Effects of categorical labels on similarity judgments: a critical analysis of similarity-based approaches

Our goal in the present study was to evaluate the claim that category labels affect children's judgments of visual similarity. We presented preschool children with discriminable and identical sets of animal pictures and asked them to make perceptual judgments in the presence or absence of labels. Our findings indicate that children who are asked to make perceptual judgments about identical items judge discriminable items less accurately when making subsequent similarity judgments. Thus, labels do not generally affect children's perceptual similarity judgments; rather, children's reliance on labels to make similarity judgments appears to be attributable to flaws in the methodological approaches used in prior studies. These results have implications for the role of perceptual and conceptual information in children's categorization and induction.     

Similarity Judgment Within and Across Categories: A Comprehensive Model Comparison

Similarity is one of the most important relations humans perceive, arguably subserving category learning and categorization, generalization and discrimination, judgment and decision making, and other cognitive functions. Researchers have proposed a wide range of representations and metrics that could be at play in similarity judgment, yet have not comprehensively compared the power of these representations and metrics for predicting similarity within and across different semantic categories. We performed such a comparison by pairing nine prominent vector semantic representations with seven established similarity metrics that could operate on these representations, as well as supervised methods for dimensional weighting in the similarity function. This approach yields a factorial model structure with 126 distinct representation-metric pairs, which we tested on a novel dataset of similarity judgments between pairs of cohyponymic words in eight categories. We found that cosine similarity and Pearson correlation were the overall best performing unweighted similarity functions, and that word vectors derived from free association norms often outperformed word vectors derived from text (including those specialized for similarity). Importantly, models that used human similarity judgments to learn category-specific weights on dimensions yielded substantially better predictions than all unweighted approaches across all types of similarity functions and representations, although dimension weights did not generalize well across semantic categories, suggesting strong category context effects in similarity judgment. We discuss implications of these results for cognitive modeling and natural language processing, as well as for theories of the representations and metrics involved in similarity.     

Framing of comparisons in research and practice

We discuss important ways in which the framing of a question or a decision can have significant effects on responses. Several areas of research and theory that indicate the scope and importance of these effects are presented. The major discussion, however, is limited to how respnses to comparative judgments and constrained by the other or sequence in which the direction of comparison occurs. Direction of comparison is first discussed with respect to similarity judgments, in which a model taking direction of comparison into account is used to explain both degrees of perceived similarity and asymmetries in judgment. A feature-matching model of comparison is then developed and is used to analyze preference judgments, judgments of change, and self-relevant judgments involving comparisons of the self with other people, with counterfactual alternatives to reality, and with the self at a different point in time. This model is employed to account for predecisional conflict and difficulty, the preference that one makes, and the degree of satisfaction or regret that one feels subsequent to making a comparison. Finally, the potential for applying this framework to areas of consumer behavior, political decision making, and psychotherapy is explored. Techniques for taking advantage of direction of comparison effects as well as methods for eliminating the biases associated with these effects are discussed.     

Effects of categorical labels on similarity judgments: a critical evaluation of a critical analysis: comment on Noles and Gelman (2012)

Noles and Gelman (2012) attempt to critically reevaluate the claim that linguistic labels affect children's judgments of visual similarity. They report results of an experiment that used a modified version of Sloutsky and Fisher's (2004) task and conclude that "labels do not generally affect children's perceptual similarity judgments; rather, children's reliance on labels to make similarity judgments appears to be attributable to flaws in the methodological approaches used in prior studies" (p. 890). In this comment we demonstrate that these conclusions are based on inadequate analyses and reporting of statistics and are inconsistent with Noles and Gelman's own data. Instead, their data show a perfect replication of Sloutsky and Fisher (2004) and present further evidence that linguistic labels affect similarity judgment of young children.     

On cross-modal similarity: the perceptual structure of pitch, loudness, and brightness

Examined how pitch and loudness correspond to brightness. In the Experiment 1, 16 Ss identified which of 2 lights more resembled each of 16 tones; in Experiment 2, 8 of the same 16 Ss rated the similarity of lights to lights, tones to tones, and lights to tones. (1) Pitch and loudness both contributed to cross-modal similarity, but for most Ss pitch contributed more. (2) Individuals differed as to whether pitch or loudness contributed more; these differences were consistent across matching and similarity scaling. (3) Cross-modal similarity depended largely on relative stimulus values. (4) Multidimensional scaling revealed 2 perceptual dimensions, loudness and pitch, with brightness common to both. A simple quantitative model can describe the cross-modal comparisons, compatible with the view that perceptual similarity may be characterized through a malleable spatial representation that is multimodal as well as multidimensional.     

