Abstract coherent categories.

Many studies have demonstrated the importance of the knowledge that interrelates features in people's mental representation of categories and that makes our conception of categories coherent. This article focuses on abstract coherent categories, coherent categories that are also abstract because they are defined by relations independently of any features. Four experiments demonstrate that abstract coherent categories are learned more easily than control categories with identical features and statistical structure, and also that participants induced an abstract representation of the category by granting category membership to exemplars with completely novel features. The authors argue that the human conceptual system is heavily populated with abstract coherent concepts, including conceptions of social groups, societal institutions, legal, political, and military scenarios, and many superordinate categories, such as classes of natural kinds.

Concept learning and feature interpretation.

Models of categorization often assume that people classify new instances directly on the basis of the presented, observable features. Recent research, however, has suggested that the coherence of a category may depend in part on more abstract features that can link together observable features that might otherwise seem to have little similarity. Thus, category learning may also involve the determination of the appropriate abstract features that underlie a category and link together the observable features. We show in four experiments that observable features of a category member are often interpreted as congruent with abstract features that are suggested by observable features of other highly available category members. Our discussion focuses on the implications of these findings for future research.

The effects of category use on learned categories.

Frequently, people learn to classify instances of a concept and later learn additional information about the concept. What is the effect of this later learning on the original classification? In five experiments, this issue was investigated with a common classification paradigm in which symptom sets were classified into disease categories. After learning to classify these sets, the subjects learned to use the category to decide what treatment should be given for a symptom set. The symptoms that were important for the treatments were later classified by disease more accurately and were generated earlier from the disease name. However, this effect occurred only if the category representation was activated during the learning of the treatments. Thus, later learning about a particular use of the concept can sometimes affect the original classification.

Induction with cross-classified categories.

One of the main functions of categories is to allow inferences about new objects. However, most objects are cross-classified, and it is not known whether and how people combine information from these different categories in making inferences. In six experiments, food categories, which are strongly cross-classified (e.g., a bagel is both a bread and a breakfast food), were studied. For each food, the subjects were told fictitious facts (e.g., 75% of breads are subject to spoilage from Aspergillus molds) about two of the categories to which it belonged and then were asked to make an inference about the food (e.g., how likely is a bagel to be subject to spoilage from Aspergillus molds?). We found no more use of multiple categories in these cases of cross-classification than in ambiguous classification, in which it is uncertain to which category an item belongs. However, some procedural manipulations did markedly increase the use of both categories in inferences, primarily those that focused the subjects' attention on the critical feature in both categories.

Food for thought: cross-classification and category organization in a complex real-world domain.

Seven studies examined how people represent, access, and make inferences about a rich real-world category domain, foods. The representation of the category was assessed by category generation, category ratings, and item sortings. The first results indicated that the high-level category of foods was organized simultaneously by taxonomic categories for the kind of food (e.g., vegetables, meats) and script categories for the situations in which foods are eaten (e.g., breakfast foods, snacks). Sortings were dominated by the taxonomic categories, but the script categories also had an influence. The access of the categories was examined both by a similarity rating task, with and without the category labels, and by a speeded priming experiment. In both studies, the script categories showed less access than the taxonomic categories, but more than novel ad hoc categories, suggesting some intermediate level of access. Two studies on induction found that both types of categories could be used to make a wide range of inferences about food properties, but that they were differentially useful for different kinds of inferences. The results give a detailed picture of the use of cross-classification in a complex domain, demonstrating that multiple categories and ways of categorizing can be used in a single domain at one time.

Distinct patterns of motor unit behavior during muscle spasms in spinal cord injured subjects.

Surface electromyograms (EMG) and force were recorded during repeated involuntary spasms of paralyzed triceps surae muscles of four men with chronic cervical spinal cord injury. The firing rates of 78 medial gastrocnemius (MG) motor units also were recorded intramuscularly with tungsten microelectrodes. Spasms typically involved a relatively rapid rise, then a more gradual fall in triceps surae EMG and torque. Motor unit firing rates either increased and then decreased with the spasm intensity (54%) or were relatively constant (26%), firing mainly at 2-10 Hz. The remaining units (20%) produced trains that included one or several doublets. Mean peak spasm firing rates were 18 +/- 9 Hz (mean +/- SD) for rate modulated units and 11 +/- 10 Hz for units with little or no rate modulation. Some motor units fired at rates comparable with those recorded previously during maximum voluntary contractions performed by intact subjects. Others fired at rates below the minimum usually seen when normal units are first recruited (< 6 Hz). Doublets (interspike interval < 10 ms) often repeated every 123-333 ms, or were interspersed in trains firing at low steady rates (< 11 Hz). This study shows that rate coding for many motor units appears to be similar whether descending motor input is intact or whether it has been reduced severely by spinal cord injury. In contrast, rate modulation in other units appears to depend mainly on voluntary motor commands.

