Across three experiments featuring naturalistic concepts (psychology concepts) and naïve learners, we extend previous research showing an effect of the sequence of study on learning outcomes, by demonstrating that the sequence of examples during study changes the representation the learner creates of the study materials. We compared participants’ performance in test tasks requiring different representations and evaluated which sequence yields better learning in which type of tests. We found that interleaved study, in which examples from different concepts are mixed, leads to the creation of relatively interrelated concepts that are represented by contrast to each other and based on discriminating properties. Conversely, blocked study, in which several examples of the same concept are presented together, leads to the creation of relatively isolated concepts that are represented in terms of their central and characteristic properties. These results argue for the integrated investigation of the benefits of different sequences of study as depending on the characteristics of the study and testing situation.
Avery, J. E., Goldstone, R. L., & Jones, M. N. (2020). Reconstructing Maps from Text. Proceedings of the 42nd Annual Conference of the Cognitive Science Society. (pp. 557-563). Toronto, CA. Cognitive Science Society.
Previous research has demonstrated that Distributional Semantic Models (DSMs) are capable of reconstructing maps from news corpora (Louwerse & Zwaan, 2009) and novels (Louwerse & Benesh, 2012). The capacity for reproducing maps is surprising since DSMs notoriously lack perceptual grounding (De Vega et al., 2012). In this paper we investigate the statistical sources required in language to infer maps, and resulting constraints placed on mechanisms of semantic representation. Study 1 brings word co-occurrence under experimental control to demonstrate that direct co-occurrence in language is necessary for traditional DSMs to successfully reproduce maps. Study 2 presents an instance-based DSM that is capable of reconstructing maps independent of the frequency of co-occurrence of city names.
A large literature suggests that the way we process information is influenced by the categories that we have learned. We examined whether, when we try to uniquely encode items in workingmemory, the information encoded depends on the other stimuli being simultaneously learned. Participants were required to memorize unknown aliens, presented one at the time, for immediate recognition of their features. Some aliens, called twins, were organized into pairs that shared every feature (nondiscriminative feature) except one (discriminative feature), while some other aliens, called hermits, did not share feature. We reasoned that if people develop unsupervised categories by creating a category for a pair of aliens, we should observe better feature identification performance for nondiscriminative features compared to hermit features, but not compared to discriminative features. On the contrary, if distinguishing features draw attention, we should observe better performance when a discriminative rather than nondiscriminative feature was probed. Overall, our results suggest that when items share features, people code items in working memory by focusing on similarities between items, establishing clusters of items in an unsupervised fashion not requiring feedback on cluster membership.
Sloman, S. J., Goldstone, R. L., & Gonzalez, C. (2019). Complex exploration dynamics from simple heuristics in a collective learning environment. Proceedings of the 41st Annual Conference of the Cognitive Science Society. (pp. 2818-2824). Montreal, Canada: Cognitive Science Society.
Effective problem solving requires both exploration and exploitation. We analyze data from a group problem-solving task to gain insight into how people use information from past experiences and from others to achieve explore-exploit trade-offs in complex environments. The behavior we observe is consistent with the use of simple, reinforcement-based heuristics. Participants increase exploration immediately after experiencing a low payoff, and decrease exploration immediately after experiencing a high or improved payoff. We suggest that whether an outcome is perceived as “high” or “low” is a dynamic function of the outcome information available to participants. The degree to which the distribution of observed information reflects the true range of possible outcomes plays an important role in determining whether or not this heuristic is adaptive in a given environment.
Goldstone, R. L., Rogosky, B. J., Pevtzow, R., & Blair, M. (2017). The construction of perceptual and semantic features during category learning. In H. Cohen & C. Lefebvre (Eds.) Handbook of Categorization in Cognitive Science. (pp. 851-882). Amsterdam: Elsevier.
Category learning not only depends upon perceptual and semantic representations; it also leads to the generation of these representations. We describe two series of experiments that demonstrate how categorization experience alters, rather than simply uses, descriptions of objects. In the first series, participants first learned to categorize objects on the basis of particular sets of line segments. Subsequently, participants were given a perceptual part/whole judgment task. Categorization training influenced participants’ part/whole judgments, indicating that whole objects were more likely to be broken down into parts that were relevant during categorization. In the second series, correlations were created or broken between semantic features of word concepts (e.g., ferocious vs. timid and group-oriented vs. solitary animals). The best transfer was found between category learning tasks that shared the same semantic organization of concepts. Together, the experiments support models of category learning that simultaneously create the elements of categorized objects’ descriptions and associate those elements with categories.
Goldstone, R. L., Kersten, A., & Carvalho, P. F. (2017). Categorization and Concepts. In J. Wixted (Ed.) Stevens’ Handbook of Experimental Psychology and Cognitive neuroscience, Fourth Edition, Volume Three: Language & Thought. New Jersey: Wiley. (pp. 275-317).
Concepts are the building blocks of thought. They are critically involved when we reason, make inferences, and try to generalize our previous experiences to new situations. Behind every word in every language lies a concept, although there are concepts, like the small plastic tubes attached to the ends of shoelaces, that we are familiar with and can think about even if we do not know that they are called aglets . Concepts are indispensable to human cognition because they take the “blooming, buzzing confusion” (James, 1890, p. 488) of disorganized sensory experiences and establish order through mental categories. These mental categories allow us to make sense of the world and predict how worldly entities will behave. We see, hear, interpret, remember, understand, and talk about our world through our concepts, and so it is worthy of reflection time to establish where concepts come from, how they work, and how they can best be learned and deployed to suit our cognitive needs.
Meagher, B. J., Carvalho, P. F., Goldstone, R. L., & Nosofsky, R. M. (2017). Organized simultaneous displays facilitate learning of complex natural science categories. Psychonomic Bulletin & Review, DOI 10.3758/s13423-017-1251-6.
Subjects learned to classify images of rocks into the categories igneous, metamorphic, and sedimentary. In accord with the real-world structure of these categories, the to-beclassified rocks in the experiments had a dispersed similarity structure. Our central hypothesis was that learning of these complex categories would be improved through observational study of organized, simultaneous displays of the multiple rock tokens. In support of this hypothesis, a technique that included the presentation of the simultaneous displays during phases of the learning process yielded improved acquisition (Experiment 1) and generalization (Experiment 2) compared to methods that relied solely on sequential forms of study and testing. The technique appears to provide a good starting point for application of cognitive-psychology principles of effective category learning to the science classroom.
Transfer of knowledge is the application of knowledge learned in one context to new, dissimilar problems or situations where the knowledge would be useful. Teachers, coaches, camp counselors, parents, and learners often have the experience of a learner showing apparent understanding when questioned about a topic in a way that closely matches how it was initially presented but showing almost no understanding when queried in a new context or with novel examples. This entry further explains the concept of knowledge transfer. It then discusses several different strategies used to support knowledge transfer.
