abstract-concept learning in relation to training set size clearly shows how set size is a variable that if used properly by researchers can allow animals to fully express the same/different abstract concept. Second, the functional relationship that underlies how PI may be combated to solve the same/different task revealed different types of memory that can be used to solve the same/different task. Together, these functional relationships show that the species tested share the qualitatively ability to learn the abstract same/different concept. However, the difference in how rhesus monkey and pigeons combat PI shows how qualitatively different memory process can produce the same ability. A pluralistic view is needed to understand how these similarities and differences in processes evolved (Krakauer, Ghazanfar, Gomez-Marin, MacIver, & Poeppel, 2017). Revealing functional relationships within and across species will be at the heart of such scientific understanding.


  • Clary, D., & Kelly, D. M. (2011). Cache protection strategies of a non-social food-caching corvid, Clark's nutcracker (Nucifraga columbiana). Animal Cognition, 14, 735–744.

    Article  Google Scholar 

  • Clary & Kelly, D. M. (2013). Are Clark’s nutcrackers (Nucifraga columbiana) able to discriminate knowledge states of human experimenters during an object-choice task? Evolutionary Psychology, 11, 628–646.

    Article  Google Scholar 

  • Clary, D., & Kelly, D. M. (2016). Graded mirror self-recognition by Clark’s nutcrackers. Scientific Reports, 6,, 2013.

  • Devkar, D. T., & Wright, A. A. (2016). Event based proactive interference by rhesus monkeys. Psychonomic Bulletin and Review, 23, 1474–1482.

  • Dunbar, R. I. (1998). The social brain hypothesis. Evolutionary Anthropology, 6, 178–190.;2-8.

  • Emery, N. J., & Clayton, N. S. (2004). The mentality of crows: convergent evolution of intelligence in corvids and apes. Science, 306, 1903–1907.

    Article  Google Scholar 

  • Gould, K. L., Kelly, D. M., & Kamil, A. C. (2010). What scatter-hoarding animals have taught us about small-scale navigation. Philosophical Transactions of the Royal Society: Biological Sciences, 365, 901–914.

    Article  Google Scholar 

  • James, W. (1890/1950). The principles of psychology, p. 459. Dover Publications, New York (Original work published in 1890).

  • Katz, J. S., & Wright, A. A. (2021). Issues in the comparative cognition of same/different abstract-concept learning. Current Opinion in Behavioral Sciences, 37, 29–34.

    Article  Google Scholar 

  • Katz, J. S., & Wright, A. A. (2006). Same/different concept learning by pigeons. Journal of Experimental Psychology: Animal Behavior Processes, 32, 80–86.

    PubMed  Google Scholar 

  • Katz, J. S., Wright, A. A., & Bachevalier, J. (2002). Mechanisms of same/different abstract-concept learning by rhesus monkeys (Macaca mulatta). Journal of Experimental Psychology: Animal Behavior Processes, 28, 358–368.

    PubMed  Google Scholar 

  • Katz, J. S., Wright, A. A., & Bodily, K. D. (2007). Issues in the comparative cognition of abstract-concept learning. Comparative Cognition and Behavior Reviews, 2, 79–92

  • Keppel, G., & Underwood, B. J. (1962). Proactive inhibition in short-term retention of single items. Journal of Verbal Learning & Verbal Behavior, 1, 153–161.

    Article  Google Scholar 

  • Krakauer, J. W., Ghazanfar, A.A., Gomez-Marin, A., MacIver, M. A., Poeppel, D. (2017). Neuroscience Needs Behavior: Correcting a Reductionist Bias. Neuron, 93, 480–493.

    Article  Google Scholar 

  • Magnotti, J. F., Katz, J. S., Wright, A. A., & Kelly, D. M. (2015). Superior abstract-concept learning by Clark's nutcrackers (Nucifraga columbiana). Biology Letters, 11, 20150148.

    Article  PubMed  PubMed Central  Google Scholar 

  • Magnotti, J. F., Wright, A. A., Leonard, K., Katz, J. S., & Kelly, D. M. (2017). Abstract-concept learning in Black-billed magpies (Pica hudsonia). Psychonomic Bulletin and Review, 24, 431–435.

    Article  Google Scholar 

  • Nakamura, T., Wright, A. A., Katz, J. S., Bodily, K. D., & Sturz, B. R. (2009). Abstract-concept learning carryover effects from the initial training set in pigeons (Columba livia). Journal of Comparative Psychology, 123, 79–89

  • Pepperberg, I. M. (1987). Acquisition of the same/different concept by an African Grey parrot (Psittacus erithacus): Learning with respect to categories of color, shape, and material. Animal Learning & Behavior, 15, 423–432.