Category-based predictions: influence of uncertainty and feature associations.

Four experiments examined how people make inductive inferences using categories. Subjects read stories in which 2 categories were mentioned as possible identities of an object. The less likely category was varied to determine if people were using it, as well as the most likely category, in making predictions about the object. Experiment 1 showed that even when categorization uncertainty was emphasized, subjects used only 1 category as the basis for their prediction. Experiments 2-4 examined whether people would use multiple categories for making predictions when the feature to be predicted was associated to the less likely category. Multiple categories were used in this case, but only in limited circumstances; furthermore, using multiple categories in 1 prediction did not cause subjects to use them for subsequent predictions. The results increase the understanding of how categories are used in inductive inference.

Human motor unit activity during induced muscle cramp.

Muscle cramp was induced in the medial head of the gastrocnemius muscle in four of seven subjects using unloaded maximal voluntary contraction of the triceps surae in the shortened position. Surface electromyography over the medial and lateral heads of gastrocnemius and the soleus muscles demonstrated that the muscle activity during cramp was localized to part or all of the medial head of the gastrocnemii. In the same muscle, a tungsten electrode was used to record from 200 motor units during 16 episodes of cramp and 871 units during 26 voluntary contractions. For the first 30 s, significantly higher motor unit firing rates were recorded during cramp compared with unloaded voluntary contractions. Motor unit firing rates were also more variable during cramp. When the cramped muscle was stretched forcibly to break the cramp, motor unit activity increased in all the triceps surae muscles. In some experiments, the Achilles tendon of five subjects was vibrated for 50 s before and after voluntary contraction or cramp. The tonic vibration reflex (TVR) was depressed or absent after four episodes of cramp but it was unchanged after voluntary contraction. These data are interpreted to indicate that motor units are involved in ordinary muscle cramp. A positive feedback loop between peripheral afferents and alpha motor neurons, mediated by changes in presynaptic input, is a possible mechanism underlying muscle cramp.

Predicting features for members of natural categories when categorization is uncertain.

An important function of concepts to allow prediction of unseen features. A Bayesian account of feature prediction suggests that people will consider all the categories an object could belong to when they judge the likelihood that the object has a feature. The judgment and decision literature suggests that they may instead use a simpler heuristic in which they consider only the most likely category. In 3 experiments, no evidence was found that participants took into account alternative categories as well as the most likely one when they judged feature probabilities for familiar objects in meaningful contexts. These results, in conjunction with those of Murphy and Ross (1994), suggest that although people may consider alternative categories in certain limited situations, they often do not. Reasons for why the use of alternative categories may be relatively rare are discussed, and conditions under which people may take alternative categories into account are outlined.

Comparison-based learning: effects of comparing instances during category learning.

When learning about a category, people often compare new instances with similar old instances and notice features common to the compared instances. Five experiments demonstrate that such comparisons cause features common to compared instances to be considered more important for the category than equally frequent features that are not common to compared instances. Experiment 1 shows that what is learned depends on which instances are compared. Experiment 2 investigates the conditions under which comparison-based learning occurs. The next experiments find that these comparisons affect subjective feature frequency (Experiment 3) and sensitivity to feature correlations (Experiment 4). Experiment 5 shows that comparisons during early learning affect what is learned from later instances. The discussion focuses on the implications for models of category representation.

Predictions from uncertain categorizations.

Eleven experiments investigated how categorization influences feature prediction. Subjects were provided with sets of categorized exemplars, which they used to make predictions about properties of new exemplars. Because the categories were provided for subjects, this method allowed a test of categorization and prediction processes, bypassing initial concept formation and memory. The experiments tested a Bayesian rule of prediction according to which (1) predictions of an object's features are based on information from multiple categories, and (2) features are treated as independent of one another. With one exception, the studies found evidence against both of these claims. Subjects did not generally alter their predictions as a function of information outside the most likely "target" category. In addition, feature relations had reliable effects on these predictions. We discuss the implications of these results for understanding how categories are used in drawing inferences.

Encoding effects of remindings.

