The ability to learn and use categories is a characteristic of intelligent behaviour. Concepts and categories allow a person or animal to generalize information to new situations or to previously unseen objects. Categories also allow for many variations of an item to be treated as the same thing, known a a behavioural equivalence class. In this course, we will review the major theoretical approaches to the psychology of concepts, conceptual behaviour, and theoretical accounts of category learning. We will also review current computational and neurobiological models of classification and categorical decision making. We will also spend some time getting into the mechanics of several computational models and we will use these models to create simulations of category learning scenarios.
My office hours are Thursdays from 12:30-2:30, room 5158 in the Western Interdisciplinary Research Building (WIRB). For most questions, email probably works best. For class related questions, I will try to respond within a 48-hour window.
Most classes will be a seminar discussion of the topic and assigned readings and each class will be led by one or more student. Seminar leaders will prepare by reading the material, creating slides (as needed), presenting a summary of the research paper, identifying new or supplemental research, designing demonstrations, and most importantly by acting as a discussion leader. Classes should be discussions and not just static presentations.
Note: Our class runs from 9-12 and the BMI Coffee break is at 10:30 on Fridays. So we’ll probably take a short break at 10:30 for coffee, tea, and snacks. Pretty convenient.
Every student is expected to read the literature prior to class, to be familiar with the key concepts, to address problems or concerns with the work, to contribute to the discussion, and to stay on topic.
The readings are all available on the course OWL site arranged by week. I have curated this reading list carefully to include a mix of review chapters that give a broad overview of some of the topics, original theory and empirical papers that have influenced the field for decades, and newer papers that develop the theories, introduce new assumptions, or challenge the ideas in the original papers. Please read the papers in the order assigned before we meet. Most classes include 3-5 papers, so I advise that you spread the reading out over a few days. Part of this class involves learning how to read papers and how to manage the reading list. This can be accomplish in many ways, but here are some suggestions to help you develop a good reading routine:
With the exception on Class 1 and Class 8, which I will lead, I expect students to lead the discussions and present papers to the class. Given the number of papers (3-5/week), the anticipated number of students (8-10) and the time (2 1/2 hours) what I envision is that a small group of students will lead each class through an integrative discussion of the week’s reading rather than a serial presentation of each paper. Work does not exist in a vacuum and we should not treat these articles and chapters as island.
There is no perfect format for this, and I am not grading your on your ability to present papers. But I do expect students to lead an interesting discussion and to present the research in a clear manner. Here are some guidelines:
Evaluation in the course is a combination of the evaluations on weekly discussion papers and a final review paper. This is a graduate seminar, so I expect everyone to do well. The evaluations will be a way to keep track of this.
Prior to each class, for classes 2-11, please complete a 1-2 page (single-spaced) discussion paper to be submitted on OWL by Thursday evening. Submit a PDF, a file, rather than a Word or Docs file. Aim for one page and try not not to exceed two pages. A good thought paper will address a key theme in the readings, suggest a new experiment in a particular area, point out problems with particular theory, etc. The goal of the discussion paper is to get you thinking about what you want to discuss in the class. Each discussion paper will be graded on a scale of 0-4 and the final discussion paper contribution to your mark is 40 points out of the 100 points possible. The discussion papers will be graded with the following rubric:
Students will complete a final project by April 24th. This project topic must be approved by me prior to March 20. Please avoid late submission or avoid the temptation to ask for an extension. I can grant an extension, but I prefer to have everyone keep to the same timeline.
The project should be an integrative or systematic review of a central topic or topics in categorization. There are many possible themes and your paper can cover an important theory, several theories, a model, a comparison of models, a paradigm, or methodology. You may wish to pick one of the topics that we covered in this course and go into greater depth. You can also pick or pick a new topic that was not covered in the course but is directly related to categorization.
The paper can be broad or narrow in focus, depending on the topic. It should be publishable, though submitting for publication is not required for this class. But don’t let that stop you from submitting it to a journal. If you decide to do this, we can discuss after the paper has been marked and the course is over. The paper should be completed with the following guidelines:
The paper will be graded with the following rubric, with the total points 60/60.
The final grade in this class will be a combination of your grade on discussion papers (40%) and the final paper (60%).
The first class will include some organization. We need to decide who is presenting which papers and organize each class. I will update the course outline and schedule with this this information Please read the short introductory chapter by Murphy and the paper by Love and we will discuss them in class. I will lead the discussion. Brad Love’s paper is really good here, and it is a good overview of a lot of what we will talk about in the class.
