Verbal feedback and the feature-positive effect in children

A pair of displays having common elements may be differentiated by the presence of a distinctive feature in one of the displays. When required to discriminate between such displays presented simultaneously, young children more readily learn to confine their responses to the display containing the distinctive feature (feature-positive condition) than to the display which does not contain the distinctive feature (feature-negative condition). The effect of explicit verbal feedback for incorrect choices on the learning of discriminations of this type was examined in 3- to 5-year-old children. In the feature-positive case, explicit feedback for incorrect responses increased the tendency to respond directly to the distinctive feature when responding on the positive display and greatly reduced errors. In the feature-negative case, explicit feedback for incorrect responses increased the tendency to avoid the distinctive feature in favor of a common feature when responding on the negative display. In this case, however, consistent avoidance of the distinctive feature on the negative display was not always followed by the development of consistent choice of the positive display, and errors continued to occur at a high rate throughout training for most subjects. These results reflect the difference in the structure of the feature-positive and feature-negative tasks.     

Effect of proximity of elements on the feature-positive effect

Eight groups of pigeons were trained to discriminate between two stimulus displays that could be differentiated only by a single distinctive feature on one of the displays. For half of the pigeons, responses to displays containing the distinctive feature were reinforced (feature-positive), and for the remaining pigeons responses to displays without the distinctive feature were reinforced (feature-negative). The pigeons were further grouped so that half were presented displays in which the distinctive feature was in close proximity to other features (compact displays) and half were presented displays in which the features were not close together (distributed displays). Pigeons in the feature-positive groups localized responses on the distinctive feature of the displays and seldom responded to displays without the distinctive feature. Pigeons in the distributed feature-negative groups localized responses on features common to the two displays and did not learn the discrimination. Compacting the displays facilitated discrimination performance for the subjects in the feature-negative condition. Tests carried out in extinction indicated that responding in the compact feature-negative group was largely controlled by pattern rather than by the individual elements on the display.     

Spatial variables and the feature-positive effect

Pigeons were trained in discriminations where S+ and S? shared common visual elements but were distinguished by the presence or absence of a visual feature. When S+ contained the distinctive feature (the feature-positive condition), the pigeons rapidly learned to withold responses to S?: The learning rate was not markedly affected by changes in the relative size of the feature and the common element. When S? contained the distinctive feature (the feature-negative condition), the pigeons learned to withhold responses to S? more rapidly when the feature was large relative to the common element than when the feature was small relative to the common element. This effect was observed when the stimuli were present at the locus of response and when the stimuli were remote from the locus of response. To explain these results, and the results of generalization tests, it was proposed that the effectiveness of the feature in suppressing responding to the common element during feature-negative training depended on the relative size of the feature and the common element.     

The non-occurrence of a stimulus as a signal

Two groups of rats were rewarded for pressing a panel following a varying number of bar presses; a signal following the bar press indicated whether or not a panel press would be rewarded if made before the next bar press. For one group the signal indicating reward was a flash of light, for the other it was the non-occurrence of the flash (in the latter case the flash thus signalled non-reward). The first group learned to withold panel presses except when reward was signalled, but the second group did not. This result is related to the “feature positive effect” of Jenkins and Sainsbury (1970), in which pigeons failed to withold pecks at a negative stimulus display when it was the same as the positive display except for the addition of a distinctive feature.     

Common and distinctive features in stimulus similarity: A modified version of the contrast model

Featural representations of similarity data assume that people represent stimuli in terms of a set of discrete properties. In this article, we consider the differences in featural representations that arise from making four different assumptions about how similarity is measured. Three of these similarity models— the common features model, the distinctive features model, and Tversky’s seminal contrast model—have been considered previously. The other model is new and modifies the contrast model by assuming that each individual feature only ever acts as a common or distinctive feature. Each of the four models is tested on previously examined similarity data, relating to kinship terms, and on a new data set, relating to faces. In fitting the models, we have used the geometric complexity criterion to balance the competing demands of data-fit and model complexity. The results show that both common and distinctive features are important for stimulus representation, and we argue that the modified contrast model combines these two components in a more effective and interpretable way than Tversky’s original formulation.     

