No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Studies of similarity
... We replicate a study by Tversky and Gati [23], which was one of several which investigated violation of symmetry in human similarity judgements. Similarity of concepts and entities is mathematically represented using distance metrics in some coordinate space. ...
... Similarity of concepts and entities is mathematically represented using distance metrics in some coordinate space. One of the chief properties of such spaces is that metric distance or vector projections are symmetric, i.e. the similarity of two entities A and B is same as similarity between B and A. A large array of work from Amol Tversky [22,23] has investigated and found that similarity judgements lose their symmetric property when made under context-sensitive scenarios. The concepts and entities considered in these studies have ranged from country-pairs to geometric figures. ...
... In this study [23], participants were shown pairs of 21 countries and asked to rate the similarity between them on a scale of 0 to 20. The participants were divided into two groups, and each group rated similarity of 21 countries. ...
Large Language Models (LLMs) have revolutionised the capability of AI models in comprehending and generating natural language text. They are increasingly being used to empower and deploy agents in real-world scenarios, which make decisions and take actions based on their understanding of the context. Therefore researchers, policy makers and enterprises alike are working towards ensuring that the decisions made by these agents align with human values and user expectations. That being said, human values and decisions are not always straightforward to measure and are subject to different cognitive biases. There is a vast section of literature in Behavioural Science which studies biases in human judgements. In this work we report an ongoing investigation on alignment of LLMs with human judgements affected by order bias. Specifically, we focus on a famous human study which showed evidence of order effects in similarity judgements, and replicate it with various popular LLMs. We report the different settings where LLMs exhibit human-like order effect bias and discuss the implications of these findings to inform the design and development of LLM based applications.
... 'A piece is like life' means that the piece can capture essential elements of life. Gati (2004Gati ( [1978, p.81), a toy train is more similar to the real train because many features of the toy train are included in the real train. On the other hand, the real train is not as similar to the toy train because many of the features of the real train are not included in the toy train. ...
... 'A piece is like life' means that the piece can capture essential elements of life. Gati (2004Gati ( [1978, p.81), a toy train is more similar to the real train because many features of the toy train are included in the real train. On the other hand, the real train is not as similar to the toy train because many of the features of the real train are not included in the toy train. ...
... Thus, according to Tversky's (1977, p.333) focus hypothesis, the asymmetry is determined by the salience of the stimulus or prototype, so that the less salient stimuli are more similar to the more salient stimuli and not vice versa, i.e., S(A, B) > S(B, A) whenever f(B) > f(A) (Tversky 1977, p.388). The similarity of (a, b) is greater than the similarity of (b, a) whenever the characteristics of 'b' are more salient than the characteristics of 'a' or whenever 'b' is more prominent than 'a', i.e., S In Gati (2004 [1978], p.82), the hypothesis of directional asymmetry, derived from the contrast model, was tested using semantic (countries) and perceptual (figures) stimuli. The two studies employed essentially the same design. ...
The study aims to use three methods of evaluation of similarities developed in cognitive psychology in the classification of distances and groupings of diatonic chords in major and minor keys, taking into account the four tonal scales (Ionian, Aeolian, Harmonic Minor and Melodic Minor), of the twenty-four tonalities, also taking into account the four tonal scales, of the forty-eight series of Alban Berg's Violin Concerto and of some ensembles representing scales in music theory. The methods employed are: the Contrast Model, which weights the common and distinct characteristics between the objects compared (Tversky 1977); the Structural Alignment, which classifies similarities by distinguishing between two types of differences and two types of commonalities (Gentner; Markman 1995, 1997; Goldstone 1994; Markman; Gentner 1990, 1993a, 1993b, 1993c, 1996); and Transformation, which evaluates similarities through the number of operations required to transform one object into another (Chater; Hahn 1997; Hahn; Chater; Richardson 2003; Hahn; Richardson; Chater 2001).
... This article is devoted to how people complete the construction and comparison of these worlds when rendering assessments of impact. Specifically, drawing on recent work on the mental simulation (Kahneman & Tversky, 1982) and judgment of similarity (Tversky, 1977;Tversky & Gati, 1978), we trace through the steps that people follow when assessing impact. From that analysis, we then test and explore a specific implication: That people perceive greater impact when their assessments are framed in terms of mental addition as opposed to mental subtraction. ...
... It is Tversky's (1977;Tversky & Gati, 1978) work on the contrast model of similarity judgments that has documented the first judgmental tendency assumed here, that the features in the subject of the comparison receive the bulk of the attention in judgments of similarity and contrast. Consider a comparison, for example, between an object that contains many features (e.g., China) and one that contains just a few (e.g., North Korea). ...
... Recall earlier that when people consider two objects and one contains many more features than the other, people perceive more contrast between the two when focused on the many-featured object (cf. Tversky, 1977;Tversky &Gati, 1978). ...
Assessing the consequences of actions and events often requires comparing a mental simulation of the world in which the action is present to one in which the action is absent. We propose that people perceive more impact when asked to assess whether an action would increase the likelihood or degree of a potential outcome (mental addition) than when asked whether it would reduce the probability or extent of a potential consequence (mental subtraction). This judgmental asymmetry occurs because people (a) give more weight to features of the particular mental simulation (the action or its absence) serving as the subject of comparison and (b) give more weight to factors that produce as opposed to inhibit the relevant outcome. In 4 studies, Ss assessed the impact of personal actions (e.g., studying for an exam). Ss perceived more impact when the assessment was placed in an additive frame (e.g., “how many more questions will you get right if you study?”) as opposed to a subtractive one (e.g., “how many fewer will you get right if you do not study?”). This effect was not influenced by the hedonic value of the event or by whether the S had actually experienced it. Discussion centers on the relevance of these results for the undoing of scenarios and causal attribution.
... Because D is a metric, it satisfies Thus, dissimilarity should be symmetric and minimal between a point and itself. There is considerable evidence (see Rosch, 1975;Tversky, 1977;Tversky & Gati, 1978) that similarity judgments reveal predictable asymmetries. In particular, a prominent object or a prototype is less similar to a nonprominent object or a variant than vice versa. ...
... The weight attached to a mustache, for example, in judging the similarity between people, varies greatly, depending on whether the task calls for the assessment of similarity with respect to appearance, character, or a combination of the two. Although the application of the contrast model to a given context requires some assumptions about the relevant feature structure, many consequences of the model can be tested without a complete specification of the features associated with the objects under study (Gati & Tversky, 1982;Tversky & Gati, 1978). The present paper investigates feature structures associated with stimuli that vary along ordered, separable attributes and applies the contrast model to the analysis of this case. ...
... Earlier work supported the hypothesis that people tend to focus on common features in judgments of similarity more than in judgments of dissimilarity. Consequently, the relative weight of the common features is greater in similarity judgments than in dissimilarity judgments (Tversky, 1977;Tversky & Gati, 1978). Because the focus on common features makes the relation of coincidence more transparent and salient, we also expect the coincidence effect to be more pronounced in similarity than in dissimilarity judgments. ...
