Serial learning by rhesus monkeys: II. Learning four-item lists by trial and error.

Three rhesus macaque monkeys were trained to produce novel 4-item lists (A-->B-->C-->D) on which all items were displayed from the start of training. Subjects were previously trained to produce 4-item lists by adding one item at a time (A, A-->B, A-->B-->C, and A-->B-->C-->D; lists K. B. Swartz et al., 1991). Those lists could be mastered by responding to each new item last. To learn lists on which all items were displayed from the start of training, subjects had to recall the consequences of errors and correct responses to each item. Errors ended the trial; correct responses to A, B, or C allowed the trial to continue. A correct response to D produced food reward. Although the probability of executing a 4-item list correctly by chance was .04, each subject mastered 4 novel 4-item lists by trial and error. The ability of monkeys to use a trial-and-error strategy to learn novel lists provides a basis for studying the development of serial expertise in animals.

Representation of the numerosities 1-9 by rhesus macaques (Macaca mulatta).

Three rhesus monkeys (Macaca mulatta) were trained to respond to exemplars of 1, 2, 3, and 4 in an ascending, descending, or a nonmonotonic numerical order (1-->2-->3-->4, 4-->3-->2--1, 3-->1-->4-->2). The monkeys were then tested on their ability to order pairs of the novel numerosities 5-9. In Experiment 1, all 3 monkeys ordered novel exemplars of the numerosities 1-4 in ascending or descending order. The attempt to train a nonmonotonic order (3-->1-->4-->2) failed. In Experiment 2A, the 2 monkeys who learned the ascending numerical rule ordered pairs of the novel numerosities 5-9 on unreinforced trials. The monkey who learned the descending numerical rule failed to extrapolate the descending rule to new numerosities. In Experiment 2B all 3 monkeys ordered novel exemplars of pairs of the numerosities 5-9. Accuracy and latency of responding revealed distance and magnitude effects analogous to previous findings with human participants (R. S. Moyer & T. K. Landaeur, 1967). Collectively these studies show that monkeys represent the numerosities 1-9 on at least an ordinal scale.

Ordering of the numerosities 1 to 9 by monkeys.

A fundamental question in cognitive science is whether animals can represent numerosity (a property of a stimulus that is defined by the number of discriminable elements it contains) and use numerical representations computationally. Here, it was shown that rhesus monkeys represent the numerosity of visual stimuli and detect their ordinal disparity. Two monkeys were first trained to respond to exemplars of the numerosities 1 to 4 in an ascending numerical order (1 --> 2 --> 3 --> 4). As a control for non-numerical cues, exemplars were varied with respect to size, shape, and color. The monkeys were later tested, without reward, on their ability to order stimulus pairs composed of the novel numerosities 5 to 9. Both monkeys responded in an ascending order to the novel numerosities. These results show that rhesus monkeys represent the numerosities 1 to 9 on an ordinal scale.

Serial learning with a wild card by pigeons (Columba livia): effect of list length.

Pigeons (Columba livia) learned 3-, 4- or 5-item lists prior to subset and wild card tests. On the latter, a novel item replaced 1 of the list items. Pigeons who learned 3-item lists responded accurately on all subset pairs (AB, BC, and AC) and on all types of 3-item wild card trials (WBC, AWC, & ABW). Pigeons who learned 4- and 5-item lists responded at chance levels of accuracy on all subsets that did not contain a start or an end item (BC, BD, & CD, respectively, on 4- and 5-item subset tests). On wild card trials, they exceeded chance levels of performance only when the wild card replaced the last item (ABCW & ABCDW trials). Monkeys (Cebus apella) trained to produce a 5-item list perform accurately on all subsets and wild cards. (M. R. D'Amato & M. Colombo, 1988, 1989). These differences provide strong evidence that pigeons and monkeys form qualitatively different representations of lists containing four or more items.

Serial learning by rhesus monkeys: I. Acquisition and retention of multiple four-item lists.

Two rhesus monkeys were trained to learn eight 4-item lists, each composed of 4 different photographs. Lists were trained in successive phases: A, A----B, A----B----C, and A----B----C----D. After List 4, retention, as measured by the method of savings, was, on average, 66% (range: 44-84%). Indeed, all 4 lists could be recalled reliably during a single session with neither a decrement in accuracy nor an increase in the latency of responding to each item. Response latencies on a subset test employing all possible 2- and 3-item subsets of each 4-item list support the hypothesis that monkeys form linear representations of a list. Latencies to Item 1 of a subset varied directly with the position of that item in the original list. On List 1, latencies to Item 2 varied directly with the number of intervening items between Item 1 and Item 2 in the original list. During the acquisition of Lists 5-8, both Ss mastered the A----B and A----B----C phases of training in the minimum number of trials possible.

Chunking during serial learning by a pigeon: III. What are the necessary conditions for establishing a chunk?