Comparison and reminding have both been shown to support learning and transfer. Comparison is thought to support transfer because it allows learners to disregard non-matching features of superficially different episodes in order to abstract the essential structure of concepts. Remindings promote memory for the individual episodes and generalization because they prompt learners to retrieve earlier episodes during the encoding of later related episodes and to compare across episodes. Across three experiments, we compared the consequences of comparison and reminding on memory and transfer. Participants studied a sequence of related, but superficially different, proverb pairs. In the comparison condition, participants saw proverb pairs presented together and compared their meaning. In the reminding condition, participants viewed proverbs one at a time and retrieved any prior studied proverb that shared the same deep meaning as the current proverb. Experiment 1 revealed that participants in the reminding condition recalled more proverbs than those in the comparison condition. Experiment 2 showed that the mnemonic benefits of reminding persisted over a one-week retention interval. Finally, in Experiment 3, we examined the ability of participants to generalize their remembered information to new items in a task that required participants to identify unstudied proverbs that shared the samemeaning as studied proverbs. Comparison led to worse discrimination between proverbs related to studied proverbs and proverbs unrelated to studied proverbs than reminding. Reminding supported better memory for individual instances and transfer to new situations than comparison.
Carvalho, P.F., Braithwaite, D.W., de Leeuw, J.R., Motz, B.A., & Goldstone, R.L. (2016). An in vivo study of self-regulated study sequencing in introductory psychology courses. PLoS ONE 11(3): e0152115.
Study sequence can have a profound influence on learning. In this study we investigated how students decide to sequence their study in a naturalistic context and whether their choices result in improved learning. In the study reported here, 2061 undergraduate students enrolled in an Introductory Psychology course completed an online homework tutorial on measures of central tendency, a topic relevant to an exam that counted towards their grades. One group of students was enabled to choose their own study sequence during the tutorial (Self-Regulated group), while the other group of students studied the same materials in sequences chosen by other students (Yoked group). Students who chose their sequence of study showed a clear tendency to block their study by concept, and this tendency was positively associated with subsequent exam performance. In the Yoked group, study sequence had no effect on exam performance. These results suggest that despite findings that blocked study is maladaptive when assigned by an experimenter, it may actually be adaptive when chosen by the learner in a naturalistic context.
Recent research in relational learning has suggested that simple training instances may lead to better generalization than complex training instances. We examined the perceptual encoding mechanisms that might undergird this Simple advantage by testing category and perceptual learning in adults with simplified and traditional (more complex) Chinese scripts. In Experiment 1, participants learned Chinese characters and their English translations, performed a memorization test, and generalized their learning to the corresponding characters written in the other script. In Experiment 2, we removed the training phase and modified the tests to examine transfer based purely on the perceptual similarities between simplified and traditional characters. We found the simple advantage in both experiments. Training with simplified characters produced better generalization than training with traditional characters when generalization relied on either recognition memory or pure perceptual similarities. On the basis of the results of these two experiments,we propose a simple processmodel to explain the perceptual mechanism that might drive this simple advantage, and in Experiment 3 we tested novel predictions of this model by examining the effect of exposure duration on the simple advantage. We found support for our model that the simple advantage is driven primarily by differences in the perceptual encoding of the information available from simple and complex instances. These findings advance our understanding of how the perceptual features of a learning opportunity interact with domain-general mechanisms to prepare learners for transfer.
Learning abstract concepts through concrete examples may promote learning at the cost of inhibiting transfer. The present study investigated one approach to solving this problem: systematically varying superficial features of the examples. Participants learned to solve problems involving a mathematical concept by studying either superficially similar or varied examples. In Experiment 1, less knowledgeable participants learned better from similar examples,while more knowledgeable participants learned better from varied examples. In Experiment 2, prior to learning how to solve the problems, some participants received a pretraining aimed at increasing attention to the structural relations underlying the target concept. These participants, like the more knowledgeable participants in Experiment 1, learned better from varied examples. Thus, the utility of varied examples depends on prior knowledge and, in particular, ability to attend to relevant structure. Increasing this ability can prepare learners to learn more effectively from varied examples.
Cavalho, P. F., Braithwaite, D. W., de Leeuw, J. R., Motz, B. A., & Goldstone, R. L. (2015). Effectiveness of learner-regulated study sequence: An in-vivo study in introductory psychology courses. Proceedings of the Thirty-Seventh Annual Conference of the Cognitive Science Society. (pp. 309-314). Pasadena, CA: Cognitive Science Society.
Study sequence can have a profound impact on learning. Previous research has often shown advantages for interleaved over blocked study, though the reverse has also been found. Learners typically prefer blocking even in situations for which interleaving is superior. The present study investigated learner regulation of study sequence, and its effects on learning in an ecologically valid context – university students using an online tutorial relevant to an exam that counted toward their course grades. The majority of participants blocked study by problem category, and this tendency was positively associated with subsequent exam performance. The results suggest that preference for blocked study may be adaptive under some circumstances, and highlight the importance of identifying task environments under which different study sequences are most effective.
Kost, A., Cavalho, P. F., & Goldstone, R. L. (2015). Can you repeat that? The effect of item repetition on interleaved and blocked study. Proceedings of the Thirty-Seventh Annual Conference of the Cognitive Science Society. (pp. 1189-1194). Pasadena, CA: Cognitive Science Society.
Three experiments explore differences between blocked and interleaved study with and without item repetition. In the first experiment we find that when items are repeated during study, blocked study results in higher test performance than interleaved study. In the second experiment we find that when there is no item repetition, interleaved and blocked study result in equivalent performance during the test phase. In the third experiment we find that when the study is passive and includes no item repetition, interleaved study results in higher test performance. We propose that learners create associations between items of the same category during blocked study and item repetition strengthens these associations. Interleaved study leads to weaker associations between items of the same category and therefore results in worse performance during test when there are item repetitions.
Weitnauer, E., Landy, D., Goldstone, R. L., & Ritter, H. (2015). A computational model for learning structured concepts from physical scenes. Proceedings of the Thirty-Seventh Annual Conference of the Cognitive Science Society. (pp. 2631-2636). Pasadena, CA: Cognitive Science Society.
Category learning is an essential cognitive mechanism for making sense of the world. Many existing computational category learning models focus on categories that can be represented as feature vectors, and yet a substantial part of the categories we encounter have members with inner structure and inner relationships. We present a novel computational model that perceives and learns structured concepts from physical scenes. The perception and learning processes happen simultaneously and interact with each other. We apply the model to a set of physical categorization tasks and promote specific types of comparisons by manipulating presentation order of examples. We find that these manipulations affect the algorithm similarly to human participants that worked on the same task. Both benefit from juxtaposing examples of different categories – especially ones that are similar to each other. When juxtaposing examples from the same category they do better if the examples are dissimilar to each other.
