Workshop on Cognitive Science Education

An Idiosyncratic View

Janet Kolodner

Georgia Institute of Technology

Introduction

In conjunction with the 1994 Cognitive Science Conference, the Cognitive Science Group at Georgia Tech held a one-day workshop on Cognitive Science Education. We had spent considerable time over the years discussing the ins and outs of putting together an undergraduate degree program in cognitive science appropriate to Georgia Tech. While several years ago, we were in agreement that the field was too immature to support the kind of program that we were comfortable with, several of us were feeling that times were changing -- that cognitive science had matured and that there was much to learn from the experiences of those offering cognitive science degree programs at other institutions. The time seemed right to some of us for reconsidering the creation of an undergraduate cognitive science degree program at Georgia Tech. The one-day workshop was proposed as part of our own preparation for putting a cognitive science program together.

But we knew we were not the only university thinking about cognitive science education. So although we had ulterior motives, we tried to put together a workshop that addressed cognitive science education more generally. Rather than confining discussion to undergraduate programs, we included discussion about graduate programs as well, and rather than discussing only degree programs, we also discussed certificate programs and minors. NSF was kind enough to support the effort with a small grant. The remainder of the cost was borne by a grant from the Georgia Tech Foundation to Georgia Tech's Cognitive Science Program.

The workshop had two parts: a poster session in which the entire cognitive science community was invited to share information on their cognitive science programs and a set of four panels in which issues relevant to cognitive science education were discussed. The first two panels explored the intellectual content of existing cognitive science programs. The first addressed the educational philosophies behind existing cognitive science programs, the programs that ensued, and the strengths and weaknesses of those programs. The second explored cognitive science's core identity -- what does every cognitive scientist need to know? We invited representatives from several ongoing cognitive science undergraduate and graduate degree and certificate programs to talk about their experiences. The third panel looked at the ways in which cognitive science research can inform us about program development. We thought it important to look into how we can transfer what our research has told us about learning to our own educational practices. The fourth panel looked at cultural issues. We felt that creating a culture of collaboration among the faculty was a crucial prerequisite or corequisite to creating an effective interdisciplinary educational program. We invited several people who were instrumental in forming cognitive science communities to discuss their views on community creation.

I'll summarize these different discussions and then turn to some important related issues that arose and were discussed. But let me warn readers first. The title of this piece speaks for itself. My summary is idiosyncratic -- not everybody who attended the workshop will agree with the points I've emphasized. And people might not remember discussions being exactly the way I am reporting them. Why? Speakers on the different panels often addressed issues that were discussed in more detail in later panels. Rather than organizing my review in sequential order, I've built my summary around the issues of each panel, regardless of when they were discussed.

Educational philosophy

It has been my impression that the earliest programs in cognitive science were of two types. One type, mostly undergraduate, produced students who were "jacks of all trades, masters of none." Students took a smattering of courses across the disciplines, but because they were taking mostly introductory courses, they never had the experience of digging deeply into anything. Not only were they not masters of any of the disciplines, they also never became masters of cognitive science. These programs failed to produce the kinds of cognitive science offspring we envisioned. The other type of program, mostly graduate, erred in the other direction, requiring students to become masters of everything. Not many students made it through this second type of program and some programs fizzled out pretty quickly.

But cognitive science education has matured since then, and we are addressing these issues better now. During the second panel, Paul Smolensky pointed out to us that learning is a life-long endeavor -- obvious but often forgotten. The implication for education is that we don't need to teach our students everything there is to know about cognitive science over the course of a four- year undergraduate program or even a longer graduate program. Rather, we need to help them learn some core and prepare them to learn the rest as they need it.

Paul's underlying concern was how do we produce the next generation of cognitive scientists -- the ones who will take cognitive science its next step forward into a unique and identifiable interdisciplinary endeavor? Based on an earlier comment by Angel Cabrera, a graduate student in the audience, he made the analogy to naturalistic language evolution where speakers from a variety of different language backgrounds, when living together in the same community, seem to develop an impoverished language, called a pidgen, that allows them to communicate with each other. The new generation born to the community picks up the pidgen and develops a new language from it, called a creole. Creoles are real languages, as structured and expressive as other languages. One can, however, see the roots, in creoles, of the languages they derived from.

Now for the analogy. We are currently a group of researchers from a variety of different disciplines trying to communicate with each other. We have developed pidgens to allow us to communicate and collaborate. Few of us, however, are native speakers of Cognitive Science. Most of us come first from an associated discipline. The argument based on language theory goes like this: If Cognitive Science is to become an autonomous discipline, with its own language and methods, then we will need to have offspring who are born into Cognitive Science, offspring for whom Cognitive Science is their first language, for whom the natural environment is one with members from the variety of disciplinary communities. The new generation will evolve the many pidgen dialects into a creole, a distinct discipline, with its own methods and issues.