An alternative analysis of geometric models of proximity data, based on a feature-matching model, leads to the coincidence hypothesis that the dissimilarity between objects that differ on 2 separable dimensions is larger than predicted from their unidimensional differences on the basis of the triangle inequality and segmental additivity. A series of studies of 2-dimensional stimuli with separable attributes (including house plants, parallelograms, schematic faces, and histograms), using judgments of similarity and dissimilarity, classification, inference, and recognition errors, all support the coincidence hypothesis. The size of the effect is determined by the separability of the stimuli, the transparency of the dimensional structure, and the discriminability of the levels within each dimension. Applications of the coincidence effect to inductive inference are investigated, and its relations to selective attention and spatial density are discussed. It is concluded that the triangle inequality and segmental additivity cannot be jointly satisfied for separable attributes. The implications of this result for multidimensional scaling are discussed. (57 ref)
... Experimental evidence for directionality in human similarity judgements has been found for both perceptual and conceptual stimuli. For instance, Tversky and Gati (1978) conducted a set of studies on undergraduate students whereby they tested similarity judgements on sets of schematic faces, cities and countries. Additionally, Tversky (1977) found significant effects of directionality in Rothkopf's (1957) Morse-code data and Krantz and Tversky (1975) report evidence for directionality in similarity-judgements of rectangles. ...
... The directionality effect can be 'cancelled' by eliminating the initial inequalities, either those associated with the cities' prominence or those associated with their grammatical roles. Due to this flexibility, the model accommodates a variety of other data sets that illustrate directionality effects, including Morse-code signals, countries, faces and geometric objects (Tversky and Gati 1978). ...
... In contrast, the decomposition of objects into discrete sets of entities that either share a target feature (e.g., being turquoise) or not fails to account for the continuous structure of perception (Barsalou 2008). On the other hand, discrete representations have proven practical to model conceptual similarities and their sensitivity to contextual changes (Shepard and Arabie 1979;Tversky and Gati 1978;Gati and Tversky 1984). For instance, Tversky (1977) explains the greater similarity between Switzerland and Italy within an expanded context including Uruguay and Brazil by appealing to the greater influence of categorical background knowledge about their continental location. ...
Tenenbaum and Griffiths ( Behavioral and Brain Sciences 24(4):629–640, 2001) have proposed that their Bayesian model of generalisation unifies Shepard’s ( Science 237(4820): 1317–1323, 1987) and Tversky’s ( Psychological Review 84(4): 327–352, 1977) similarity-based explanations of two distinct patterns of generalisation behaviours by reconciling them under a single coherent task analysis. I argue that this proposal needs refinement: instead of unifying the heterogeneous notion of psychological similarity, the Bayesian approach unifies generalisation by rendering the distinct patterns of behaviours informationally relevant. I suggest that generalisation as a Bayesian inference should be seen as a complement to, instead of a replacement of, similarity-based explanations. Furthermore, I show that the unificatory powers of the Bayesian model of generalisation can contribute to the selection of one of these models of psychological similarity.
... perceptions of the average fit between rival extensions and their respective parent categories and "focal fit" is defined as the perceptions of the similarity between a focal extension and its parent category. Then, building upon research on context effects (Tversky and Gati, 1978), the paper argues for the existence of a backdrop effect, whereby the evaluation of a focal extension is conditional on context fit. ...
... In a somewhat different context, Tversky and Gati (1978) have previously examined country dyads such as Canada-U.S.A. and Italy-Switzerland. Eight of the dyads in their study consisted of American countries and eight consisted of European countries. ...
... Situations when consumers evaluate brand extensions as part of making purchase decisions have some commonalities and some differences with the set-up of Tversky's and Gati's (1978) experiment. One important dissimilarity is that with brand extensions the dyadic relationship between parent and extension categories is strictly implicit -there is no dyad per se; rather, the purchase situation itself activates consumers' consideration of the parent category. ...
Given the widespread acceptance of extensions-led brand strategies on the one hand, and a reduction in the size of brand portfolios by marketers and retailers on the other, many product categories now consist largely of extensions of brands from other categories. This paper examines consumers’ evaluation of a brand extension when its competitive context consists of other extensions. The results from two studies show that consumers’ perceptions of the fit of a focal extension with its brand’s core category and their acceptance of the focal extension are conditional on the relationship of the competing extensions with their respective parent categories. This ‘backdrop effect’ is mediated by consumers’ perceptions of fit between the competing extensions’ parent and extension categories and moderated by consumers’ differentiation versus integration mindsets.
... In contrast, previous research into the processing of comparative claims (Tversky, 1977;Tversky & Gati, 1978) suggests that the direction of comparison should influence whether the target shown in the photo facilitates or impairs fluent processing, resulting in differential effects on the perceived truth of the associated claim. Evaluating the truth of any comparative claim requires that the subject of the claim is compared to its referent (e.g., in the claim Turtles have better hearing than sealions, turtles constitute the subject and sealions the referent). ...
... Evaluating the truth of any comparative claim requires that the subject of the claim is compared to its referent (e.g., in the claim Turtles have better hearing than sealions, turtles constitute the subject and sealions the referent). As Tversky (1977;Tversky & Gati, 1978) demonstrated, the comparison process begins with an assessment of features of the subject, which are then checked against features of the referent. Hence, switching subject and referent results in qualitatively different judgments, such as the memorable observations that North Korea is judged as being more similar to China than China is to North Korea (Tversky & Gati, 1978). ...
... As Tversky (1977;Tversky & Gati, 1978) demonstrated, the comparison process begins with an assessment of features of the subject, which are then checked against features of the referent. Hence, switching subject and referent results in qualitatively different judgments, such as the memorable observations that North Korea is judged as being more similar to China than China is to North Korea (Tversky & Gati, 1978). Similarly, female teachers are judged as more empathetic than male teachers when they serve as the subject of the comparison, but as less empathetic when they serve as the referent of the comparison, reflecting that the direction of comparison influences which features come to represent the attribute "empathetic" (Wänke et al., 1995). ...
Claims are more likely to be judged true when presented with a related nonprobative photo (Newman, Garry, Bernstein, Kantner, & Lindsay, 2012). According to a processing fluency account, related photos facilitate processing and easy processing fosters acceptance of the claim. Alternatively, according to an illusion-of-evidence account, related photos may increase acceptance of the claim because they are treated as tentative supportive evidence. We disentangle these potential mechanisms by using comparative claims. In forming comparative judgments, people first assess attributes of the linguistic subject of comparison and subsequently compare them to attributes of the referent (Tversky, 1977). Hence, photos of the linguistic subject in a sentence should facilitate, but photos of the linguistic referent impair, fluent processing of this sequence. In contrast, a photo of either the subject or the referent can be perceived as tentative evidence. In two experiments (total N = 1200), photos of the subject increased acceptance of comparative claims relative to a no-photo condition (a truthiness effect), but only when the subject was otherwise difficult to visualize. Photos of the referent decreased acceptance of comparative claims relative to a no-photo condition (a falsiness effect), but only when the subject of comparison was otherwise easy to visualize. All results are consistent with a context-sensitive fluency account: increases in fluency foster, and decreases in fluency impair, acceptance of a claim as true. The results provide no support for an illusion-of-evidence account.