Brunel, L., Carvalho, P. F., & Goldstone, R. L. (2015). It does belong together: Cross-modal correspondences influence cross-modal integration during perceptual learning. Frontiers in Psychology, 6: 358. doi: 10.3389/fpsyg.2015.00358
Experiencing a stimulus in one sensory modality is often associated with an experience in another sensory modality. For instance, seeing a lemon might produce a sensation of sourness. This might indicate some kind of cross-modal correspondence between vision and gustation. The aim of the current study was to provide explore whether such cross-modal correspondences influence cross-modal integration during perceptual learning. To that end, we conducted 2 experiments. Using a speeded classification task, Experiment 1 established a cross-modal correspondence between visual lightness and the frequency of an auditory tone. Using a short-term priming procedure, Experiment 2 showed that manipulation of such cross-modal correspondences led to the creation of a crossmodal unit regardless of the nature of the correspondence (i.e., congruent, Experiment 2a or incongruent, Experiment 2b). However, a comparison of priming-effects sizes suggested that cross-modal correspondences modulate cross-modal integration during learning and thus leading to new learned units that have different stability over time. We discuss the implications of our results for the relation between cross-modal correspondence and perceptual learning in the context of a Bayesian explanation of cross-modal correspondences.
Weitnauer, E., Carvalho, P. F., Goldstone, R. L., & Ritter, H. (2014). Similarity-based ordering of instances for efficient concept learning. Proceedings of the Thirty-Sixth Annual Conference of the Cognitive Science Society. (pp. 1760-1765). Quebec City, Canada: Cognitive Science Society.
Theories in concept learning predict that interleaving instances of different concepts is especially beneficial if the concepts are highly similar to each other, whereas blocking instances belonging to the same concept provides an advantage for learning low-similarity concept structures. This suggests that the performance in concept learning tasks can be improved by grouping the instances of given concepts based on their similarity. To explore this hypothesis, we use Physical Bongard Problems, a rich categorization task with an open feature space, to analyze the combined effects of comparing dissimilar and similar instances within and across categories. We manipulate the within- and between-category similarity of instances presented close to each other in blocked, interleaved and simultaneous presentation schedules. The results show that grouping instances to promote dissimilar within- and similar between category comparisons improves the learning results, to a degree depending on the strategy used by the learner.
Braithwaite, D. W., & Goldstone, R. L. (2014). Benefits of variation increase with preparation. Proceedings of the Thirty-Sixth Annual Conference of the Cognitive Science Society. (pp. 1940-1945). Quebec City, Canada: Cognitive Science Society.
Abstract concepts are characterized by their underlying structure rather than superficial features. Variation in the examples used to teach abstract concepts can draw attention towards shared structure and away from superficial detail, but too much variation can inhibit learning. The present study tested the possibility that increasing attention to underlying structural relations could alleviate the latter difficulty and thereby increase the benefits of varied examples. Participants were trained with either varied or similar examples of a mathematical concept, and were then tested on their ability to apply the concept to new cases. Before training, some participants received pre training aimed at increasing attention to the structural relations underlying the concept. The relative advantage of varied over similar examples was increased among participants who received the pretraining. Thus, preparation that promotes attention to the relations underlying abstract concepts can increase the benefits of learning from varied examples.
Carvalho, P. F., & Goldstone, R. L., (2014). Putting category learning in order: category structure and temporal arrangement affect the benefit of interleaved over blocked study. Memory & Cognition, 42(3), 481-495.
Recent research in inductive category learning has demonstrated that interleaved study of category exemplars results in better performance than does studying each category in separate blocks. However, the questions of how the category structure influences this advantage and how simultaneous presentation interacts with the advantage are open issues. In this article, we present three experiments. The first experiment indicates that the advantage of interleaved over blocked study is modulated by the structure of the categories being studied. More specifically, interleaved study results in better generalization for categories with high within- and between-category similarity, whereas blocked presentation results in better generalization for categories with low within- and between-category similarity. In Experiment 2, we present evidence that when presented simultaneously, between-category comparisons (interleaved presentation) result in a performance advantage for high-similarity categories, but no differences were found for low-similarity categories. In Experiment 3, we directly compared simultaneous and successive presentation of low-similarity categories. We again found an overall benefit for blocked study with these categories. Overall, these results are consistent with the proposal that interleaving emphasizes differences between categories, whereas blocking emphasizes the discovery of commonalities among objects within the same category.
The terms concreteness fading and progressive formalization have been used to describe instructional approaches to science and mathematics that use grounded representations to introduce concepts and later transition to more formal representations of the same concepts. There are both theoretical and empirical reasons to believe that such an approach may improve learning outcomes relative to instruction employing only grounded or only formal representations (Freudenthal, 1991; Goldstone & Son, 2005; McNeil & Fyfe, 2012; but see Kaminski, Sloutsky, & Heckler, 2008). Two experiments tested the effectiveness of this approach to instruction in the mathematical domain of combinatorics, using outcome listing and numerical calculation as examples of grounded and formal representations, respectively. The study employed a pretest-training, posttest design. Transfer performance, that is, participants’ improvement from pretest to posttest, was used to assess the effectiveness of instruction received during training. In Experiment 1, transfer performance was compared for 4 types of instruction, which differed only in the types of representation they employed: pure listing (i.e., listing only), pure formalism (i.e., numerical calculation only), list fading (i.e., listing followed by numerical calculation), and formalism first (i.e., listing introduced after numerical calculation). List fading instruction led to transfer performance on par with pure formalism instruction and higher than formalism first and pure listing instruction. In Experiment 2, an enhanced version of list fading training was again compared to pure formalism. However, no difference in transfer performance due to training was found. The results suggest that combining grounded and formal representations can be an effective approach to combinatorics instruction but is not necessarily preferable to using formal representations alone. If both grounded and formal representations are employed, the former should precede rather than follow the latter in the instructional sequence.
Jones, M., & Goldstone, R. L. (2013). The structure of integral dimensions: Contrasting topological and Cartesian representations. Journal of Experimental Psychology: Human Perception and Performance, 39, 111-132.
Diverse evidence shows that perceptually integral dimensions, such as those composing color, are represented holistically. However, the nature of these holistic representations is poorly understood. Extant theories, such as those founded on multidimensional scaling or general recognition theory, model integral stimulus spaces using a Cartesian coordinate system, just as with spaces defined by separable dimensions. This approach entails a rich geometrical structure that has never been questioned but may not be psychologically meaningful for integral dimensions. In particular, Cartesian models carry a notion of orthogonality of component dimensions, such that if 1 dimension is diagnostic for a classification or discrimination task, another can be selected as uniquely irrelevant. This article advances an alternative model in which integral dimensions are characterized as topological spaces. The Cartesian and topological models are tested in a series of experiments using the perceptual-learning phenomenon of dimension differentiation, whereby discrimination training with integral-dimension stimuli can induce an analytic representation of those stimuli. Under the present task design, the 2 models make contrasting predictions regarding the analytic representation that will be learned. Results consistently support the Cartesian model. These findings indicate that perceptual representations of integral dimensions are surprisingly structured, despite their holistic, unanalyzed nature.