    Article  Google Scholar 

  • Premack, D. (1978). On the abstractness of human concepts: Why it would be difficult to talk to a pigeon. In S. H. Hulse, H. Fowler, & W. K. Honig (Eds.), Cognitive processes in animal behavior (pp. 423–451). Hillsdale, NJ: Erlbaum.

    Google Scholar 

  • Premack, D. (1983). Animal cognition. Annual Review of Psychology, 34, 351–362.

    Article  Google Scholar 

  • Premack, D. & Premack, A. J. (1983). The mind of an ape. New York: W.W. Norton & Co.

    Google Scholar 

  • Prior, H., Schwarz, A., & Güntürkün, O. (2008) Mirror-Induced Behavior in the Magpie (Pica pica): Evidence of self-recognition. PLoS Biol, 6(8):e202.

  • Tomback, D. F. (1998) Clark’s nutcracker (Nucifraga columbiana), The Birds of North America Online. Ithaca: Corbell Lab of Ornithology.

  • Tornick, J. K., Gibson, B. M., Kispert, D., & Wilkinson, M. (2011). Clark’s nutcrackers (Nucifraga columbiana) use gestures to identify the location of hidden food. Animal Cognition, 14, 117–125.

    Article  Google Scholar 

  • Trost, C. H. (1999). In A. Poole (Ed.), Black-billed Magpie (Pica hudsonia), The Birds of North America Online. Ithaca: Cornell Lab of Ornithology.

  • Vander Wall, S. B. (1982). An experimental analysis of cache recovery in Clark's nutcracker. Animal Behaviour, 30, 84–94.

    Article  Google Scholar 

  • Vander Wall, S. B., & Balda, R. P. (1977). Coadaptations of the Clark’s nutcracker and the pinyon pine for efficient seed harvest and dispersal. Ecological Monographs, 47, 89–111.

    Article  Google Scholar 

  • Wright, A. A. (2013). Functional relationships for investigating cognitive processes. Behavioural Processes, 93, 4–24.

    Article  Google Scholar 

  • Wright, A. A., Cook, R. G., Rivera, J. J., Sands, S. F., & Delius, J. D. (1988). Concept learning by pigeons: Matching to sample with trial-unique video picture stimuli. Animal Learning & Behavior, 16, 436–444.

  • Wright, A. A., & Katz, J. S. (2006). Mechanisms of same/different concept learning in primates and avians. Behavioural Processes, 72, 234–254.

    Article  Google Scholar 

  • Wright, A. A., & Katz, J. S. (2007). The generalization hypothesis of abstract-concept learning: Learning strategies and related issues in Macaca mulatta, Cebus apella, and Columba livia. Journal of Comparative Psychology, 121, 387–397.

    Article  Google Scholar 

  • Wright, A. A., Katz, J. S., & Ma, W. (2012). How to be proactive about interference: Lessons from animal memory. Psychological Science, 23, 453–458.

    Article  PubMed  PubMed Central  Google Scholar 

  • Wright, A.A., Kelly, D.M., & Katz, J. S. (2018). Comparing cognition by integrating concept learning, proactive interference, and list memory. Learning & Behavior, 46, 107–123.

  • Wright, A. A., Magnotti, J. F., Katz, J. S., Leonard, K., & Kelly, D. M. (2016). Concept learning set-size functions for Clark's nutcrackers. Journal of the Experimental Analysis of Behavior, 105, 76–84.

    Article  PubMed  Google Scholar 

  • Wright, A. A., Magnotti, J. F., Katz, J. S., Leonard, K., Vernouillet, A., & Kelly, D. M. (2017). Corvids outperform pigeons and primates in learning a basic concept. Psychological Science, 28, 437–444.

    Article  Google Scholar 

  • Wright, A. A., Rivera, J. J., Katz, J. S., & Bachevalier, J. (2003). Abstract-concept learning and list-memory processing by capuchin and rhesus monkeys. Journal of Experimental Psychology: Animal Behavior Processes, 29, 184–198.

    PubMed  Google Scholar 

Download references

Author information

Authors and Affiliations


Corresponding author

Correspondence to Debbie M. Kelly.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Wright, A.A., Kelly, D.M. & Katz, J.S. Same/different concept learning by primates and birds. Learn Behav 49, 76–84 (2021).

Download citation

  • Accepted:

  • Published:

  • Issue Date:

  • DOI:


  • Concept learning
  • Interference
  • Working memory
  • Black-billed magpies
  • Clark’s nutcrackers
  • Pigeons
  • Primates