Understanding requires prior knowledge. Most theories and empirical work on understanding have focused on the use of general prior knowledge. In the experiments presented here, we examined the role of specific prior episodes in understanding. Subjects read ambiguous stories that had some superficial cue to make them think back to an earlier story. Experiment 1 shows that a superficial cue to an earlier story affects how the new story is interpreted. Experiment 2, in which a reading-time measure was used, provides evidence that this effect of the earlier story occurs during the initial understanding of the new story. We discuss the implications of these results for theories of understanding and the use of remindings in analogy.

Relation between power and endurance for treadmill running of short duration.

An exercise test was devised to investigate the relationship between power and endurance for treadmill running. The subjects were 19 males aged 21-25 yr (11 distance runners and 4 sprinters of provincial grade, and 4 non-competitive runners). Each subject ran to exhaustion on a treadmill at 15 km hr-1 at five different inclinations (31%-9%), giving maximum performance times in the range 10 s to 3 min. An iterative least-squares procedure was used to fit the following exponential model to each subject's data: It = I infinity + (I0-I infinity) exp (-t/tau), where It, I0 and I infinity are inclinations at time t = t, t = 0 and t----infinity, and tau is a time constant. The fit was excellent (r2 = 0.96-1.00). I0 and I infinity are interpreted as measures of maximum anaerobic (instantaneous) and maximum aerobic (continuous) power respectively. Inclinations corresponding to performance times of 10-180 s (I10-I180) were calculated from these parameters. Test-retest reliability was highest for I0-I30 (intraclass r = 0.97-0.94), lower for I60-I infinity (r = 0.89-0.84), and lowest for tau (r = 0.78). Good correlations were observed between I0-I30 and peak power in a 30 s all-out test on a cycle ergometer (r = 0.73-0.81), and between I180, I infinity and maximum oxygen consumption (r = 0.87, 0.81). The test may be useful for ranking or monitoring running performance for events of up to 1 min duration.

The access of relevant information for solving problems.

Two issues were investigated concerning the retrieval of recently acquired relevant information for solving riddle problems. First, although earlier research indicated that problem solvers often fail to retrieve relevant information unless they are explicitly informed of its relevance, Bowden (1985) suggested that uninformed subjects would benefit greatly from additional problem-solving time. In two experiments, we found that uninformed subjects solved more problems than did subjects who did not receive the information, but a simple model attributed this difference to these uninformed subjects "catching on" to the information's relevance after solving some problems. Second, the retrieval characteristics were examined by varying the proportion of problems for which clues were given. Informed subjects given clues for only half of the problems benefited (though perhaps not fully) from being informed, without incurring any cost from being misdirected on the unclued problems. The discussion focuses on some ways in which accessing relevant information may affect problem solving.

Human motor-unit recruitment during isometric contractions and repeated dynamic movements.

Spike-triggered averaging was used to determine the twitch tensions and contraction times of motor units in the abductor pollicis brevis muscle of two human subjects for two directions of isometric contraction: abduction and opposition of the thumb. During isometric contractions in each direction, the threshold force for motor-unit recruitment and the twitch amplitude were correlated linearly. These data suggested that an orderly pattern of recruitment, according to increasing twitch size, describes the function of the human abductor pollicis brevis muscle for each contraction direction. Rank order of motor-unit recruitment in each isometric contraction direction was correlated, but not identical. All units contributed tension in each direction of contraction, so no clear evidence was found for task-dependent motor units. In two subjects, motor-unit recruitment order during isometric contraction of the first dorsal interosseous and abductor pollicis brevis muscles was then compared with that of motor-unit pairs in both muscles during repetitive dynamic movements. Recruitment according to increasing twitch size was largely preserved during the repetitive opening and closing of scissors. The recruitment reversals that were observed were usually between pairs of units with similar thresholds.

Motor-unit recruitment in human first dorsal interosseous muscle for static contractions in three different directions.

Spike-triggered averaging was used to extract the twitch tensions and contraction times of 144 motor units from the first dorsal interosseous muscle of four subjects for three different directions of static contraction: abduction of the index finger, flexion of the index finger, and adduction of the thumb coupled with flexion of the index finger (hereafter referred to as adduction). Although the twitch tensions were generally largest for the abduction contraction, all units contributed tension to all three directions of contraction. A linear correlation was found for twitch tensions of motor units for the three directions of static contractions. Linear correlations were also found between twitch tension and threshold force of these motor units for each direction, which suggests that an orderly pattern of recruitment, according to increasing twitch size, adequately describes the function of human first dorsal interosseous muscle for all contraction directions. No clear evidence was found for separate groups of motor units in the muscle that were selectively activated for the different tasks. Rank order of recruitment for motor units in the three directions of contraction was correlated, but was not identical. The scatter in our data is discussed in relation to earlier reports of altered motor-unit recruitment during different movements.