Typical disjunctive artificial classification tasks require participants to sort stimuli according to rules such as “x likes cars only when black and coupe OR white and SUV.” For cate-gories like this, increasing the salience of the diagnostic dimensions has two simultaneous effects: increasing the distance between members of the same category and increas-ing the distance between members of opposite categories. Potentially, these two effects respectively hinder and facilitate classification learning, leading to competing predictions for learning. Increasing saliency may lead to members of the same category to be consid-ered less similar, while the members of separate categories might be considered more dissimilar. This implies a similarity-dissimilarity competition between two basic classifica-tion processes. When focusing on sub-category similarity, one would expect more difficult classification when members of the same category become less similar (disregarding the increase of between-category dissimilarity); however, the between-category dissimi-larity increase predicts a less difficult classification. Our categorization study suggests that participants rely more on using dissimilarities between opposite categories than finding similarities between sub-categories.We connect our results to rule- and exemplar-based classification models.The pattern of influences of within- and between-category similarities are challenging for simple single-process categorization systems based on rules or exem-plars. Instead, our results suggest that either these processes should be integrated in a hybrid model, or that category learning operates by forming clusters within each category.
An enormous amount of ink has been spilled in the psychology literature on the topic of similarity. There are two reasons that this seemingly intuitive and prosaic concept has been the subject of such intense scrutiny. First, there is virtually no area of cognitive processing in which similarity does not seem to play a role. William James observed that “This sense of Sameness is the very keel and backbone of our thinking” (James 1890/1950: 459). Ivan Pavlov first noted that dogs would generalize their learned salivation response to new sounds as a function of their similarity to the original tone, and this pattern of generalization appears to be ubiquitous across species and stimuli. People group things together based on their similarity, both during visual processing and categorization. Research suggests that memories are retrieved when they involve similar features or similar processing to a current situation. Problem solutions are likely to be retrieved from similar prior problems, inductive inference is largely based on the similarity between the known and unknown cases, and the list goes on and on.
Download PDF version of this paper
Carvalho, P. F., & Goldstone, R. L. (2013). How to present exemplars of several categories? Interleave during active learning and block during passive learning. Proceedings of the Thirty-Fifth Annual Conference of the Cognitive Science Society. (pp. 1982-1987). Berlin, Germany: Cognitive Science Society.
Research on how information should be presented during inductive category learning has identified both interleaving of categories and blocking by category as beneficial for learning. Previous work suggests that this mixed evidence can be reconciled by taking into account within- and between-category similarity relations. In this paper we present a new moderating factor. One group of participants studied categories actively, either interleaved or blocked. Another group studied the same categories passively. Results from a subsequent generalization task show that active learning benefits from interleaved presentation while passive learning benefits from blocked presentation.
Weitnauer, E., Carvalho, P. F., Goldstone, R. L., & Ritter, H. (2013). Grouping by Similarity Helps Concept Learning. Proceedings of the Thirty-Fifth Annual Conference of the Cognitive Science Society. (pp. 3747-3752). Berlin, Germany: Cognitive Science Society.
In inductive learning, the order in which concept instances are presented plays an important role in learning performance. Theories predict that interleaving instances of different concepts is especially beneficial if the concepts are highly similar to each other, whereas blocking instances belonging to the same concept provides an advantage for learning lowsimilarity concept structures. This leaves open the question of the relative influence of similarity on interleaved versus blocked presentation. To answer this question, we pit withinand between-category similarity effects against each other in a rich categorization task called Physical Bongard Problems. We manipulate the similarity of instances shown temporally close to each other with blocked and interleaved presentation. The results indicate a stronger effect of similarity on interleaving than on blocking. They further show a large benefit of comparing similar between-category instances on concept learning tasks where the feature dimensions are not known in advance but have to be constructed.
Carvalho, P. F., & Goldstone, R. L. (2012). Category structure modulates interleaving and blocking advantage in inductive category acquisition. Proceedings of the Thirty-Fourth Annual Conference of the Cognitive Science Society. (pp. 186-191). Sapporo, Japan: Cognitive Science Society.
Research in inductive category learning has demonstrated that interleaving exemplars of categories results in better performance than presenting each category in a separate block. Two experiments indicate that the advantage of interleaved over blocked presentation is modulated by the structure of the categories being presented. More specifically, interleaved presentation results in better performance for categories with high within- and between-category similarity while blocked presentation results in better performance for categories with low within- and between-category similarity.
This interaction is predicted by accounts in which blocking promotes discovery of features shared by the members of a category whereas interleaving promotes discovery of features that discriminate between categories.
Download PDF version of this paper
Engle, J. T. Feng, Y., Goldstone, R. L. (2012). Mining relatedness graphs for data integration. Proceedings of the Thirty-Fourth Annual Conference of the Cognitive Science Society. (pp. 1524-1529). Sapporo, Japan: Cognitive Science Society.
In this paper, we present the AbsMatcher system for schema matching which uses a graph based approach. The primary contribution of this paper is the development of new types of relationships for generating graph edges and the effectiveness of integrating schemas using those graphs. AbsMatcher creates a graph of related attributes within a schema, mines similarity between attributes in different schemas, and then combines all information using the ABSURDIST graph matching algorithm. The attribute-to-attribute relationships this paper focuses on are semantic in nature and have few requirements for format or structure. These relationships sources provide a baseline which can be improved upon with relationships specific to formats, such as XML or a relational database. Simulations demonstrate how the use of automatically mined graphs of within-schema relationships, when combined with cross-schema pair-wise similarity, can result in matching accuracy not attainable by either source of information on its own.
Download PDF version of this paper
Terai, A., & Goldstone, R. L. (2012). An Experimental Examination of Emergent Features in Metaphor Interpretation Using Semantic Priming Effects. Proceedings of the Thirty- Fourth Annual Conference of the Cognitive Science Society. (pp. 2399-2404). Sapporo, Japan: Cognitive Science Society.
In comprehension of the metaphor “TOPIC is VEHICLE,” emergent features in the interpretation of metaphors are characteristic neither of the topic nor the vehicle. An experiment examines the hypothesis that new features emerge as metaphoric interpretations through association with nonemergent features connected with the topic, vehicle, or both. In the experiment, participants were presented with a nonemergent feature as a prime, a metaphor, and an emergent feature, sequentially. Participants were then asked to respond as to whether the emergent feature is an appropriate interpretation of the metaphor. The results showed that primed non-emergent features derived from the vehicle facilitate the recognition of emergent features. The results support an account in which new features emerge through two processes – non-emergent features are recognized as interpretations of the metaphor and then these non-emergent features facilitate the recognition of emergent features.