... In the literature, semantic similarity is applicable in different fields of computer science, particularly novel applications focused on the information retrieval task to increase the precision and recall [8][9][10][11]; to find matches between ontology concepts [12,13]; to assure or restore ontology alignment [14]; for question-answering systems [15]; for tasks for natural language processing, such as tokenization, stopwords removing, lemmatization, word sense disambiguation, lemmatization, and named entity recognition [16,17]; for recommender systems [18,19]; for data and feature mining [20][21][22]; for multimedia content search [23]; for semantic data and intelligent integration [24,25]; for ontology learning based on web-scrapping techniques, where new definitions connected to existing concepts should be acquired from document resources [26]; for text clustering [27]; for biomedical context [28][29][30][31]; and for geographic information and cognitive sciences [6,[32][33][34][35]. In a pragmatic perception, the semantic similarity helps us to comprehend human judgment, cognition, and understanding to categorize and classify various conceptualizations [36][37][38]. Thus, similitude is an essential theoretical foundation in semantic-processing tasks [39,40]. ...
... A similar principle is the Tversky's model [38]; this principle states that the similarity degree between two concepts can be calculated using a function that supports taxonomic information. Further, [76] indicated that a straightforward terminolog-ical pairing between ontological concepts addresses issues relating to integrating diverse sources of information. ...
This research uses the computing of conceptual distance to measure information content in Wikipedia categories. The proposed metric, generality, relates information content to conceptual distance by determining the ratio of the information that a concept provides to others compared to the information that it receives. The DIS-C algorithm calculates generality values for each concept, considering each relationship’s conceptual distance and distance weight. The findings of this study are compared to current methods in the field and found to be comparable to results obtained using the WordNet corpus. This method offers a new approach to measuring information content applied to any relationship or topology in conceptualization.
... That is, we see evidence of assimilation when individuals consider a target first and relate themselves to that person. For example, if a person is asked "how similar is Claire's opinion to yours?" they match features of Claire with features of themselves (Tversky, 1977;Tversky & Gati, 1978). If assimilation effects bolster similarity, and similarity drives self-knowledge activation, then it is possible that inducing assimilation effects will increase self-knowledge activation. ...
... In contrast, when individuals consider their own characteristics first, and subsequently relate a target to themselves, people are more likely to focus on differences between themselves and the other (Fabrigar & Krosnick, 1995;Karniol, 2003;Mullen et al., 1985;Mullen & Smith, 1990). For example, if a person is asked "how similar is your opinion to Claire's?" individuals draw on their elaborate and specific self-knowledge, making it difficult to match features, thereby decreasing perceived similarity (Tversky, 1977;Tversky & Gati, 1978). Thus, we expect stronger SIM effects when considering a similar target before the self. ...
Simulating other people can shift one’s self-concept, an effect known as simulation-induced malleability. How does imagining others shift the self? We propose that the activation of self-knowledge is the key factor by which simulation of others alters one’s self-concept. We test this possibility across four studies that each manipulate self-knowledge activation indirectly during simulation and measure the impact on subsequent self-ratings. Results demonstrate that increasing activation of self-knowledge during simulation is associated with increased self-concept change. People experienced greater self-concept change when simulating similar others (Studies 1 and 2). People also generalized simulation-induced changes to aspects of the self-concept that were semantically similar to the simulated content (Study 3). Finally, people who are less likely to recruit self-knowledge (i.e., older adults) during simulation were less susceptible to self-concept change (Study 4). These studies highlight self-knowledge activation as an essential component of the effects of simulation on self-rated change.
... The geometric approach has been successfully applied to a wide variety of psychological data (e.g., Carroll & Arabic, 1980;Shepard, 1974Shepard, , 1980; but its theoretical adequacyas distinct from its function as a method for data reduction'-can be questioned. In recent articles (Tversky, 1977;Tversky & Gati, 1978), we discussed some limitations of the geometric approach and explored an alternative (set-theoretical) framework in which objects are described as measurable sets of features, and similarity of objects is expressed as a linear combination, or a contrast, of the measures of their common and distinctive features. The present article applies a simplified version of the contrast model to a two-dimensional factorial structure with ordered attributes. ...
... The contrast model has been used to uncover, analyze, and explain several empirical phenomena such as the relation between similarity and dissimilarity judgments, the effect of context on similarity and classification, and the presence of asymmetric proximities (see Tversky, 1977;Tversky & Gati, 1978). The present article investigates feature structures associated with stimuli that vary along ordered attributes and applies the contrast model to the analysis of this case. ...
Geometric representations of psychological dimensions were analyzed and compared to an alternative set-theoretical approach. Judgments of similarity between forms and figures by 244 high school and undergraduate students revealed the following effects: (a) Qualitative attributes were curved relative to quantitative attributes, contrary to intradimensional subtractivity; (b) quantitative attributes augmented differences in qualitative attributes, contrary to interdimensional additivity; and (c) adding a new dimension with a fixed value increased similarity, contrary to translation invariance. The implications of these results for multidimensional representations of proximity data are discussed. (24 ref)
... Similarity is the basis of metaphor, and it also constrains the selection of particular linguistic expression to talk about something else (Kovecses, 2002). Similarity plays an important role in human perception ( Melara, 1992;Tversky & Gati, 1978) and information organization and retrieval. ...
... Namely, the salience of features may vary widely depending on implicit or explicit instructions and on the object set under consideration. Tversky & Gati (1978) proposed the salience of features is determined in part of by their diagnosticity, i.e., classificatory significance. A feature may acquire diagnostic value in a particular context if it severs as a basis for classification in that particular context, then the value may become more salient. ...
Metaphor is based on the similarity of two things.
It’s the center of creative thinking. The purpose of this study
is to understand humans’ perception about similarity. This
study uses Tversky's contrast model to discuss the similarity
of product images with three issues: the importance of
common features for similarity and dissimilarity; the
asymmetry of similarity; the degree of similarity of two
images in different context. Through three experimental
investigations with chair and product images as material,
different assumptions were tested respectively. The results
show: (1) common features are more important while
choosing similar source; (2) more salient images are
regarded as a referent and comparison of similarity is
asymmetry; (3) salience of features might be changed in
different context. The results can be applied to metaphorical
products and creative thinking to perceive common and
salient fetures between different domains, and make a
innovative connection.
... Proximity Measure: The final but crucial step in this framework constitutes the comparison of the transformed data by quantifying their shared or discriminating features, i.e., object commonalities and differences [84]. This is achieved by applying an appropriately chosen proximity measure, i.e., (dis-)similarity measure [66], which has to be consistent with the transformed data as well as with the subsequent machine learning approach. ...
... Hence, both definitions state that an object is most similar to itself as a basic principle in agreement with interpretations in cognitive science [84]. A more strict dissimilarity measure is a distance or, equivalently, a metric. ...
The encounter of large amounts of biological sequence data generated during the last decades and the algorithmic and hardware improvements have offered the possibility to apply machine learning techniques in bioinformatics. While the machine learning community is aware of the necessity to rigorously distinguish data transformation from data comparison and adopt reasonable combinations thereof, this awareness is often lacking in the field of comparative sequence analysis. With realization of the disadvantages of alignments for sequence comparison, some typical applications use more and more so-called alignment-free approaches. In light of this development, we present a conceptual framework for alignment-free sequence comparison, which highlights the delineation of: 1) the sequence data transformation comprising of adequate mathematical sequence coding and feature generation, from 2) the subsequent (dis-)similarity evaluation of the transformed data by means of problem-specific but mathematically consistent proximity measures. We consider coding to be an information-loss free data transformation in order to get an appropriate representation, whereas feature generation is inevitably information-lossy with the intention to extract just the task-relevant information. This distinction sheds light on the plethora of methods available and assists in identifying suitable methods in machine learning and data analysis to compare the sequences under these premises.