Murphy, G. (2004). Chapter 1: Introduction. In The Big Book of Concepts, pages 1–9. MIT Press, Cambridge, MA
Love, B. C. (2017). Concepts, Meaning, and Conceptual Relationships. In Chipman, S. E. F., editor, The Oxford Handbook of Cognitive Science. Oxford University Press
This class covers an introduction to theories and models that are usually called “classical models”. These models assume strict class membership. These models may seem too strict to account for the available data but they have motivated the development of alternative models. Please read the articles in the order assigned and be prepared to discuss even if you are not leading the discussion. Remember to submit your first discussion paper prior to the class.
Presented by Carolina and Josh
Murphy, G. (2004). Chapter 2: Typicality and the classical view of categories. In The Big Book of Concepts, chapter 2, pages 11–40. MIT Press, Cambridge, MA
Markman, A. B. and Dietrich, E. (2000). Extending the classical view of representation. Trends in Cognitive Sciences, 4(12):470–475
Margolis, E. (1994). A reassessment of the shift from the classical theory of concepts to prototype theory. Cognition, 51(1):73–89
Family resemblance theory was seen as a counter to classical models and also influenced the development of subsequent ideas like prototype theory. We will also discuss the idea that some concepts may be organized in the mind as relational hierarchies.
Presented by Bailey and Ana
Rosch, E. and Mervis, C. B. (1975). Family resemblances: Studies in the internal structure of categories. Cognitive Psychology, 7(4):573–605
Medin, D. L., Wattenmaker, W. D., and Hampson, S. E. (1987). Family resemblance, conceptual cohesiveness, and category construction. Cognitive Psychology, 19(2):242–279
Murphy, G. (2004). Chapter 7: Taxonomic organization and the basic level of concepts. In The Big Book of Concepts, pages 199–242. MIT Press: Cambridge, MA
In this class, the readings include one of the original papers in the field Reed (1972) along with Murphy’s chapter and Minda and Smith’s updated computational work. Bowman and Zeithamova (2018) explore the idea of prototypes in brain imaging.
Presentd by Claudia and Maz
Murphy, G. (2004). Chapter 3: Theories. In The Big Book of Concepts, chapter 3, pages 41–71. MIT Press, Cambridge,MA
Reed, S. K. (1972). Pattern recognition and categorization. Cognitive Psychology, 3(3): 382–407
Minda, J. P. and Smith, J. D. (2001). Prototypes in category learning: the effects of category size, category structure, and stimulus complexity. Journal of Experimental Psychology: Learning, Memory, and cognition,27(3):775–799
Bowman, C. R. and Zeithamova, D. (2018). Abstract memory representations in the ventromedial prefrontal cortex and hippocampus support concept generalization. The Journal of Neuroscience, 38(10):2605–2614
There is no meeting this week. This is the Psychology Department’s annual graduate recruitment day. If I had had my act together in the summer and requested a different time for our class, we would not have ended up on Friday morning. But I didn’t and we did.
Presented by Toka and Evgenii
Exemplar theory came about in the late 1970 and 1980s as an strong alternative to prototype theory. It has been influential and is one of the most extensively researched representation models in psychology. Please read the papers in order and consider carefully some of Murphy’s criticism.
Medin, D. L. and Schaffer, M. M. (1978). Context theory of classification learning. Psychological Review, 85(3):207
Nosofsky, R. M. (1986). Attention, similarity, and the identification–categorization relationship. Journal of Experimental Psychology: General, 115(1):39
Nosofsky, R. M., Sanders, C. A., and McDaniel, M. A. (2018). Tests of an exemplar-memory model of classification learning in a high-dimensional natural-science category domain. Journal of Experimental Psychology: General, 147(3):328–353
Murphy, G. L. (2016). Is there an exemplar theory of concepts? Psychonomic Bulletin & Review, 23(4):1035– 1042
The winter break is primarily for undergraduate classes, but we will also take the week off. Engage in some self care. Work on a side project. Work ahead on your thesis or plan out what you want to write for the final project. Enjoy the winter.
I will lead this class and discuss the formalization of the prototype model and the exemplar model. We will explore how each model works, how to generate predictions, how to simulate data, and how to fit data by minimizing parameters. I will use Python for this, but no direct experience is needed. This will be a demonstration. Please read the two chapters and submit a discussion paper.
Presented by Paul * Nosofsky, R. M. (2011). The generalized context model: An exemplar model of classification. In Pothos, E. M. and Wills, A. J., editors, Formal Approaches in Categorization, pages 18–39. Cambridge University Press, Cambridge, UK
These models are all built around the idea that concepts and categories can be learned and represented within a distributed neural network. The earlier papers are especially good because the models are straightforward and clear. Kruschke’s model is an attempt to combine the exemplar representations of Nosofsky’s GCM with a neural-network architecture.