Download PDF version of this paper
Goldstone, R. L., Kersten, A., & Cavalho, P. F. (2012). Concepts and Categorization. In A. F. Healy & R. W. Proctor (Eds.) Comprehensive handbook of psychology, Volume 4: Experimental psychology. (pp. 607-630). New Jersey: Wiley.
Issues related to concepts and categorization are nearly ubiquitous in psychology because of people’s natural tendency to perceive a thing as something. We have a powerful impulse to interpret our world. This act of interpretation, an act of “seeing something as X” rather than simply seeing it (Wittgenstein, 1953), is fundamentally an act of categorization. The attraction of research on concepts is that an extremely wide variety of cognitive acts can be understood as categorizations (Murphy, 2002).
Click here to learn more about this project
Son, J. Y., Smith, L. B., & Goldstone, R. L. (2011). Connecting instances to promote children’s relational reasoning. Journal of Experimental Child Psychology, 108, 260-277.
The practice of learning from multiple instances seems to allow children to learn about relational structure. The experiments reported here have focused on two issues regarding relational learning from multiple instances: (1) what kind of perceptual situations foster such learning and (2) how particular object properties, such as complexity or 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, instances are viewed one at a time, one right after the other. We examine 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 to a common sample) and a variable match-to-sample task (where new instances are compared to 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 showed 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 (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.
Terai, A. & Goldstone, R. L. (2011). Processing emergent features in metaphor comprehension. Proceedings of the Thirty-Third Annual Conference of the Cognitive Science Society. (pp. 2043-2048). Boston, Massachusetts: Cognitive Science Society.
This study examines the processing of emergent features in metaphors. Emergent features are metaphoric interpretations that are characteristic neither of the target nor the vehicle. In the first experiment, participants were asked to respond as to whether a verbal feature is an appropriate interpretation of the metaphor, which was presented as a prime. They are asked to respond immediately after a tone is presented which has a variable temporal lag after the feature. The timing of each tone controlled the participants’ response times. The results show that the response deadline given to the participants only slightly affected their judgments. In a second experiment, the time to interpret a metaphor was controlled by varying the pre- sentation time of the metaphor. The results showed that emer- gent features require more time for recognition as a metaphoric interpretation than do non-emergent features. The results sup- port the hypothesis that interaction among features causes feature emergence.
Previous research has consistently found that spontaneous analogical transfer is strongly tied to concrete and contextual similarities between the cases. However, that work has largely failed to acknowledge that the relevant factor in transfer is the similarity between individuals’ mental representations of the situations, rather than the overt similarities between the cases themselves. Across several studies, we find that participants are able to transfer strategies learned from a perceptually concrete simulation of a physical system to a task with very dissimilar content and appearance. This transfer is reflected in better performance on the transfer task when its underlying dynamics are consistent rather than inconsistent with the preceding training task. Our data indicate that transfer in these tasks relies on the perceptual and spatial nature of the training task, but does not depend on direct interaction with the system, with participants performing equally well after simply observing the concrete simulation. We argue that participants in these studies are using the concrete, spatial, dynamic information presented in the training simulation as the basis for a concretely similar mental model of the dissimilar transfer task. Unexpectedly, our data consistently showed that transfer was independent of reported recognition of the analogy between tasks: while such recognition was associated with better overall performance, it was not associated with better transfer (in terms of applying an appropriate strategy). Together, these findings suggest that analogical transfer between overtly dissimilar cases may be much more common—and much more relevant to our cognitive processing—than is generally assumed.
Carvalho, P. F., & Goldstone, R. L. (2011). Sequential similarity and comparison effects in category learning. Proceedings of the Thirty-Third Annual Conference of the Cognitive Science Society. (pp. 2977-2982). Boston, Massachusetts: Cognitive Science Society.
Order effects in category learning have been previously demonstrated. Specifically, alternation between exemplars of two categories has been shown to improve category learning and discrimination, compared to presenting exemplars of each category in separate blocks. However, the mechanisms under- lying order effects are still not completely known. Remaining issues pertain to the relevance of within and between category similarities, and the role of comparing sequentially presented objects. We present two experiments: in Experiment 1 within- and between-category similarity are manipulated simultane- ously with presentation schedule. In Experiment 2, alternation between categories is compared to two blocked conditions: one in which very similar stimuli are presented successively, and another in which they are dissimilar. Our results show a clear overall advantage of low similarity in categorization performance, but no effect of presentation schedule. Also, al- ternation between categories is shown to result in better per- formance than the blocked condition with more dissimilar stimuli.
Day, S., & Goldstone, R. L. (2010). The Effects of Similarity and Individual differences on Comparison and Transfer. Proceedings of the Thirty-Second Annual Conference of the Cognitive Science Society. (pp. 465-470). Portland, Oregon: Cognitive Science Society.
Prior research has found that while people are generally quite poor at recognizing when a new situation is structurally similar to a known case, comparison of two analogous cases greatly improves the likelihood of achieving such recognition. Our study examines the effects of varying the similarity between these compared cases, both featurally and structurally. We find that between-case similarity has a significant impact on transfer, and that these effects interact with characteristics of the learner.
Goldstone, R. L., Hills, T. T., & Day, S. B. (2010). Concept formation. In. I. B. Weiner & W. E. Craighead (Eds.) The Corsini Encyclopedia of Psychology. New York: John Wiley & Sons. (pp. 381-383).
A concept is a mentally possessed idea or notion that can be used to categorize information or objects. Over the course of each person’s lifetime, thousands of concepts are learned, for nouns like corkscrew, justice, and doorknob, adjectives like green, symmetric, and beautiful, and verbs like kick, climb, and eschew. While some philosophers have maintained that we do not genuinely learn new concepts through induction (Fodor, 1988), most psychologists believe that concepts can be learned, and that the representational capacity of the learner increases as they acquire new concepts. Most efforts have been spent developing accounts of how people acquire and represent concepts, including models based on: rules, prototypes, exemplars, boundaries, and theories.
Goldstone, R. L., Day, S., & Son, J. Y. (2010). Comparison. In B. Glatzeder, V. Goel, & A. von Müller (Eds.) On thinking: Volume II, towards a theory of thinking. Heidelberg, Germany: Springer Verlag GmbH. (pp. 103-122).
It might not be immediately clear why the topic of comparison warrants a whole chapter in a book on human thinking. Of course, we are often required to make decisions that involve comparing two or more alternatives and assessing their relative value. Which car should I buy? Which job is more suited to my long-term goals? Would I rather have the soup or the salad? But in the grand scheme of human cognition, it might seem that such processes could be relegated to a subheading in a chapter on decision making. In fact, comparison is one of the most integral components of human thought. Along with the related construct of similarity, comparison plays a crucial role in almost everything that we do. Furthermore, comparison itself is a powerful cognitive tool—in addition to its supporting role in other mental processes, research has demonstrated that the simple act of comparing two things can produce important changes in our knowledge.