... In an earlier paper (van der Leeuw, 2019), I have argued that the core dynamic underpinning perception and decision-making is categorization, in which first open, exploratory (groups) and subsequently closed, entities (classes) are created that are exploited to grapple with the unknown. In technical terms, based on theoretical and experimental work of Tversky and Gati (1978), the evolution of pattern recognition is here seen as a shift from extrinsically circumscribed, polythetic open categories to intrinsically defined, monothetic, closed categories. That perspective is chosen because the distinction between open and closed categories has been widely discussed in the cognitive sciences (e.g., Cohen & Lefebvre, 2018), but also in cultural anthropology, sociology (e.g., El Guindi, 1972Guindi, , 1973Selby & El Guindi, 1976) and other disciplines (e.g., Davis-Floyd, 2018). ...
The paper considers narratives as dynamic memory banks and shifts understanding from emphasizing the origins of the present to the emergence of the present. In the construction of reality, imagined futures articulate with knowledge obtained in the past.
In another inversion, rather than explain change and consider stability as the norm, it focuses on change as the norm and investigates the creation of stability to explain, for example, why our societies are so slow in acting on climate change.
The creation of meaning is the result of an interaction between thinking and experience, like the interaction between a map and the territory it represents. It reduces the complexity of the territory to the simplicity of the map, shaping simultaneously the cognitive map and the territory it represents. Such cognitive structures evolve into dense networks of cognitive dimensions.
Tipping points emerge as a particular cognitive structure is no longer enabling a society to deal with its changing environment because it does not fully trace the logical and functional nature of the relationship between the two. To facilitate that, we need to understanding noise as signals for which no interpretative conceptual and cognitive structure has yet been identified.
... For example, when people form associations between categories and their attributes, more attention and weight are given to attributes that heighten between-category differences rather than between-category similarities (Krueger & Clement, 1994;Krueger et al., 1989). Distinct attributes tend to define a category (Tversky & Gati, 2004). ...
We investigate self-appraisals over time using a cognitive–ecological approach. We assume that ecologically, negative person attributes are more diverse than positive ones, while positive person attributes are more frequent than negative ones. We combine these ecological properties with the cognitive process of similarity- and differences-based social comparisons to predict temporal self-appraisals. The resulting cognitive–ecological model predicts that people should evaluate similarities with themselves over time positively, whereas differences would be evaluated more negatively. However, because positive attributes are reinforced over time relative to negative attributes, we predicted an asymmetry to emerge such that distinct attributes of the past self (past differences) would be most negative. Six experiments (total N = 1,796) and an integrative data analysis confirmed the cognitive–ecological model’s predictions for temporal self-appraisals. However, we found no evidence for motivated self-perception across time. We discuss the implications of these findings for temporal self-appraisal theory and other aspects of self and identity.
... However, studies have shown that evaluations of stimuli, including outcomes, are influenced by the context in which they are considered. A great deal of research has shown that context influences the aspects of stimuli as well as concepts that are brought to mind at any given time and thus may influence evaluations of those stimuli (e.g., Bransford, 1979;Tversky & Gati, 1978;Tversky & Kahneman, 1974). ...
The influence of positive affect on expectancy motivation was investigated in 2 studies. The results of Study 1 indicated that positive affect improved people's performance and affected their perceptions of expectancy and valence. In Study 1, in which outcomes depended on chance, positive affect did not influence people's perceptions of instrumentality. In Study 2, in which the link between performance and outcomes was specified, positive affect influenced all 3 components of expectancy motivation. Together, the results of Studies 1 and 2 indicated that positive affect interacts with task conditions in influencing motivation and that its influence on motivation occurs not through general effects, such as response bias or general activation, but rather through its influence on the cognitive processes involved in motivation.
... Indeed, it has been suggested that any newly encountered person is likely to be compared with preexisting notions in memory that are similar to the encountered person, based on a simple similarity estimate between the person and the various representations, perhaps yielding a match in relation to a particular representation (Tversky, 1977;Tversky & Gati, 1978; see also Cantor & Mischel, 1979;Nosofsky, 1986;Rosch, 1978). Such a match may thus lead to complex interpretive activity about the new person by triggering the application of the preexisting representation to the person-presumably with little cognitive effort (e.g., Bargh, 1989;Bargh & Tota, 1988;Andersen, Spielman, & Bargh, 1992;Smith & Lerner, 1986) or conscious awareness (e.g., Bargh & Pietromonaco, 1982;; see also Bornstein & Pittman, 1992;Kihlstrom, 1987Kihlstrom, , 1990 and based on spreading activation through elaborate networks (Cantor & Kihlstrom, 1987; see also Collins & Loftus, 1975). ...
Research has shown that the activation and application of a significant-other representation to a new person, or transference, occurs in everyday social perception (S. M. Andersen & A. Baum, 1994; S. M. Andersen & S. W. Cole, 1990). Using a combined idiographic and nomothetic experimental paradigm, two studies examined the role of chronic accessibility of significant-other representations in transference. After learning about 4 fictional people, 1 of whom resembled a significant other, participants' recognition memory was assessed. Both studies showed greater false-positive memory in the significant-other condition, relative to control, even in the absence of priming. Study 2 showed that although the effect was greater when the significant-other representation was concretely applicable to the target information, it occurred even when no such applicability was present. Results implicate the chronic accessibility of significant-other representations in transference.
... Interestingly, this appears to be a case in which the prototypical member of the category and the most common member of the category are not the same (Rosch & Mervis, 1975;Smith & Medin, 1981;Tversky & Gati, 1978). If subjects' judgments of typicality had followed the male-female ratio in the population, these judgments should have favored women over men. ...
The hypothesis that explanations for differences between prototypical and nonprototypical members of categories would focus more on attributes of the latter than on those of the former was examined. Explanations for alleged gender differences in the behavior of voters, elementary school teachers, and college professors were elicited. As predicted, explanations for gender differences within the 3 categories emphasized the qualities of the “deviant” member. Ss' explanations of alleged gender gaps in the behavior of voters and college professors focused more on qualities of women than on qualities of men. In contrast, Ss' explanations of an alleged gender gap in the behavior of elementary school teachers focused more on qualities of men than on qualities of women. The results are interpreted in terms of Kahneman and Miller's (1986) norm theory.
... Experimental research stimulated by this conception has uncovered several factors that control the relative weight of these components. Tversky and Gati (1978) showed that W is larger in judgments of similarity than in judgments of dissimilarity, presumably because the former focus on shared features, whereas the latter focus on distinctive features. Gati and Tversky (1984) showed that the value of W associated with a particular component, added to one or two subjects, is inversely related to the baseline similarity of the objects. ...
We investigated possible explanations of the finding that the relative weight (W) of common components in similarity judgments is higher for verbal than for pictorial stimuli. A serial presentation of stimulus components had no effect on verbal stimuli; it increased the impact of both common and distinctive components of pictorial stimuli but did not affect their relative weight. On the other hand, W was increased by manipulations that reduced the cohesiveness of composite pictures, such as separating, scrambling, and mixing their components. Furthermore, W was decreased by manipulations that enhanced the cohesiveness of composite verbal stimuli by imposing structure on their components. Verbal and pictorial representations of the same stimuli yielded no systematic differences in W.