The effect of a fading procedure upon the acquisition of control by an overshadowed auditory feature

Pigeons' key-pecking responses were reinforced in the presence of a compound stimulus that consisted of an auditory feature (a tone) and a visual feature (a light) and non-reinforced in the presence of a compound stimulus that was either a noise and a dark key, or noise and a light. In the condition where reinforcement trials differed from non-reinforcement trials on the basis of both auditory and visual features, the tone exerted very little control over responding on test. In the condition where reinforcement differed from non-reinforcement trials solely on the basis of the auditory features, an abrupt and a gradual introduction of the visual feature of the negative stimulus, a light, were compared for their effect upon control in the compounds. The tone acquired strong control in both cases. Evidence indicated that the tone had acquired control in the gradual condition without the occurrence of responses to the negative stimulus. An incidental finding was that when the negative stimulus consisted of a noise and a light, which was introduced abruptly, responding over the light dimension with tone, on test, was peaked at a point other than that light value used as positive and negative during training.     

Happy with a difference, unhappy with an identity: Observers’ mood determines processing depth in visual search

Visual search for feature targets was employed to investigate whether the mechanisms underlying visual selective attention are modulated by observers’ mood. The effects of induced mood on overall mean reaction times and on changes and repetitions of target-defining features and dimensions across consecutive trials were measured. The results showed that reaction times were significantly slower in the negative than in the positive and neutral mood groups. Furthermore, the results demonstrated that the processing stage that is activated to select visual information in a feature search task is modulated by the observer’s mood. In participants with positive or neutral moods, dimension-specific, but no feature-specific, intertrial transition effects were found, suggesting that these observers based their responses on a salience signal coding the most conspicuous display location. Conversely, intertrial effects in observers in a negative mood were feature-specific in nature, suggesting that these participants accessed the feature identity level before responding.     

Distinctive feature control of decision time: Same-different judgments of simultaneously heard phonemes

Ss decided whether dichotically presented consonant or vowel phonemes were “same” or “different” in a reaction time (RT) task. Results indicated that “different” responses were made on the basis of a serial, self-terminating scan of distinctive feature differences between phonemes. “Same” responses were considered too fast to be accounted for by this process and were discussed as a separate parallel process. Recognition of dichotically presented syllables appears to involve a third process in which phonemes are also identified in terms of their distinctive features.     

Pictorial experience as a factor in the recognition of incomplete pictures

The study investigated the effects of distinctive feature emphasis and familiarity with pictorial materials on the recognition of incomplete pictures of common objects in three cultural groups. Contrary to a familiarity hypothesis Kxoe Bushman pre-school children were superior to both Zulu and white English-speaking Sth African pre-school children in recognizing fragmented pictures. Cultural differences were interpreted in terms of response strategies mediated by ecological and social variables. Children in each cultural group made effective use of distinctive features to aid recognition of fragmented pictures. The implications of these findings for the development of teaching materials were examined.     

Differential autoshaping to common and distinctive elements of positive and negative discriminative stimuli

The learning by hungry pigeons of a discrimination between two successively presented compound visual stimuli was investigated using a two-key autoshaping procedure. Common and distinctive stimulus elements were simultaneously presented on separate keys and either followed by food delivery, S+, or not, S−. The subjects acquired both between-trial and within-trial discriminations. On S+ trials, pigeons pecked the distinctive stimulus more than the common stimulus; before responding ceased on S− trials, they pecked the common stimulus more than the distinctive one. Mastery of the within-display discrimination during S+ trials preceded mastery of the between-trials discrimination. These findings extend the Jenkins-Sainsbury analysis of discriminations based upon a single distinguishing feature to discriminations in which common and distinctive elements are associated with both the positive and negative discriminative stimuli. The similarity of these findings to other effects found in autoshaping—approach to signals that forecast reinforcement and withdrawal from signals that forecast nonreinforcement—is also discussed.     