What are the minimal conditions for the formation of chunks by a pigeon learning an arbitrary list? Experiment 1 compared the acquisition of two types of chunkable list (each composed of colors and achromatic geometric forms): A----B----C----D'----E' (or A'----B'----C'----D----E) and A----B----C'----D'----E' (or A'----B'----C----D----E). The first type of list was acquired more rapidly than the second. On both lists, however, evidence of chunking did not emerge until the four-item phase of training (e.g., pauses at the end of one category of list item). In Experiment 2, chunking was shown to occur on four-item lists in which colors and forms were segregated (A----B----C'----D' and A'----B'----C----D), but not on lists in which the two types of items were interspersed (A----B'----C'----D and A'----B----C----D'). As in Experiment 1, evidence of chunking (pauses at chunk boundaries) did not appear until the fourth item was added.

Chunking during serial learning by a pigeon: I. Basic evidence.

Chunking by pigeons was demonstrated by comparing performance on different types of lists. Experiment 1 showed that Groups II and IV (who learned lists in which colors and achromatic geometric forms were segregated: A----B----C----D'----E' and A----B----C----D----E', respectively) executed lists more rapidly than did Group I (who learned a homogeneous list of colors: A----B----C----D----E) or Groups II and III (who learned lists consisting of unsegregated colors and forms: A----B'----C----D'----E and A----B----C'----D----E, respectively). Experiment 2 showed that Groups II and IV tolerated interruptions of the list better than did Groups I, III, and V. The accuracy of responding of Groups I, III, and V decreased as a function of the duration of the interruption and the point in the sequence at which it occurred. The performance of Group II was unaffected by interruptions; Group IV was minimally affected. These results indicate that Groups II and IV organized their lists as ordered chunks.

Chunking during serial learning by a pigeon: II. Integrity of a chunk on a new list.

Can a group of items that a pigeon chunks on one list function as such on a second list? In Experiment 1, the ordinal position of the chunk was held constant across both lists. Following training on a list of colors and achromatic geometric forms (A----B----C----D'----E'), the integrity of the color chunk [A----B----C] was maintained on a list of five colors (A----B----C----F----G) even though the basis for establishing that chunk was eliminated. The integrity of the chunk [A----B----C] was also maintained on a new list of colors and forms (A----B----C----D*----E*). In Experiment 2, the ordinal position of a chunk established on list1 (A----B----C'----D') was changed on list 2. As shown by positive transfer between lists 1 and 2, the integrity of the chunks [A----B] and [C'----D'] was maintained on lists X'----A----B----Y' and X'----C'----D'----Y', respectively. Conversely, the heterogeneous list X'----B----C'----Y' took longer to learn than the original list.

Chunking by a pigeon in a serial learning task.

A basic principle of human memory is that lists that can be organized into memorable 'chunks' are easier to remember. Memory span is limited to a roughly constant number of chunks and is to a large extent independent of the amount of informaton contained in each chunk. Depending on the ingenuity of the code used to integrate discrete items into chunks, one can substantially increase the number of items that can be recalled correctly. Newly developed paradigms for studying memory in non-verbal organisms allow comparison of the abilities of human and non-human subjects to memorize lists. Here I present two types of evidence that pigeons 'chunk' 5-element lists whose components (colours and achromatic geometric forms) are clustered into distinct groups. Those lists were learned twice as rapidly as a homogeneous list of colours or heterogeneous lists in which the elements are not clustered. The pigeons were also tested for knowledge of the order of two elements drawn from the 5-element lists. They responded in the correct order only to those subsets that contained a chunk boundary. Thus chunking can be studied profitably in animal subjects; the cognitive processes that allow an organism to form chunks do no presuppose linguistic competence.

On the nature of animal thinking.

Recent attempts to teach apes rudimentary grammatical skills have produced negative results. The basic obstacle appears to be at the level of the individual symbol which, for apes, functions only as a demand. Little evidence exists that apes can use symbols as names, that is, as a means of simply transmitting information. Even though non-human animals lack linguistic competence, much evidence has accumulated recently that a variety of animals can form representations of particular features of their environment. What then is the non-verbal nature of animal representations? This question will be discussed with reference to recent investigations of animal memory--in particular, studies concerned with a pigeon's ability to learn "lists" of arbitrary elements. While learning to produce a particular sequence of four elements (colors), pigeons also acquire knowledge about the relation between non-adjacent elements and about the ordinal position of a particular element. Learning to produce a particular sequence also facilitates the discrimination of that sequence from other sequences.

Physical restraint produces rapid acquisition of the pigeon's key peck.

The acquisition and maintenance of autoshaped key pecking in pigeons was studied as a function of intertrial interval. At each of six intervals, which ranged from 12 seconds to 384 seconds, four pigeons were physically restrained during training while four other pigeons were not restrained. Restrained subjects acquired key pecking faster and with less intragroup variability at each interval. The effects of restraint were specific to acquisition and were not evident in maintained responding after five postacquisition sessions.

Can an ape create a sentence?