... This effect is frequently described in the literature on similarity and categorization studies. Aside from studies confirming the effect by Krantz and Tversky (1975) and Tversky and Gati (1978), Krumhansl (1978) identifies evidence for directionality on the similarities of focal colors. Rips (1975) shows that subjects are more likely to attribute a disease to an atypical species if the typical species carries the disease, as opposed to attributing the disease to a typical species if the infected species is atypical. ...
The feature-matching approach pioneered by Amos Tversky remains a groundwork for psychological models of similarity and categorization but is rarely explicitly justified considering recent advances in thinking about cognition. While psychologists often view similarity as an unproblematic foundational concept that explains generalization and conceptual thought, long-standing philosophical problems challenging this assumption suggest that similarity derives from processes of higher-level cognition, including inference and conceptual thought. This paper addresses three specific challenges to Tversky’s approach: (i) the feature-selection problem, (ii) the problem of cognitive implausibility, and (iii) the problem of unprincipled tweaking. It subsequently supports key insights from Tversky’s account based on recent developments in Bayesian modeling of cognition. A novel computational view of similarity as inference is proposed that addresses each challenge by considering the contrast class as constitutive of similarity and selecting for highly informative features. In so doing, this view illustrates the ongoing promise of the feature-matching approach in explaining perception, generalization and conceptual thought by grounding them in principles of probabilistic inference.
... We also assume that this association is asymmetric and exists to a lesser extent from target to source (cf. Tversky and Gati (2004)). ...
Transformer-based models are now predominant in NLP. They outperform approaches based on static models in many respects. This success has in turn prompted research that reveals a number of biases in the language models generated by transformers. In this paper we utilize this research on biases to investigate to what extent transformer-based language models allow for extracting knowledge about object relations (X occurs in Y; X consists of Z; action A involves using X). To this end, we compare contextualized models with their static counterparts. We make this comparison dependent on the application of a number of similarity measures and classifiers. Our results are threefold: Firstly, we show that the models combined with the different similarity measures differ greatly in terms of the amount of knowledge they allow for extracting. Secondly, our results suggest that similarity measures perform much worse than classifier-based approaches. Thirdly, we show that, surprisingly, static models perform almost as well as contextualized models -- in some cases even better.
... We also assume that this association is asymmetric and exists to a lesser extent from target to source (cf. Tversky and Gati (2004)). ...
Transformer-based models are now predominant in NLP.
They outperform approaches based on static models in many respects.
This success has in turn prompted research that reveals a number of biases in the language models generated by transformers.
In this paper we utilize this research on biases to investigate to what extent transformer-based language models allow for extracting knowledge about object relations (X occurs in Y; X consists of Z; action A involves using X).
To this end, we compare contextualized models with their static counterparts.
We make this comparison dependent on the application of a number of similarity measures and classifiers.
Our results are threefold:
Firstly, we show that the models combined with the different similarity measures differ greatly in terms of the amount of knowledge they allow for extracting.
Secondly, our results suggest that similarity measures perform much worse than classifier-based approaches.
Thirdly, we show that, surprisingly, static models perform almost as well as contextualized models -- in some cases even better.
... A particularly important problem is the proper selection of the appropriate membership function (in the case of the product of fuzzy sets). This takes place in the construction of algorithms for supporting diagnostic decisions using similarity models and pattern recognition [9,11,28]. The following considerations will be limited to examining the operations of the product of fuzzy sets and the consequences of adopting various formulas of belonging to the fuzzy set. ...
The paper presents a several new definitions of concepts regarding the properties of fuzzy sets in the aspect of their use in decision support processes. These are concepts such as the image and counter – image of the fuzzy set, the proper fuzzy set, the fuzzy support and the ranking of fuzzy set. These concepts can be important in construction decision support algorithms. Particularly a lot of space was devoted to the study of the properties of membership function of the fuzzy set as a result of operations on fuzzy sets. Two additional postulates were formulated that should be fulfilled by the membership function product of fuzzy sets in decision making.
... For example, based on the diagnosticity effect (Evers & Lakens, 2014;Tversky, 1977), we would expect that the same target outcomes could be included in the same mental account or different mental accounts depending on features of additional outcomes provided, even when those outcomes are clearly unrelated to the target outcomes. To illustrate, Tversky and Gati (1978) found that participants perceive England and France to be much more similar when surrounded by Australia and New Zealand, as compared to when they were surrounded by Australia and Italy. We would expect similar effects to occur for the posting of outcomes into mental accounts. ...
The theory of mental accounting is often used to understand how people evaluate multiple outcomes or events. However, a model predicting which outcomes are associated with the same mental account and evaluated jointly, versus different accounts and evaluated separately, has remained elusive. We develop a framework that incorporates an online, bottom-up process of similarity and categorization into mental accounting operations. In this categorization-based model of mental accounting, outcomes that overlap on salient attributes are automatically categorized and assigned to the same mental account while outcomes that do not overlap on salient attributes are assigned to different accounts. We use this model to derive the hedonic accounting hypothesis, which generates testable behavioral predictions on people's preferences over the timing of outcomes given similarity-based constraints on mental accounting operations. Six studies provide support for the predictions: People prefer to experience similar losses close together in time and spread dissimilar losses apart; the reverse is true for gains, with a preference for dissimilar gains close together in time and similar gains spread apart across time. Importantly, our model is able to rationalize prior evidence that has found only limited support for the predictions of mental accounting and hedonic editing. Once the psychological process of similarity and categorization is explicitly incorporated into a formal model of mental accounting, its predictions are supported by the data. (PsycInfo Database Record (c) 2021 APA, all rights reserved).
... A way to harness similarity's multifaceted nature might be by assuming that its flexibility derives from contextual factors, so that, if such contextual factors can be understood, then we would likewise understand how similarity can appear to vary 'arbitrarily' (Goodman, 1972). Tversky (1977;Tversky & Gati, 1978) presented a highly cited example of the dependency of similarity on context. Consider a choice task of selecting the country most similar to Austria, amongst different choice sets. ...
Tversky's (1977) famous demonstration of a diagnosticity effect indicates that the similarity between the same two stimuli depends on the presence of contextual stimuli. In a forced choice task, the similarity between a target and a choice, appears to depend on the other choices. Specifically, introducing a distractor grouped with one of the options would reduce preference for the grouped option. However, the diagnosticity effect has been difficult to replicate, casting doubt on its robustness and our understanding of contextual effects in similarity generally. We propose that the apparent brittleness of the diagnosticity effect is because it is in competition with an opposite attraction effect. Even though in both the similarity and decision-making literatures there are indications for such a competition, we provide the first direct experimental demonstration of how an attraction effect can give way to a diagnosticity one, as a distractor option is manipulated. (PsycInfo Database Record (c) 2021 APA, all rights reserved).
... Such findings imply that the perceptual space differs from the physical space and that due to its asymmetric nature its properties cannot be captured by Euclidean geometry (e.g., distances in the vowel formant space). Asymmetry in perception has been investigated for various types of stimuli including color, line orientation, numbers (Rosch, 1975), geometric figures (Tversky and Gati, 1978) as well as vowels (Polka andBohn, 2003, 2011;Eulitz and Lahiri, 2004;De Jonge and Boersma, 2015), and consonants (Schluter et al., 2016;Cummings et al., 2017;Højlund et al., 2019). Vowel perception asymmetry has been studied by means of reaction time or accuracy in discrimination tasks, where a reversed order of stimuli led to the significant difference in the measured parameters. ...