Features of graded category structure: Generalizing the family resemblance and polymorphous concept models

Many real-world categories contain graded structure: certain category members are rated as more typical or representative of the category than others. Research has shown that this graded structure can be well predicted by the degree of commonality across the feature sets of category members. We demonstrate that two prominent feature-based models of graded structure, the family resemblance (Rosch & Mervis, 1975) and polymorphous concept models (Hampton, 1979), can be generalized via the contrast model (Tversky, 1977) to include both common and distinctive feature information, and apply the models to the prediction of typicality in 11 semantic categories. The results indicate that both types of feature information play a role in the prediction of typicality, with common features weighted more heavily for within-category predictions, and distinctive features weighted more heavily for contrast-category predictions. The same pattern of results was found in additional analyses employing rated goodness and exemplar generation frequency. It is suggested that these findings provide insight into the processes underlying category formation and representation.     

Feature-based search asymmetries in pigeons and humans

Pigeon and human subjects searched for one target item amidst a number of identical distractors. Simple line forms were used. The target differed from the distractors only in terms of the presence or absence of a feature (a line or a gap); in some experimental series, the feature was present in the target; in others, the feature was in the distractors. The pigeons pecked at the target; the human subjects either reported the presence of the target or pointed to it with a light pen. The time between display onset and this response was recorded. Varied across experimental conditions were the number of distractors in the display, the nature of the stimulus forms, and certain procedural parameters; five conditions were run with pigeons and three with humans. Under all test conditions, the results from the human subjects replicated the previously reported search-asymmetry effect. That is, search speed was greater and decreased less with display size when the target bore the feature (line or gap) than when the distractors bore the feature; both yes/no and localization-response conditions yielded this effect. However, pigeons failed to show search asymmetry; neither line nor gap in a target facilitated search. The results suggest that early visual processing differs for pigeons and humans, that pigeon features differ from human features, or that search asymmetry was eliminated by the long practice given the pigeons.     

Associative learning in baboons (Papio papio) and humans (Homo sapiens): species differences in learned attention to visual features

We examined attention shifting in baboons and humans during the learning of visual categories. Within a conditional matching-to-sample task, participants of the two species sequentially learned two two-feature categories which shared a common feature. Results showed that humans encoded both features of the initially learned category, but predominantly only the distinctive feature of the subsequently learned category. Although baboons initially encoded both features of the first category, they ultimately retained only the distinctive features of each category. Empirical data from the two species were analyzed with the 1996 ADIT connectionist model of Kruschke. ADIT fits the baboon data when the attentional shift rate is zero, and the human data when the attentional shift rate is not zero. These empirical and modeling results suggest species differences in learned attention to visual features. Received: 30 May 1998 / Accepted after revision: 13 September 1998     

Mechanisms of aural encoding: VI. Consonants and vowels are remembered as subsets of distinctive features

An attempt was made to examine the manner in which consonants and vowels are coded in short-term memory. under identical recall conditions. Ss were presented with sequences of consonant-vowel digrams for serial recall. Sequences were composed of randomly presented consonants paired with/a/ or randomly presented vowels paired with /d/. Halle’s distinctive feature system was used to generate predictions concerning the frequency of intrusion errors. among phonemes. These predictions were based on the assumption that phonemes are discriminated in memory in terms of their component distinctive features, so that intrusions should most frequently occur between phonemes sharing similar distinctive features. The analysis of intrusion errors revealed that each consonant and vowel phoneme was coded m short-term memory by a particular combination of distinctive features which differed from one phoneme to another. A given phoneme was coded by the same set of distinctive features regardless of the number of syllables in the sequence. However, distinctive feature theories were not able to predict the frequency of intrusion errors for phonemes presented in the middle serial positions of a sequence with 100% accuracy. The results of the experiment support the notion that consonant and vowel phonemes are coded in a similar manner in STM and that this coding involves the retention of a specific set of distinctive features for each phoneme.     