More than 19,000 multisign utterances of an infant chimpanzee (Nim) were analyzed for syntactic and semantic regularities. Lexical regularities were observed in the case of two-sign combinations: particular signs (for example, more) tended to occur in a particular position. These regularities could not be attributed to memorization or to position habits, suggesting that they were structurally constrained. That conclusion, however, was invalidated by videotape analyses, which showed that most of Nim's utterances were prompted by his teacher's prior utterance, and that Nim interrupted his teachers to a much larger extent than a child interrupts an adult's speech. Signed utterances of other apes (as shown on films) revealed similar non-human patterns of discourse.

Serial learning in the pigeon.

Three pigeons learned to peck four colors in a particular sequence, regardless of how these colors were positioned on four response keys and without feedback following each response. This demonstrates that serial learning is possible in subprimate animals.

Autoshaping, random control, and omission training in the rat.

The role of the stimulus-reinforcer contingency in the development and maintenance of lever contact responding was studied in hooded rats. In Experiment I, three groups of experimentally naive rats were trained either on autoshaping, omission training, or a random-control procedure. Subjects trained by the autoshaping procedure responded more consistently than did either random-control or omission-trained subjects. The probability of at least one lever contact per trial was slightly higher in subjects trained by the omission procedure than by the random-control procedure. However, these differences were not maintained during extended training, nor were they evident in total lever-contact frequencies. When omission and random-control subjects were switched to the autoshaping condition, lever contacts increased in all animals, but a pronounced retardation was observed in omission subjects relative to the random-control subjects. In addition, subjects originally exposed to the random-control procedure, and later switched to autoshaping, acquired more rapidly than naive subjects that were exposed only on the autoshaping procedure. In Experiment II, subjects originally trained by an autoshaping procedure were exposed either to an omission, a random-control, or an extinction procedure. No differences were observed among the groups either in the rate at which lever contacts decreased or in the frequency of lever contacts at the end of training. These data implicate prior experience in the interpretation of omission-training effects and suggest limitations in the influence of stimulus-reinforcer relations in autoshaping.

Evidence for the innate basis of the hue dimension in the duckling.

Different groups of ducklings reared under sodium monochromatic light (589 nanometers) and under white light were trained to discriminate between the stimulus correlated with reinforcement (589 nanometers), and the stimulus correlated with extinction, whole value was either 570 or 610 nanometers. The peaks of subsequently obtained gradients of wavelength generalization of both groups were displaced away from the stimulus correlated with extinction. The peaks of the groups trained not to respond to 570 nanometers were located at 600 nanometers. The peaks of the groups trained not to respond to 610 nanometers were located at 580 nanometers. These results (in agreement with earlier data of Rudolph and Honig, 1972) suggest that ducklings have an innate basis for ordering stimuli of different wavelengths along the hue dimension.

On the minimal conditions for the development of a peak-shift and inhibitory stimulus control.

Failures to obtain a peak-shift and inhibitory stimulus control following massed extinction in the presence of intra- and interdimensional stimuli confirm earlier results reported by Honig, Thomas, and Guttman (1959) and by Weisman and Palmer (1969). Peak-shifts and inhibitory stimulus control were observed when any of the following procedures intervened for 3 min between massed extinction and generalization testing: (1) S+ presented, responding reinforced; (2) S+ presented, responding not reinforced; or, (3) noncontingent food presented in the presence of a dark key. Behavioral contrast was shown not to be a necessary antecedent for the occurrence of peak-shift or inhibitory stimulus control.

On the nature of non-responding in discrimination learning with and without errors.

In human subjects, discrimination learning with errors results in active responding incompatible with the reinforced response. The direction of such incompatible behavior is opposite to that of the reinforced response. Responding occurs only during the stimulus correlated with extinction. The frequency of active non-responding is maximal shortly after the start of discrimination training (the time at which the frequency of errors decreases most rapidly) and approaches zero as discrimination training continues. The magnitude of behavioral contrast is not related systematically to the number of errors. Instead it is related directly to the frequency of active non-responding. Active non-responding appears to be motivated by the aversiveness of self-produced frustration, in the sense that active non-responding allows the subject to avoid the aversiveness of non-reinforced responding.

Extinction of a discriminative operant following discrimination learning with and without errors.

Different groups of pigeons were trained to respond to red and not to green with and without errors (responses to green) under a free operant procedure, in which responding to red was intermittently reinforced, and under a trial procedure in which all responses to red were reinforced. The response to red was then extinguished under a procedure in which the discriminative stimuli were successively alternated as during discrimination training. The performances of those birds that learned the discrimination without errors under the trial procedure were seriously disrupted during extinction; the birds persistently responded to green for the first time. The performances of those subjects that learned the discrimination without errors under the free operant procedure were not disrupted during extinction. In a second experiment, the same discrimination was trained without errors under a trial procedure in which the response to red was intermittently reinforced. Extinction did not disrupt discrimination performance. Thus, errorless discrimination performance was shown to remain intact during extinction so long as the response to red was intermittently reinforced during discrimination training.