Neural discrimination of auditory contrasts is usually studied via the mismatch negativity (MMN) component of the event-related potentials (ERPs). In the processing of speech contrasts, the magnitude of MMN is determined by both the acoustic as well as the phonological distance between stimuli. Also, the MMN can be modulated by the order in which the stimuli are presented, thus indexing perceptual asymmetries in speech sound processing. Here we assessed the MMN elicited by two types of phonological contrasts, namely vowel quality and vowel length, assuming that both will elicit a comparably strong MMN as both are phonemic in the listeners’ native language (Czech) and perceptually salient. Furthermore, we tested whether these phonemic contrasts are processed asymmetrically, and whether the asymmetries are acoustically or linguistically conditioned. The MMN elicited by the spectral change between /a/ and /ε/ was comparable to the MMN elicited by the durational change between /ε/ and /ε:/, suggesting that both types of contrasts are perceptually important for Czech listeners. The spectral change in vowels yielded an asymmetrical pattern manifested by a larger MMN response to the change from /ε/ to /a/ than from /a/ to /ε/. The lack of such an asymmetry in the MMN to the same spectral change in comparable non-speech stimuli spoke against an acoustically-based explanation, indicating that it may instead have been the phonological properties of the vowels that triggered the asymmetry. The potential phonological origins of the asymmetry are discussed within the featurally underspecified lexicon (FUL) framework, and conclusions are drawn about the perceptual relevance of the place and height features for the Czech /ε/-/a/ contrast.
... During LI acquisition, the child establishes perceptual constancy for sound categories by identifying a wide range of acoustically different sounds as categorically equivalent (Kuhl & Miller, 1982). The young child is able to attend lo just those sensory attributes or features of a sound important to its category identity while learning to ignore other attri-butes that arc irrelevant, less important, or which provide conflicting information concerning categorical identity (Kuhl, 1980;Tversky & Gati, 1978). ...
This chapter had three goals: (1) to identify and develop theoretical issues of importance regarding the production and perception of sounds found in a foreign language; (2) to review research dealing with these issues up to 1984, when this chapter was completed; and (3) and to identify topics of special interest for future research.
... It may also be worth following the lead of Tversky and Gati's (1978) work on similarity relations. These authors pointed out that to say that a toy train is like a real train is much more acceptable than to say that a real train is like a toy train, because most features of the toy train are contained within the real train, but the toy train only has a small number of features exhibited by the real train. ...
This article proposes that predicate order and coherence relations are the two major determining factors in copredication licensing, resolving a long-standing puzzle over the criteria for constructing acceptable copredications. The effects of predicate ordering are claimed to be anchored around semantic complexity, such that copredications with semantically Simple–Complex predicate orderings are more acceptable than the reverse. This motivates a parsing bias, termed Incremental Semantic Complexity. Particular ways of implementing this parsing bias are discussed. The effect of predicate coherence is claimed to be anchored around a sense of causality and featural commonality. Lastly, a hierarchy of possible copredications is outlined (the Copredication Hierarchy), helping to delimit the modelling of copredications to a greater extent than has previously been possible.
... We note that the theory on which this work is based makes several assumptions, such as the notions that the distance data meets the assumptions of positivity and symmetry (cf. Tversky & Gati, 1978) and that the use of vector addition to predict the effect of messages works reasonably well. ...
Using Galileo theory and method of multidimensional scaling (MDS), we compared the psychological distances between concepts related to two pandemic viruses, Zika and COVID-19. Surveys (Zika, N = 410; COVID-19, N = 291) were used to investigate the role of media use and interpersonal communication on the relationship between 10 concepts in multidimensional spaces. We asked these four research questions: Do the two spaces represent the two pandemics similarly? What is the relationship of me and of people to each pandemic? What is the effect of virus-related media use and interpersonal talk on the pandemic space? What are optimal messages for moving me closer to Zika and to COVID-19? Media use influenced the distances for both pandemics: With greater media use, the concepts were closer in the Zika space and further apart in the COVID-19 space. Interpersonal communication was associated with few differences in the spaces. Based on the psychological distances between concepts, optimal messages were identified: For Zika, a message with two concepts, people and women, is predicted to be most effective to move Zika to the concept me, whereas for COVID-19, a message with people is predicted to be most effective to move COVID-19 to me.
... To understand how, consider the statements "Zebras are similar to horses" and "Horses are similar to zebras." Although both statements express similarity between horses and zebras, adults who are asked which statement "sounds better" will overwhelmingly choose "Zebras are similar to horses" (Bowdle & Medin, 2001;Chestnut & Markman, 2016;Gleitman et al., 1996;Roese et al., 1998;Rosch, 1975;Shen, 1989;Tversky, 1977;Tversky & Gati, 1978). These asymmetrical framing preferences have been found not only for similarity, but also for a number of other symmetrical predicates such as near, meet, and marry. ...
How do children learn gender stereotypes? Although people commonly use statements like "Girls are as good as boys at math" to express gender equality, such subject-complement statements subtly perpetuate the stereotype that boys are naturally more skilled. The syntax of such statements frames the item in the complement position (here, boys) as the standard for comparison or reference point. Thus, when the statement concerns ability, listeners infer that this item is naturally more skilled than the item in the subject position (here, girls). In 2 experiments, we ask whether subject-complement statements could not only reinforce preexisting gender stereotypes, but also teach them. The participants were 288 adults (51% women, 49% men) and 337 children ages 7 to 11 (50% girls, 50% boys; of the 62% who reported race, 44% self-declared as White; from primarily middle-class to upper middle-class families). Participants were provided with subject-complement statements about either novel abilities (e.g., "Girls are as good as boys at trewting") or nonstereotyped activities (e.g., "Boys are as good as girls at snapping"). Both adults and children inferred that the gender in the complement position was naturally more skilled than the gender in the subject position. Using subject-complement statements to express gender equality (e.g., "Girls are as good as boys at math") could thus backfire and teach children that boys have more natural ability. (PsycInfo Database Record (c) 2020 APA, all rights reserved).
... These assumptions are controversial. Tversky and Gati (1978) demonstrated that individuals assessing the similarity of concepts could violate these assumptions. They developed a feature-theoretical approach to assess the similarity or difference between concepts in contrast to the geometric approach used here. ...
Galileo multidimensional scaling (MDS) provides a way of mapping conceptual configurations of attitudes, beliefs, behaviors, and emotions in individuals, groups, and cultures, as well as a way to compare such configurations to predict and observe their change over time. Given this analytical power, Galileo MDS is well suited for studying media psychology regarding how media and the changing media environment relate to human cognition, emotion, and behavior. This entry explains the assumptions related to Galileo MDS as well as the steps needed to use this model effectively. Finally, it provides examples that elucidate the analytical potential of this approach.
... Moreover, the truthiness and falsiness effects should depend on the changes in the ease of processing produced by the photos. Research on comparative judgments shows that a comparison begins with an assessment of features of the subject, which are then checked against features of the referent (e.g., Tversky, 1977;Tversky & Gati, 1978). Put simply, the subject is the focus of the claim and people tend to process the subject before the referent. ...