Presented by Carolina, Lindsay, Ana
Gluck, M. A. and Bower, G. H. (1988). From conditioning to category learning: an adaptive network model. Journal of Experimental Psychology: General, 117(3):227–247
Gluck, M. A. (1991). Stimulus generalization and representation in adaptive network models of category learning. Psychological Science
Kruschke, J. K. (1992). ALCOVE: an exemplar-based connectionist model of category learning. Psychological Review, 99(1):22–44
Lee, M. D. and Navarro, D. J. (2002). Extending the ALCOVE model of category learning to featural stimulus domains. Psychonomic Bulletin & Review, 9(1):43–58
Kutrz, K. J. (2007). The divergent autoencoder (DIVA) model of category learning. Psychonomic Bulletin & Review, 14(4):560–576
Each of these papers looks at how a rational system organizes objects into categories. A key features of all of these theories and models is that concepts have a structure that is sensitive and to the demands of the environment and the learning situation. These models are not necessarily counter to the representational models, such as the prototype or exemplar model, but they make different assumptions about how objects go together.
Presented by Toka and Maz * Anderson, J. R. (1991). The adaptive nature of human categorization. Psychological Review, 98(3):409–429
Feldman, J. (2000). Minimization of Boolean complexity in human concept learning. Nature, 407(6804):630–633
Love, B. C., Medin, D. L., and Gureckis, T. M. (2004). SUSTAIN: a network model of category learning. Psychological Review, 111(2):309–332
Weeks 11 and 12 will be devoted to the study of how concepts are learned and represented at the neural level. This area is almost big enough for its own course, so I’ve constrained the discussion to models that looks at how the brain learns to classify. As with the rational approaches above, these theories are not necessarily counter to the formal accounts above (prototype, exemplar, network) but are rather trying to answer questions about how the brain instantiates concepts and category learning.
Presented by Evgenii, Claudia, Bailey * Ashby, F. G., Alfonso-Reese, L. A., Turken, A. U., and Waldron, E. M. (1998). A neuropsychological theory of multiple systems in category learning. Psychological Review, 105(3):442–481
Smith, E. E. and Grossman, M. (2008). Multiple systems of category learning. Neuroscience and Biobehavioral Reviews, 32(2):249–264
Seger, C. A. and Miller, E. K. (2010). Category learning in the brain. Annual Review of Neuroscience, 33:203–-219
Martin, C. B., Douglas, D., Newsome, R. N., Man, L. L., & Barense, M. D. (2018). Integrative and distinctive coding of visual and conceptual object features in the ventral visual stream. eLife, 7, 1-29
Zeithamova, D., Mack, M. L., Braunlich, K., Davis, T., Seger, C. A., van Kesteren, M. T. R., and Wutz, A. (2019). Brain mechanisms of concept learning. The Journal of Neuroscience, 39(42): 8259–-8266
Some of these articles push back against one of the most popular and influential neurobiological theories (COVIS), explore weaknesses in that approach, and suggest alternatives.
Presented by Josh, Lindsay
Nosofsky, R. M. and Johansen, M. K. (2000). Exemplar-based accounts of “multiple-system” phenomena in perceptual categorization. Psychonomic Bulletin & Review, 7(3):375–402 [
Dunn, J. C., Newell, B. R., and Kalish, M. L. (2012). The effect of feedback delay and feedback type on perceptual category learning: the limits of multiple systems. Journal of experimental psychology. Learning, memory, and cognition, 38(4):840–859
Smith, J. D., Boomer, J., Zakrzewski, A. C., Roeder, J. L., Church, B. A., and Ashby, F. G. (2014). Deferred feedback sharply dissociates implicit and explicit category learning. Psychological Science, 25(2):447–457
Le Pelley, M. E., Newell, B. R., & Nosofsky, R. M. (2019). Deferred Feedback Does Not Dissociate Implicit and Explicit Category-Learning Systems: Commentary on Smith et al. (2014). Psychological Science, X(XX) p. 1-7
Most of the theories we have discussed so far ignore the idea of how preexisting knowledge affects concepts and categories. But everything we learn is part of the broader context of our existing theories about the world. The earlier paper by Lee Brooks explores this within the context of exemplars and instances. Murphy’s and Rehder’s explores this within an organized system of theories and reasoning.
Presented by Ana and Toka * Brooks, L. R. (1978). Nonanalytic Concept Formation and Memory for Instances. In Rosch, E. and Lloyd, B., editors, Cognition and Categorization, pages 3–170. Lawrence Elbaum Associates
Murphy, G. L. and Medin, D. L. (1985). The role of theories in conceptual coherence. Psychological Review,92(3):289–316
Rehder, B. and Murphy, G. L. (2003). A knowledge-resonance (KRES) model of category learning. Psychonomic Bulletin & Review, 10(4):759–784
Rehder, B. (2003). Categorization as causal reasoning. Cognitive Science, 27(5):709–748