Prior Attitudes,Salient Value Similarity,and Dimensionality: Toward an Integrative Model of Trust in Risk Regulation1

Although it is widely recognized that trust plays an important role in people's responses to various risks, there is still considerable conceptual disagreement about the different aspects of trust. There are at least 3 different approaches to trust: (a) dimensional, (b) salient value similarity, and (c) associationist. Three British datasets on genetically modified food were used to test the plausibility of a causal model that integrates these approaches. It appears that value similarity can be predicted by a combination of prior attitudes and perceived attitudes of the other, and that value similarity precedes other important trust judgments. The study suggests that various risk‐relevant judgments are expressions of a more general attitude toward genetically modified food, and raises questions about the usefulness of detailed modelling.     

相似性判断和差异性判断的不对称性及其机制探索

本文阐述了相似性判断和差异性判断不对称性的相关研究。首先介绍了不对称性的发展及理论解释。其次,根据事件相关电位的研究结果,阐述了图-词信息的双编码模型对相似性判断的影响。接着分析和介绍了相似性判断的双编码过程及其时间进程。最后,作者提出有待深入研究的几个问题：（1）差异性策略在分类和推理中的应用;（2）相似性双编码过程的实证研究及其应用;（3）从发展认知神经科学的角度探讨不对称的认知神经机制。     

Are the effects of attention on speed judgments genuinely perceptual?

A number of studies in recent years have suggested that exogenous and endogenous attention might enhance the perceived magnitude of various perceptual attributes, such as contrast and motion speed. Those studies have generally used comparative judgments as a measure to assess the point of subjective equality; however, similarity judgments have been proposed as possibly less prone to decision biases (Schneider & Komlos, 2008). In three experiments, using a similarity judgment task, we did not find any evidence of motion speed enhancement by exogenous attention. We suggest that the effect revealed by comparative judgments arises at the decisional, rather than the perceptual, stage.     

Similarity and proximity: When does close in space mean close in mind?

People often describe things that are similar as close and things that are dissimilar as far apart. Does the way people talk about similarity reveal something fundamental about the way they conceptualize it? Three experiments tested the relationship between similarity and spatial proximity that is encoded in metaphors in language. Similarity ratings for pairs of words or pictures varied as a function of how far apart the stimuli appeared on the computer screen, but the influence of distance on similarity differed depending on the type of judgments the participants made. Stimuli presented closer together were rated more similar during conceptual judgments of abstract entities or unseen object properties but were rated less similar during perceptual judgments of visual appearance. These contrasting results underscore the importance of testing predictions based on linguistic metaphors experimentally and suggest that our sense of similarity arises from our ability to combine available perceptual information with stored knowledge of experiential regularities.     

Exploring children’s face-space: A multidimensional scaling analysis of the mental representation of facial identity

We explored differences in the mental representation of facial identity between 8-year-olds and adults. The 8-year-olds and adults made similarity judgments of a homogeneous set of faces (individual hair cues removed) using an “odd-man-out” paradigm. Multidimensional scaling (MDS) analyses were performed to represent perceived similarity of faces in a multidimensional space. Five dimensions accounted optimally for the judgments of both children and adults, with similar local clustering of faces. However, the fit of the MDS solutions was better for adults, in part because children’s responses were more variable. More children relied predominantly on a single dimension, namely eye color, whereas adults appeared to use multiple dimensions for each judgment. The pattern of findings suggests that children’s mental representation of faces has a structure similar to that of adults but that children’s judgments are influenced less consistently by that overall structure.     