This open-access book examines the phenomenon of fake news by bringing together leading experts from different fields within psychology and related areas, and explores what has become a prominent feature of public discourse since the first Brexit referendum and the 2016 US election campaign. Thanks to funding from the Swiss National Science Foundation, all chapters can be downloaded free of charge at the publisher's website:
https://www.taylorfrancis.com/books/e/9780429295379
There is also an Amazon Kindle edition that's free of charge:
https://www.amazon.com/Psychology-Fake-News-Correcting-Misinformation-ebook-dp-B08FF54H53/dp/B08FF54H53/ref=mt_other?_encoding=UTF8&me=&qid=
... Sonesson believes that similarity as a semiotic concept, i.e. our common sense notion and perception of it, is also to be regarded within this Lifeworld. To substantiate his view Sonesson refers to experiments in cognitive psychology (Rosch 1975, Rosch, Simpson and Miller 1976, Tversky 1977, Tversky and Gati 1978 which demonstrate the asymmetry of similarity by means of prototypicality effects. In one popular experiment carried out by Tversky (1977) subjects preferred statements such as 'North Korea is similar to Red China' to the opposite 'Red China is similar to North Korea' . ...
Iconicity has become a popular notion in contemporary linguistic research. This book is the first to present a synthesis of the vast amount of scholarship on linguistic iconicity which has been produced in the previous decades, ranging from iconicity in phonology and morpho-syntax to the role of iconicity in language change. An extensive analysis is provided of some basic but nonetheless fundamental questions relating to iconicity in language, including: what is a linguistic sign and how are linguistic signs different from signs in general? What is an iconic sign and how may iconicity be involved in language? How does iconicity pertain to the relation between language and cognition? This book offers a new and comprehensive theoretical framework for iconicity in language. It is argued that the linguistic sign is fundamentally arbitrary, but that iconicity may be involved on a secondary level, adding extra meaning to an utterance.
... Traditional MDS methods consider that distances among objects are symmetric, although this consideration is not always satisfied. For instance, [137] and [138] explained the characters of similarity among objects studied with psychological scale and concluded that cognitive similarity is mostly asymmetric. The motivation behind these asymmetric MDS methods is to remove the shortcomings of traditional MDS methods, i.e., in case where similarity or dissimilarity matrices are asymmetric in nature as they are based on the supposition that similarity or dissimilarity matrices can be associated with inter-point distances in a given metric space [126], [129], [139]. ...
Current and future wireless applications strongly rely on precise real-time localization. A number of applications, such as smart cities, Internet of Things (IoT), medical services, automotive industry, underwater exploration, public safety, and military systems require reliable and accurate localization techniques. Generally, the most popular localization/positioning system is the global positioning system (GPS). GPS works well for outdoor environments but fails in indoor and harsh environments. Therefore, a number of other wireless local localization techniques are developed based on terrestrial wireless networks, wireless sensor networks (WSNs), and wireless local area networks (WLANs). Also, there exist localization techniques which fuse two or more technologies to find out the location of the user, also called signal of opportunity-based localization. Most of the localization techniques require ranging measurements, such as time of arrival (ToA), time difference of arrival (TDoA), direction of arrival (DoA), and received signal strength (RSS). There are also range-free localization techniques which consider the proximity information and do not require the actual ranging measurements. Dimensionality reduction techniques are famous among the range free localization schemes. Multidimensional scaling (MDS) is one of the dimensionality reduction technique which has been used extensively in the recent past for wireless networks localization. In this paper, a comprehensive survey is presented for MDS and MDS-based localization techniques in WSNs, IoT, cognitive radio networks, and 5G networks.
The roles of function, reminding, and exemplar variability in categorization of a physically dissimilar object were studied with 3-month-old infants trained to move a crib mobile by kicking. Performance on a transfer test with a motionless novel object provided evidence of categorization. In Experiments 1 and 2, infants, like adults, initially categorized novel objects on the basis of physical appearance, but only if trained with multiple exemplars, after delays of 1 and 7 days. In Experiment 3, prior knowledge of an object's functional properties overrode physical dissimilarity as the basis for categorization and enabled reminding of the classification response 2 weeks later. In Experiment 4, postevent contingency information overrode physical and functional properties as the basis for categorization. These findings indicate that expectations and goals influence infants' category decisions and raise the possibility that infants of 3 months respond by analogy.
In the value chain of cultural and creative industry, where metaphor is a common tool, design of culture-inspired products is one of the most important parts. This study proposes a new approach to metaphorical design method for the cultural and creative products. Integrating the notions of contrast model, salience imbalance model and structure-mapping theory, this method uses features as the primitive of metaphor, relations as the dynamic links between the features, and similarity judgment as the framework. It provides practical feature-mapping charts and feature-context charts for designers’ similarity judgment and feature retrieval. To verify the effectivity of the proposed model, three design projects with LED lighting as metaphor target, and sun-shooting myth, Tanglulu dessert and traditional pinball game as metaphor sources, are implemented. Results reveal that the behavior-level relational similarity between the source and target is positive to the symbolic meaning of the designs, whereas the action-level relational similarity is supportive to the functional innovation. The relational similarity is useful for de-composing and re-combining the elements about users, products and cultural context to transform into a creative and metaphorical solution. The method proposed is helpful for designers to restructure the abstract concepts and concrete objects from culture domain in order to design and develop products with creative form and function, as well as cultural significance.
Are species worth saving? Can they be resurrected by technology? What is the use of species in biomedicine? These questions all depend on a clear definition of the concept of 'species', yet biologists have long struggled to define this term. In this accessible book, John S. Wilkins provides an introduction to the concept of 'species' in biology, philosophy, ethics, policymaking and conservation. Using clear language and easy-to-understand examples throughout, the book provides a history of species and why we use them. It encourages readers to appreciate the philosophical depth of the concept as well as its connections to logic and science. For any interested reader, this short text highlights the complexities of a single idea in biology, the problems with the concept of 'species' and the benefits of it in helping us to answer the bigger questions and understand our living world.
Semantic similarity is a fundamental concept in computational linguistics. The models used for the representation of text have a major role in similarity computation. The text with multilingual and multimodal components shows the need for computing similarity based on different characteristics of text. This chapter studies various aspects of semantic similarity of linguistic units, cross-level similarity, semantic models, and similarity measures. One of the main motivations of this chapter is to analyze semantic similarity models such as geometric models, feature-based models, graph-based models, vector space models, and formal concept analysis models. In addition, a composite summary score based on words and hashtags is applied for the tweet summarization task which is effective when compared with other measures.
Similarity is a fundamental concept in artificial intelligence and cognitive sciences. Despite all the efforts made to study similarity, defining and measuring it between concepts or images remains challenging. Fortunately, measuring similarity is comparable to answering “why”/“in which respects” two stimuli are similar. While most related works done in computer sciences try to measure the similarity, we propose to analyze it from a different angle and retrospectively find such respects. In this paper, we provide a pipeline allowing us to find, for a given dataset of image pairs, what are the different concepts generally compared and why. As an indirect evaluation, we propose generating an automatic explanation of clusters of image pairs found using a model pre-trained on texts/images. An experimental study highlights encouraging results toward a better comprehension of visual similarity.