Son, J. Y., Doumas, L. A., & Goldstone, R. L. (2010). When do words promote analogical transfer? The Journal of Problem Solving, 3, 52-92.
The purpose of this paper is to explore how and when verbal labels facilitate relational reasoning and transfer. We review the research and theory behind two ways words might direct attention to relational information: (1) words generically invite people to compare and thus highlight relations (the Generic Tokens [GT] hypothesis), and/or (2) words carry semantic cues to common structure (the Cues to Specific Meaning [CSM] hypothesis). Four experiments examined whether learning Signal Detection Theory (SDT) with relational words fostered better transfer than learning without relational words in easily alignable and less alignable situations (testing the GT hypothesis) as well as when the relational words matched and mismatched the semantics of the learning situation (testing the CSM hypothesis). The results of the experiments found support for the GT hypothesis because the presence of relational labels produced better transfer when two situations were alignable. Although the CSM hypothesis does not explain how words facilitate transfer, we found that mismatches between words and their labeled referents can produce a situation where words hinder relational learning.
Day, S. B., & Goldstone, R. L. (2009). Analogical transfer from interaction with a simulated physical system. Proceedings of the Thirty-First Annual Conference of the Cognitive Science Society, 1406-1411. Amsterdam, Netherlands: Cognitive Science Society.
In two studies, we find that participants are able to transfer strategies learned while interacting with a simulated physical system to a dissimilar and less perceptually-concrete domain. Interestingly, performance on the transfer task was completely unrelated to explicit knowledge of the structural correspondences between the systems. We suggest that direct interaction with a concrete system may lead to a kind of procedural knowledge that provides a good basis for analogical transfer.
Baldwin D. & Goldstone R.L. (2007). Finding analogies within systems: Constraints on unsegmented matching. Proceedings of the Workshop on Analogies: Integrating Multiple Cognitive Abilities (AnICA07). Nashville, Tennessee.
The complex structure and organization of knowledge in the human mind is one of the key facets of thought. One of the fundamental cognitive processes that oper- ates over that structure is analogy. A typical compu- tational model of analogy might juxtapose a source do- main and a target domain, such as the solar system and the Bohr-Rutherford (BR) model of an atom (Gentner, 1983). The goal is to find a correspondence mapping between these two domains. Determining a mapping between the source and target domains of a non-trivial size would be intractable without a set of constraints to restrict the set of correspondences that are considered by a human reasoner. Moreover, the mere presence of domains serve as a constraint on mapping. In this paper, we study an alternative problem called unsegmented mapping – correspondence without specification of domains. We show a series of three formal constraints that allow for analogical-like mappings without explicit segmentation. The result, correspondence is possible without domains, has implications for models of analogical reasoning as well as schema induction and inference.
Feng, Y., Goldstone, R. L., & Menkov, V. (2005). A Graph Matching Algorithm and its Application to Conceptual System Translation. International Journal on Artificial Intelligence Tools, 14, 77-100.
ABSURDIST II, an extension to ABSURDIST, is an algorithm using attributed graph matching to find translations between conceptual systems. It uses information about the internal structure of systems by itself, or in combination with external information about concept similarities across systems. It supports systems with multiple types of weighted or unweighted, directed or undirected relations between concepts. The algorithm exploits graph sparsity to improve computational efficiency.We present the results of experiments with a number of conceptual systems, including artificially constructed random graphs with introduced distortions.
Goldstone, R. L., Feng, Y., & Rogosky, B. (2005). Connecting concepts to the world and each other. In D. Pecher & R. Zwaan (Eds.) Grounding cognition: The role of perception and action in memory, language, and thinking. Cambridge: Cambridge University Press. (pp. 292-314)
How can well tell that two people both have a concept of dog, gold, or car despite differences in their conceptual knowledge? Two kinds of information can be used to translate between the concepts in two persons’ minds: the internal relations between concepts within each person’s mind, and external grounding of the concepts. We present a neural network model called ABSURDIST (Aligning Between Systems Using Relations Derived Inside Systems Themselves) that integrates internal and external determinants of conceptual meaning to find translations across people or other systems. The model shows that appropriate translations can be found by considering only similarity relations among concepts within a person. However, simulations also indicate synergistic interactions between internal and external sources of information. ABSURDIST is then applied to analogical reasoning, dictionary translation, translating between web-based ontologies, subgraph matching, and object recognition. The performance of ABSURDIST suggests the utility of concepts that are simultaneously externally grounded and enmeshed within a conceptual system.
Human assessments of similarity are fundamental to cognition because similarities in the world are revealing. The world is an orderly enough place that similar objects and events tend to behave similarly. This fact of the world is not just a fortunate coincidence. It is because objects are similar that they will tend to behave similarly in most respects. It is because crocodiles and alligators are similar in their external form, internal biology, behavior, diet, and customary environment that one can often successfully generalize from what one knows of one to the other. As Quine (1969) observed, “Similarity, is fundamental for learning, knowledge and thought, for only our sense of similarity allows us to order things into kinds so that these can function as stimulus meanings. Reasonable expectation depends on the similarity of circumstances and on our tendency to expect that similar causes will have similar effects (p. 114).” Similarity thus plays a crucial role in making predictions because similar things usually behave similarly.
ABSURDIST II, an extension to ABSURDIST, is an algorithm using attributed graph matching to find translations between conceptual systems. It uses information about the internal structure of systems by itself, or in combination with external information about concept similarities across systems. It supports systems with multiple types of weighted or unweighted, directed or undirected relations between concepts. The algorithm exploits graph sparsity to improve computational efficiency. We present the results of experiments with a number of conceptual systems, including artificially constructed random graphs with introduced distortions.
Goldstone, R. L., & Rogosky, B. J. (2002). Using relations within conceptual systems to translate across conceptual systems, Cognition, 84, 295-320.
We explore one aspect of meaning, the identification of matching concepts across systems (e.g. people, theories, or cultures). We present a computational algorithm called ABSURDIST (Aligning Between Systems Using Relations Derived Inside Systems for Translation) that uses only within-system similarity relations to find between-system translations. While illustrating the sufficiency of within-system relations to account for translating between systems, simulations of ABSURDIST also indicate synergistic interactions between intrinsic, within-system information and extrinsic information.
Here is a brief description and commentary on ABSURDIST:
Dietrich, E. (2003). An ABSURDIST model vindicates a venerable theory. Trends in Cognitive Science, 7, 57-59.
Goldstone, R. L., & Rogosky, B. J. (2002). The role of roles in translating across conceptual systems, Proceedings of the Twenty-fourth Annual Conference of the Cognitive Science Society. Hillsdale, New Jersey: Lawrence Erlbaum Associates. (pp. 369-374).