Identification versus discrimination of distinctive features in speech perception

Dichotically presented CV syllables were presented to 24 normal subjects under two contrasting performance requirements: consonant identification and consonant discrimination. In the identification task, subjects made more errors identifying stimuli distinguished by two as opposed to one distinctive feature. Conversely, in the discrimination task, subjects made more errors for stimuli distinguished by one as opposed to two distinctive features. It was proposed that the results of the identification task reflect the degree of information load in short-term memory. The results of the discrimination task, on the other hand, reflect the degree of perceptual similarity between contrasting stimuli. Both results were accounted for within a distinctive feature framework. Analyses of the individual features which comprised one and two feature distinctions demonstrated the perceptual prominence of the feature [voice] in contrast to the two place features [compact] and [grave].     

Search asymmetry: a diagnostic for preattentive processing of separable features

The search rate for a target among distractors may vary dramatically depending on which stimulus plays the role of target and which that of distractors. For example, the time required to find a circle distinguished by an intersecting line is independent of the number of regular circles in the display, whereas the time to find a regular circle among circles with lines increases linearly with the number of distractors. The pattern of performance suggests parallel processing when the target has a unique distinguishing feature and serial self-terminating search when the target is distinguished only by the absence of a feature that is present in all the distractors. The results are consistent with feature-integration theory (Treisman & Gelade, 1980), which predicts that a single feature should be detected by the mere presence of activity in the relevant feature map, whereas tasks that require subjects to locate multiple instances of a feature demand focused attention. Search asymmetries may therefore offer a new diagnostic to identify the primitive features of early vision. Several candidate features are examined in this article: Colors, line ends or terminators, and closure (in the sense of a partly or wholly enclosed area) appear to be functional features; connectedness, intactness (absence of an intersecting line), and acute angles do not.     

The contribution of category relationships to similarity ratings.

Two experiments produced evidence that category relationships contribute to similarity ratings. In Experiment 1, participants gave similarity ratings with respect to a semantic category (VEGETABLE) and a set of exemplars, some of which were members of the category (e.g., BROCCOLI) and some of which were not (e.g., BANANA). A regression analysis was used to predict the similarity ratings in terms of numbers of common and distinctive features, as reported by other participants. Perceived similarity was greater for examplars that were members of the category, independently of feature overlap. Experiment 2 examined similarity ratings with respect to pairs of exemplars. In some cases, both exemplars were members of the same category (e.g., BROCCOLI/CUCUMBER). In other cases, one exemplar was a member of the category and the other was not (e.g., BROCCOLI/BANANA). A regression analysis was used to predict the similarity ratings in terms of numbers of common and distinctive features. Perceived similarity was greater when both exemplars were members of the same category, independently of feature overlap.     

Contrasting effects of feature-based statistics on the categorisation and basic-level identification of visual objects

Conceptual representations are at the heart of our mental lives, involved in every aspect of cognitive functioning. Despite their centrality, a long-standing debate persists as to how the meanings of concepts are represented and processed. Many accounts agree that the meanings of concrete concepts are represented by their individual features, but disagree about the importance of different feature-based variables: some views stress the importance of the information carried by distinctive features in conceptual processing, others the features which are shared over many concepts, and still others the extent to which features co-occur. We suggest that previously disparate theoretical positions and experimental findings can be unified by an account which claims that task demands determine how concepts are processed in addition to the effects of feature distinctiveness and co-occurrence. We tested these predictions in a basic-level naming task which relies on distinctive feature information (Experiment 1) and a domain decision task which relies on shared feature information (Experiment 2). Both used large-scale regression designs with the same visual objects, and mixed-effects models incorporating participant, session, stimulus-related and feature statistic variables to model the performance. We found that concepts with relatively more distinctive and more highly correlated distinctive relative to shared features facilitated basic-level naming latencies, while concepts with relatively more shared and more highly correlated shared relative to distinctive features speeded domain decisions. These findings demonstrate that the feature statistics of distinctiveness (shared vs. distinctive) and correlational strength, as well as the task demands, determine how concept meaning is processed in the conceptual system.