Connecting instances to promote children's relational reasoning

The practice of learning from multiple instances seems to allow children to learn about relational structure. The experiments reported here focused on two issues regarding relational learning from multiple instances: (a) what kind of perceptual situations foster such learning and (b) how particular object properties, such as complexity and similarity, interact with relational learning. Two kinds of perceptual situations were of interest here: simultaneous view, where instances are viewed at once, and sequential view, where instances are viewed one at a time (one right after the other). We examined the influence of particular perceptual situations and object properties using two tests of relational reasoning: a common match-to-sample task, where new instances are compared with a common sample, and a variable match-to-sample task, where new instances are compared with a sample that varies on each trial. Experiments 1 and 2 indicate that simultaneous presentation of even highly dissimilar instances, one simple and one complex, effectively connects them together and improves relational generalization in both match-to-sample tasks. Experiment 3 shows that simple samples are more effective than complex ones in the common match-to-sample task. However, when one instance is not used a common sample and various pairs of instances are simply compared, as in Experiment 4, simple and rich instances are equally effective at promoting relational learning. These results bear on our understanding of how children connect instances and how those initial connections affect learning and generalization.     

Constructing Semantic Models From Words,Images, and Emojis

A number of recent models of semantics combine linguistic information, derived from text corpora, and visual information, derived from image collections, demonstrating that the resulting multimodal models are better than either of their unimodal counterparts, in accounting for behavioral data. Empirical work on semantic processing has shown that emotion also plays an important role especially in abstract concepts; however, models integrating emotion along with linguistic and visual information are lacking. Here, we first improve on visual and affective representations, derived from state-of-the-art existing models, by choosing models that best fit available human semantic data and extending the number of concepts they cover. Crucially then, we assess whether adding affective representations (obtained from a neural network model designed to predict emojis from co-occurring text) improves the model's ability to fit semantic similarity/relatedness judgments from a purely linguistic and linguistic–visual model. We find that, given specific weights assigned to the models, adding both visual and affective representations improves performance, with visual representations providing an improvement especially for more concrete words, and affective representations improving especially the fit for more abstract words.     

Intertemporal Similarity: Discounting as a Last Resort

Standard models of intertemporal choice assume that individuals discount future payoffs by integrating reward amounts and time delays to generate a discounted value. Alternative models propose that, rather than integrate across them, individuals compare within attributes (amounts and delays) to determine if differences in one attribute outweigh differences in another attribute. For instance, the similarity model 1) compares the two reward amounts to determine whether they are similar, 2) compares the similarity of the two time delays, and then 3) makes a decision based on these similarity judgments. Here, I tested discounting models against attribute‐based models that use similarity judgments to make choices. I collected intertemporal choices and similarity judgments for the reward amounts and time delays from participants in three experiments. All experiments tested the ability of discounting and similarity models to predict intertemporal choices. Model generalization analyses showed that the best predicting models started with similarity judgments and then, if similarity failed to make a prediction, resorted to discounting models. Similarity judgments also matched intertemporal choice data demonstrating both the magnitude and sign effects, thereby accounting for behavioral data that contradict many discounting models. These results highlight the possibility that attribute‐based models such as the similarity models provide alternatives to discounting that may offer insights into the process of making intertemporal choices. Copyright © 2015 John Wiley & Sons, Ltd.     

Bilingualism and the perception of professional concepts

A selected set of professional concepts was subjected to analysis through two separate multidimensional scaling techniques, the INDSCAL and TORSCA models, to evaluate the intergroup perceptual differences of four experimental groups, made up of unilingual French, unilingual English and bilingual students. The linguistic relativism thesis provided the research hypotheses on the relationship between language access and usage and concept perception. The multidimensional scaling techniques were applied to the matrix of subjects' similarity judgments on pairs of concepts, thus enabling the identification of three dimensions. The dimensions were labelled as conjunctive, relational and disjunctive, and were assumed to be related to the criteria employed by the subjects in their similarity rankings. An analysis of variance of the individual saliences on each dimension provided evidence of linguistic relativism for both the relational and disjunctive dimensions. These findings support the contention that unilingual speakers of separate languages differ from each other and from bilingual speakers in their perception of professional concepts.     