According to an “external grounding” theory of meaning, a concept’s meaning depends on its connection to the external world. By a “conceptual web” account, a concept’s meaning depends on its relations to other concepts within the same system. We explore one aspect of meaning, the identification of matching concepts across systems (e.g. people, theories, or cultures). We present a computational algorithm called ABSURDIST (Aligning Between Systems Using Relations Derived Inside Systems for Translation) that uses only within-system similarity relations to find between-system translations. While illustrating the sufficiency of a conceptual web account for translating between systems, simulations of ABSURDIST also indicate powerful synergistic interactions between intrinsic, within-system information and extrinsic information. Applications of the algorithm to issues in object recognition, shape analysis, automatic translation, human analogy and comparison making, pattern matching, neural network interpretation, and statistical analysis are described.
Goldstone, R. L, & Steyvers, M. (2001). The Sensitization and Differentiation of Dimensions During Category Learning. Journal of Experimental Psychology: General, 130,116-139.
The reported experiments explore two mechanisms by which object descriptions are flexibly adapted to support concept learning: selective attention and dimension differentiation. Arbitrary dimensions were created by blending photographs of faces in different proportions, and mixing these blends together. Consistent with learned selective attention, positive transfer was found when initial and final categorizations shared either relevant or irrelevant dimensions, and negative transfer was found when previously relevant dimensions became irrelevant. Unexpectedly good transfer was observed when both irrelevant dimensions became relevant and relevant dimensions became irrelevant, and was explained in terms of participants learning to isolate one dimension from another. This account was further supported by experiments indicating that conditions expected to produce positive transfer via dimension differentiation produced better transfer than conditions expected to produce positive transfer via selective attention, but only when stimuli were composed of highly integral and overlapping dimensions. We discuss the relation between dimension differentiation and selective attention, mechanisms that may underlie these processes, and implications for category learning research.
An ability to assess similarity lies close to the core of cognition. In the time-honored tradition of legitimizing fields of psychology by citing William James, `This sense of Sameness is the very keel and backbone of our thinking` (James, 1890/1950; p. 459). Similarity plays an indispensable foundational role in theories of cognition. People`s success in problem solving depends on the similarity of previously solved problems to current problems. Categorization depends on the similarity of objects to be categorized to category members. Memory retrieval depends on the similarity of retrieval cues to stored memories. Inductive reasoning is based on the principle that if an event is similar to a previous event, then similar outcomes are predicted. An understanding of these cognitive processes requires that we understand how humans assess similarity. Four major psychological models of similarity are: geometric, featural, alignment-based, and transformational.
Goldstone, R. L. (1998). Hanging Together: A connectionist model of similarity. In J. Grainger & A. M. Jacobs (Eds.) Localist Connectionist Approaches to Human Cognition. (pp. 283 – 325). Mahwah, NJ: Lawrence Erlbaum Associates.
Human judgments of similarity have traditionally been modelled by measuring the distance between the compared items in a psychological space, or the overlap between the items` featural representations. An alternative approach, inspired jointly by work in analogical reasoning (D. Gentner, 1983; K. T. Holyoak & P. Thagard, 1989) and interactive activation models of perception (J. L. McClelland & D. E. Rumelhart, 1981), views the process of judging similarity as one of establishing alignments between the parts of compared entities. A localist connectionist model of similarity, SIAM, is described wherein units represent correspondences between scene parts, and these units mutually and concurrently influence each other according to their compatability. The model is primarily applied to similarity rating tasks, but is also applied to other indirect measures of similarity, to judgments of alignment between scene parts, to impressions of comparison difficulty, and to patterns of perceptual sensitivity for matching and mismatching features.
Similarity comparisons are highly sensitive to judgment context. Three experiments explore context effects that occur within a single comparison rather than across several trials. Experiment 1 shows reliable intransitivities in which a target is judged to be more similar to stimulus A than to stimulus B, more similar to B than to stimulus C, and more similar to C than to A. Experiment 2 explores the locus of Tversky`s (1977) diagnosticity effect in which the relative similarity of two alternatives to a target is influenced by a third alternative. Experiment 3 demonstrates reliable, though occasional, violations of an assumption of monotonicity. The observed violations of common assumptions to many models of similarity can be accomodated in terms of dynamic property weighting processes based on specific forms of diagnosticity, and contrast sets that are generated when a comparison is presented.
(Translated into Japanese as: Spencer-Smith, J., & Goldstone, R. L. (2001). The dynamics of similarity. in A. Ohnishi and H. Suzuki (Eds.) Ruii kara mita kokoro (Similarity-based approach to mind). Tokyo, Japan: Kyoritsu Shuppan.)
Similarity depends on representations of stimuli that are constructed and changed during comparison-making. Specific features may be selectively weighted during comparison, and the features used in a comparison may themselves be a product of the comparison process. Traditional models of similarity and analogy rely on representations that are assumed to exist prior to comparison and are inflexible. Evidence from previous research indicates that weighting of features in similarity judgments may vary dynamically during processing (Goldstone, 1994; Goldstone & Medin, 1994). SIAM (Goldstone, 1994), a model providing an account of dynamic weighting, is discussed. Additional studies indicate that features may be developed or introduced during similarity judgments. A methodology for examining process-oriented models that may account for flexible representations is proposed.
According to the assumption of monotonicity in similarity judgments, adding a shared feature in common to two items should increase or leave unchanged, but should never decrease, their similarity. Violations of monotonicity are not predicted by feature- or dimension-based models, but can be accommodated by alignment-based models. According to alignment-based models, when structured displays are compared, the parts of one compared display must be aligned, or placed in correspondence with the parts of the other display. In two experiments, evidence for nonmonotonicities is obtained that is generally, although not entirely, consistent with the alignment-based model SIAM (Similarity as Interactive Activation and Mapping; Goldstone, 1994). The primary assumption of the model is that the calculation of similarity involves an interactive activation process whereby correspondences between the parts of compared displays mutually and concurrently influence each other. As SIAM predicts, the occurrence of nonmonotonicities depends on the perceptual similarity of features and the duration of presented comparisons.
Most studies of the categorization of emotions tests the prototype model against the classical model, concluding that the prototype model offers the better explanation. Prototype models, as with all similarity-based models, posit that categorization depends on the similarity between the instance to be categorized and the category representation. However, we find that emotion similarity judgments and categorization judgments sometimes diverge. Specifically, information about changes in a person`s status and/or potency is weighted more heavily in categorization decisions than it is in similarity decisions. We argue that a knowledge-based model, rather than a similarity-based model, offers the best account of emotion categorization when information about status and potency changes is available.
In the first part of this article, empirical evidence is reviewed that suggests a substantial amount of flexibility and context-sensitivity in peopleUs judgments of similarity. Four examples of flexible similarity from our laboratory are considered in detail. In the second part of the article, evidence for relatively constrained, invariant similarity assessments is considered. In the final section, a resolution to these apparently contradictory views on similarity is proposed. Assessments of similarity are used to make inferences from one entity to another. In some situations, flexible similarity is needed to tailor inferences to oneUs knowledge of the entities and their relations. In other situations, particularly those in which specific knowledge is missing or unavailable, a relatively constant similarity is needed to establish generally permissible inferences. Thus, the flexibility and stability of similarity may reflect its different cognitive uses.