I use standard setters’ description of comparability to conceptualize two forms of incomparability in financial reporting. Using a laboratory experiment, I construct states of both comparability and incomparability and investigate investors’ capital allocation decisions under each. Benchmarking against states of comparable reporting, I find that incomparable reporting making different economics look alike results in lower-quality investment decisions, relative to incomparable reporting making similar economics look different. My evidence suggests the difference in investment quality is due to the reduced ability of supplemental disclosures to compensate for one type of incomparability relative to the other type. The study uses a setting holding other reporting characteristics, such as faithful representation, constant, and results are robust to alternative configurations of incomparability. Collectively, my results suggest that the type of incomparability has differential effects on the quality of investment decisions. My findings should inform standard setters aiming to help users allocate capital.
JEL Classifications: G11; M41.
The species of classical taxonomy are examined with a view to their future role in integrative taxonomy. Taxonomic species are presented as the products of a cyclic workflow between taxonomists and biologists in general, and as the essential means to express the results of evolutionary biological research in a cognitive form which can be widely understood outside the systematics research community. In the first part, the procedures underlying the formation and structure of classical species taxon concepts are analysed and discussed, and this involves some passing reference to mental concepts as understood by cognitive psychologists. The second part considers the need for methodological advances in classical taxonomy in the form of computational modelling. It is argued that in order to accomplish this, species taxon concepts will need to be expressed as computable matrices in parallel to their conventional form, expanding their role in integrative taxonomy, facilitating the feedback from evolutionary biological research and potentially accelerating the update and modification of their delimitation as knowledge increases. The third part treats another, more immediate methodological issue: some kinds of data already produced by taxonomic revisions could be provided as standard online outputs but are not yet part of the canonical published format. The final part consists of a discussion of the gradually emerging global online framework of taxonomic species and its importance as a general reference system. A glossary of terms is provided.
We propose and axiomatize the categorical thinking model (CTM) in which the framing of the decision problem affects how agents categorize alternatives, that in turn affects their evaluation of it. Prominent models of salience, status quo bias, loss-aversion, inequality aversion, and present bias all fit under the umbrella of CTM. This suggests categorization is an underlying mechanism of key departures from the neoclassical model of choice. We specialize CTM to provide a behavioral foundation for the salient thinking model of Bordalo et al. [2013] that highlights its strong predictions and distinctions from other models.
This article aims to clarify fundamental aspects of the process of assigning fuzzy scores to conditions based on family resemblance (FR) structures by considering prototype and set theories. Prototype theory and set theory consider FR structures from two different angles. Specifically, set theory links the conceptualization of FR to the idea of sufficient and INUS (Insufficient but Necessary part of a condition, which is itself Unnecessary but Sufficient for the result) sets. In contrast, concept membership in prototype theory is strictly linked to the notion of similarity (or resemblance) in relation to the prototype, which is the anchor of the ideational content of the concept. After an introductive section where I elucidate set-theoretic and prototypical aspects of concept formation, I individuate the axiomatic properties that identify the principles of transforming FR structures into fuzzy sets. Finally, I propose an algorithm based on the power mean that is able to operationalize FR structures considering both set-theoretic and prototype theory perspectives.
We investigate how predicates expressing symmetry, asymmetry and non-symmetry are encoded in a newly emerging sign language, Central Taurus Sign Language (CTSL). We find that predicates involving symmetry (i.e., reciprocal and symmetrical actions) differ from those involving asymmetry (i.e., transitive) in their use of the morphological devices investigated here: body segmentation, mirror-image articulators and double perspective. Symmetrical predicates also differ from non-symmetrical ones (i.e., intransitive) in their use of mirror-image configuration. Furthermore, reciprocal actions are temporally sequenced within a linear structure, whereas symmetrical actions are not. Thus, our data reveal that CTSL expresses each type of action with a particular combination of linguistic devices to encode symmetry, asymmetry, and non-symmetry. Furthermore, differences in the use of these devices across age cohorts of CTSL suggest that some have become more conventionalized over time. The same semantic distinctions have been observed – though with different realization – in another emerging sign language, Nicaraguan Sign Language (NSL). This converging suggests that natural human language learning capacities include an expectation to distinguish symmetry, asymmetry and non-symmetry.
This monograph offers a critical introduction to current theories of how scientific models represent their target systems. Representation is important because it allows scientists to study a model to discover features of reality. The authors provide a map of the conceptual landscape surrounding the issue of scientific representation, arguing that it consists of multiple intertwined problems. They provide an encyclopaedic overview of existing attempts to answer these questions, and they assess their strengths and weaknesses. The book also presents a comprehensive statement of their alternative proposal, the DEKI account of representation, which they have developed over the last few years. They show how the account works in the case of material as well as non-material models; how it accommodates the use of mathematics in scientific modelling; and how it sheds light on the relation between representation in science and art. The issue of representation has generated a sizeable literature, which has been growing fast in particular over the last decade. This makes it hard for novices to get a handle on the topic because so far there is no book-length introduction that would guide them through the discussion. Likewise, researchers may require a comprehensive review that they can refer to for critical evaluations. This book meets the needs of both groups.
Questions the metric and dimensional assumptions that underlie the geometric representation of similarity on both theoretical and empirical grounds. A new set-theoretical approach to similarity is developed in which objects are represented as collections of features and similarity is described as a feature-matching process. Specifically, a set of qualitative assumptions is shown to imply the contrast model, which expresses the similarity between objects as a linear combination of the measures of their common and distinctive features. Several predictions of the contrast model are tested in studies of similarity with both semantic and perceptual stimuli. The model is used to uncover, analyze, and explain a variety of empirical phenomena such as the role of common and distinctive features, the relations between judgments of similarity and difference, the presence of asymmetric similarities, and the effects of context on judgments of similarity. The contrast model generalizes standard representations of similarity data in terms of clusters and trees. It is also used to analyze the relations of prototypicality and family resemblance. (39 ref) (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Three experiments were performed in an investigation of how differences in size and inclination combine to determine the over-all similarity between otherwise identical visual stimuli. Similarity was defined both in terms of direct subjective judgments of over-all resemblance and in terms of the frequencies with which the stimuli were actually confused during identification learning. The results were incompatible with the static, Euclidean metric assumed by multidimensional scaling models. Apparently, when the stimuli vary along perceptually distinct dimensions, the psychological metric changes as subjects shift their attention more to one dimension or the other. The interstimulus similarities for any one state of attention, however, appear to conform with a Minkowski metric somewhere between the Euclidean and “city-block” varieties.
Conclusion
After struggling with the problem of representing structure in similarity data for over 20 years, I find that a number of challenging problems still remain to be overcome—even in the simplest case of the analysis of a single symmetric matrix of similarity estimates. At the same time, I am more optimistic than ever that efforts directed toward surmounting the remaining difficulties will reap both methodological and substantive benefits. The methodological benefits that I forsee include both an improved efficiency and a deeper understanding of “discovery” methods of data analysis. And the substantive benefits should follow, through the greater leverage that such methods will provide for the study of complex empirical phenomena—perhaps particularly those characteristic of the human mind.
Germany Italy England Spain -12 .1 30
- W France
France
W. Germany
Italy
England
Spain
-12.1
30.3
74.6
On the relation between similarity and dissimilari timates
- J Hosman
- T Kuennapas
Hosman, J., and Kuennapas, T. On the relation between similarity and dissimilari
timates. Report No. 354, Psychological Laboratories, The University of Stock
1972.
Psychology of judgment and choice Rosch, E. Cognitive reference points
- F Restle
Restle, F. Psychology of judgment and choice. New York : Wiley, 1961.
Rosch, E. Cognitive reference points. Cognitive Psychology, 1975, 7, 532-547.