On the relation between perceptual priming and recognition memory

Two experiments evaluated the hypothesis that perceptual fluency is used to infer prior occurrence. Subjects heard (Experiment 1) or saw (Experiment 2) a list of words and then were presented in the same modality with both these and other words twice in succession: first in a more or less impoverished fashion, and then in clear fashion. For the first of these two presentations, the subjects tried to identify the word; for the second, they gave a recognition judgement. As predicted by the perceptual fluency hypothesis, and as has been found in previous research, the recognition judgments were more positive for identified words than for unidentified words. However, degree of impoverishment, by which apparent perceptual fluency was brought under experimental control, did not affect the recognition judgments. The perceptual fluency hypothesis was therefore not supported, and the observed relation between identification and recognition was attributed to an item selection effect.     

"Ratio" and "difference" judgments for length, area, and volume: are there two classes of sensory continua?

Subjects were required to judge ratios and differences of (a) line length for pairs of lines, (b) area for pairs of squares, and (c) volume for pairs of cubes. Nonmetric analyses of these judgments indicated that all subjects were able to make consistent ratio judgments for all three continua. Many of the subjects, when asked to judge subjective differences, however, performed as if they were judging subjective ratios rather than differences. The data for the few subjects who appeared to be judging subjective differences were not consistent across subjects and conditions. Previous studies of ratio and difference judgments of loudness and heaviness, on the other hand, showed the opposite pattern, in that subjects most often behaved as if they were judging sensory differences when asked to judge sensory ratios. We propose that ratio judgments are more natural to perceptual continua along which stimuli are easily "decomposed" into a number of smaller perceptual units.     

The relative sensitivities of same-different and identification judgment models to perceptual dependence

Probabilistic models of same-different and identification judgments are compared (within each paradigm) with regard to their sensitivity to perceptual dependence or the degree to which the underlying psychological dimensions are correlated. Three same-different judgment models are compared. One is a step function or decision bound model and the other two are probabilistic variants of a similarity model proposed by Shepard. Three types of identification models are compared: decision bound models, a probabilistic multidimensional scaling model, and probabilistic models based on the Shepard-Luce choice rule. The decision bound models were found to be most sensitive to perceptual dependence, especially when there is considerable distributional overlap. The same-different model based on the city-block metric and an exponential decay similarity function, and the corresponding identification model were found to be particularly insensitive to perceptual dependence. These results suggest that if a Shepard-type similarity function accurately describes behavior, then under typical experimental conditions it should be difficult to see the effects of perceptual dependence. This result provides strong support for a perceptualindependence assumption when using these models. These theoretical results may also play an important role in studying different decision rules employed at different stages of identification training.We thank Robert Melara, Jerome Busemeyer and three anonymous reviewers for comments on an earlier draft of this paper.     

Multidimensional similarity: an experimental and theoretical comparison of vector,distance, and set theoretical models : I. Models and internal consistency of data

In an experiment (conducted by the senior author) subjects were instructed to give three kinds of quantitative judgment on pairs of simple geometrical stimuli: similarity, commonality ratio and magnitude ratio. It was assumed that the different kinds of judgments could be interpreted in terms of one cognitive structure, and that formal models for each of these judgments should be validated by predicting one kind of judgment from knowledge of another kind of judgment. The present report proposes two different, though related, systems of formal models for similarity, commonality and magnitude ratio judgments. One, called E-model, interprets the judgments in terms of set- and vector-representations, and connections between them; the other, called R-model, interprets the judgments in terms of set- and distance-representations and their connections. The results are slightly in favor of the E-model. In a subsequent paper, the authors will report the results of multidimensional analyses for the same data, which slightly favor the R-model.     

Shape constancy and a perceptual bias towards symmetry

In attempting to judge whether a briefly presented rotated figure was symmetrical (circle or square) or asymmetrical (ellipse or rectangle), a bias towards perceiving symmetry was evident. This bias appears to be perceptual rather than judgmental in origin in that it was affected neither by variations in the probability of occurrence of symmetrical figures nor by familiarization with the stimuli. A characterization of the stimulus similarity space based on the shape-slant invariance hypothesis is presented, and a model which incorporates both the similarity space and the symmetry bias is proposed. The model can account for our paradoxical finding that the more asymmetrical some projected images, the more likely they will be perceived as symmetrical. This model can also account for Massaro’s (1973, 1975) results on shape constancy. However, Massaro’s theory was not supported in that extension in depth did not account for increased difficulty of shape judgments.