Medin, D. L., & Goldstone, R. L. (1995). The predicates of similarity. In C. Cacciari (Ed.), Similarity in Language, Thought, and Perception. (pp. 83-110). Brussels: BREPOL.
Research and theory in decision making and in similarity judgment have developed along parallel paths. We review and analyze phenomena in both domains that suggest that similarity processing and decision making share important correspondences. The parallels are explored at the level of empirical generalizations and underlying processing principles. Important component processes that are shared by similarity judgments and decision making include generation of alternatives, recruitment of reference points, dynamic weighting of aspects, creation of new descriptors, development of correspondences between items, and justification of judgement.
Four experiments investigated the influence of categorization training on perceptual discriminations. Ss were trained according to 1 of 4 different categorization regimes. Subsequent to category learning, Ss performed a Same-Different judgement task. Ss` sensitivities (d`s) for discriminating between items that varied on a category(ir)relevant dimensions were measured. Evidence for acquired distinctiveness (increased perceptual sensitivity for items that are categorized differently) was obtained. One case of acquired equivalence (decreased perceptual sensitivity for items that are categorized together) was found for separable, but not integral, dimensions. Acquired equivalence within a categorization-relevant dimension was never found for either integral or separable dimensions. The relevance of the results for theories of perceptual learning, dimensional attention, categorical perception, and categorization are discussed.
The relation between similarity and categorization has recently come under scrutiny from several sectors. The issue provides an important inroad to questions about the contributions of high-level thought and lower-level perception in the development of people`s concepts. Many psychological models base categorization on similarity, assuming that thing belong in the same category because of their similarity. Empirical and in-principle arguments have recently raied objections to this connection, on the grounds that similarity is too unconstrained to provide an explanation of categorization, and similarity is not sufficiently sophisticated to ground most categories. Although these objections have merit, a reassesment of evidence indicates that similarity can be sufficiently constrained and sophisticated to provide at least a partial account of many categories. Principles are discussed for incorporating similarity into theories of category formation.
The question of “what makes things seem similar?” is important both for the pivotal role of similarity in theories of cognition and for an intrinsic interest in how people make comparisons. Similarity frequently involves more than listing the features of the things to be compared and comparing the lists for overlap. Often, the parts of one thing must be aligned or placed in correspondence with the parts of the other. The quantitative model with the best overall fit to human data assumes an interactive activation process whereby correspondences between the parts of compared things mutually and concurrently influence each other. An essential aspect of this model is that matching and mismatching features influence similarity more if they belong to parts that are placed in correspondence. In turn, parts are placed in correspondence if they have many features in common and if they are consistent with other developing correspondences.
Measurements of similarity have typically been obtained through the use of rating, sorting, and perceptual confusion tasks. In the present paper, a new method of measuring similarity is described, in which subjects rearrange items so that their proximity on a computer screen is proportional to their similarity. This method provides very efficient data collection. If a display has nobjects, then, after subjects have rearranged the objects (requiring slightly more than n movements), n(n-1)/2 pairwise similarityes can be recorded. As long as the constraints imposed by two-diminsional space are not too different from those intrinsic to psychological similarity, the technique apears to offer an efficient, user-friendly, and intuitive process for measuring psychological similarity.
Similarity as interactive activation and mapping (SIAM), a model of the dynamic course of similarity comparisons, is presented. According to SIAM, when structured scenes are compared, the parts of one scene must be aligned, or placed in correspondence, with the parts from the other scene. Emerging correspondences influence each other in a manner such that, with sufficient time, the strongest correspondences are those that are globally consistent with other correspondences. Relative to globally inconsistent feature matches, globally consistent feature matches influence similarity more when greater amounts of time are given for a comparison. A common underlying process model of scene alignment accounts for commonalities between different task conditions. Differences between task conditions are accounted for by principled parametric variation within the model.
Goldstone, R.L., & Medin, D.L. (1994). Interactive activation, similarity, and mapping: An overview. in K. Holyoak and J. Barnden (Eds.) Advances in Connectionist and Neural Computation Theory, Vol. 2: Analogical Connections. (pp. 321-362). Ablex : New Jersey.
The organization of this chapter is as follows. First, we review the role of mapping and global consistency in both low-level visual perception and abstract analogy and then suggest that mapping and consistency also apply to similarity assessment. Next, we review current models of similarity and note that they have little to say about processes by which corresponding properties are aligned. We then describe some experiments on alignment processes associated with comparisons. We account for these results with an interactive activation model of alignment and contrast this model with a number of alternative.s Finally, we asses the role of mapping or alignment in comparisons more generally and offer some conclusions.
This article reviews the status of similarity as an explanatory construct with a focus on similarity judgements. For similarity to be a useful construct, one must be able to specify the ways or respects in which two things are similar. One solution to this problem is to restruct the notion of similarity to hard-wired perceptual processes. It is argued that this view is too narrow and limiting. Instead, it is proposed that an important source of constraints derives from the similarity comparison process itself. Both new experiments and other evidence are described that support the idea that respects are determined by processes internal to comparisons.
Four experiments examined the hypothesis that simple attributional features and relational features operate differently in the determination of similarity judgements. Forced choice similarity judgments (“Is X or Y more similar to Z?”) and similarity rating tasks demonstrate that making the same featural change in two geometric stimuli unequally affects their judged similarity to a third stimulus (the comparison stimulus). More specifically, a featural change that causes stimuli to be more superficially similar and less relationally similar increases judged similarity if it occurs in stimuli that already share many superficial attributes, and decreases similarity if it occurs in stimuli that do not share as many superficial attributes. These results argue against an assumption of feature independence which asserts that the degree to which a feature shared by two objects affects similarity is independent of the other features shared by the objects. The MAX hypothesis is introduced, in which attributional and relational similarities are separately pooled, and shared features affect similarity more if the pool they are in is already relatively large. The results support claims that relations and attributes are psychologically distinct and that formal measures of similarity should not treat all types of matching features equally.
Conventional wisdom and previous research suggest that similarity judgements and difference judgements are inverses of one another. An exception to this rule arises when both relational similarity and attributional similarity are considered. When presented with choices that are relationally or attributionally similar to a standard, human subjects tend to pick the relationally similar choice as more similar to the standard and as more different from the standard. These results not only reinforce the general distinction between attributes and relations but also show that attributes and relations are dynamically distinct in the processes that give rise to similarity and difference judgments.
Goldstone, R. L., Gentner, D., & Medin, D. L. (1989). Relations Relating Relations. Proceedings of the Eleventh Annual Conference of the Cognitive Science Society. (pp. 131-138). Hillsdale, New Jersey: Lawrence Erlbaum Associates.