Those of us who have been in cognitive science for some time had hoped that our interdisciplinary collaborations and the pidgen dialects we developed to communicate across the disciplines would evolve into a creole --a distinctive, real, hybrid discipline -- but it hasn't happened yet. Why not, and how can we aim towards a creole? (1) Paul says that perhaps we should look at the language learning literature. Developmentalists who study language learning tell us that there are critical periods during which certain things can be learned, and the child who misses out on being exposed to the right things during those critical periods will never attain native-speaker mastery of a language. Though learning is life-long, the right foundations need to be laid early on to attain mastery later. Continuing the language analogy, Paul suggests that we identify critical periods in the learning of a discipline, identify what learners need to be exposed to during those times in order to master a discipline, and design our educational programs on that model.

Paul suggested that one of the most important things for students to experience early in their intellectual careers (i.e., as undergraduates) is "hard" science (e.g., physics). He wants students to understand what a theory is because he wants us to be aiming toward having hard theories in cognitive science. He suggested that graduate school is a time for learning to think like an X (some kind of cognitive scientist), and that cognitive scientists might need to learn how to think like several kinds of X's. And while factual knowledge can be acquired over a lifetime, he thought that methods needed to be learned and honed under the direction of experts over long periods of time -- one might begin to learn the rudiments of some methods as an undergrad, learn additional methods and learn more about some that are already known as a grad student, and learn still other methods in post-doc years. Why is this so necessary? I think because our methods and the ways in which we use them are the embodiments of our paradigmatic assumptions and intellectual commitments. In a sense, thinking like an X means using the methods of an X to answer the kinds of questions an X finds it important to ask.

Neil Stillings addressed the curriculum and suggested two frameworks for setting up cognitive science curricula:

Have a strong core, required for everyone, with methods courses, skills courses, and the opportunity to do research that ties the disciplines together.
Have a weaker core, fewer required courses, and give the students more options.

The Stanford undergraduate cognitive science degree program, called the Symbolic Systems Program (2), follows the first course, providing students with a wealth of methodology courses and issues courses and much experience doing cognitive science. The Stanford program, as Paul Smolensky suggests, was derived by asking what students will need to master to cross disciplinary boundaries in answering the important questions of cognitive sciences. Stanford focuses its core on the computer and its role in studying mind and intelligence. The core introduces students to philosophical and logical foundations and linguistic theories and techniques, and helps students acquire skill in the theory of computation and manipulation and use of computers. Beyond the core, students can choose to concentrate in one of ten disciplinary or topical areas.

The Hampshire undergraduate degree program follows the second route. If not well managed, students following this second route can turn out like the jack of all trades, masters of none discussed earlier. Hampshire solves this problem by giving close attention to all of its students and providing them with options that allow them to gain deep insights. This requires a lot of faculty time.

Core identity

Of course, in order to put the core of a program together, we must first decide what the core is. Much discussion in the first two panels focused on cognitive science's core identity, the role that the disciplines play, and the extent to which students need to know about the disciplines.

The most entertaining part of this discussion, perhaps, was an interchange between Neil Stillings and Roger Schank. Neil, who was on the core identity panel, came to the workshop wearing two hats. He was an organizer of another cognitive science education workshop that dealt with the core curriculum; he also came to us as a faculty member from Hampshire College, a fairly student-centered liberal arts college. With his workshop organizer hat on, he gave us a long list of topics that needed to be covered in a cognitive science curriculum -- the history of cognitive science, its core methods, its core perspectives, several pieces of core knowledge. Neil talked about the difficulty of covering these topics, given the many points of view on each, and how an interdisciplinary team of teachers was needed for each course. Some in the audience wondered how many years students would have to remain in school to learn all of this and whether, having learned all these facts, they would know how to do anything. Roger asked Neil how much of his list he covered in the curriculum at Hampshire College. Neil's answer: They don't do it that way at Hampshire. At Hampshire, they tend not to have a lot of required courses -- rather, their courses are more student-centered, getting in depth at some things, but sometimes missing breadth.

Paul Thagard suggested that at cognitive science's core is a theory called CRUM -- Computational/Representational Understanding of Mind. We are striving to understand thinking in a computational and representational way -- what are the algorithms that our minds use, what types of knowledge do they require, and how is that knowledge represented? Since CRUM is the child of interdisciplinary thinking, our educational philosophy, he says, should be to (a) teach the basics of CRUM and (b) be interdisciplinary about it.

What are the basics of CRUM? Excitement about the study of mind, an approach to studying the mind that is interdisciplinary, a core that emphasizes mind's computational nature, the realization that the mind is multifarious, and the realization that we don't yet know how it all fits together and what might be missing from our conception. By multifarious, he means that thinking requires many different inferential methods (e.g., logical deduction, schema application, analogical reasoning, imaging).

Later it was suggested that disciplines, in general, have a methodology or small set of methodologies that are central to the way they do things, and there was much discussion about whether cognitive science has such a core methodology. Much of that discussion centered on the many different methodologies that the disciplines bring to cognitive science -- empirical methods, programming, ethnography, neural networks, ... Some folks were adamant about the need to not rule any of those out as we define our core. Roger Schank again hit the nail on the head in this discussion (my opinion, not shared by everyone). He said that the field is trying to be too inclusive, letting anyone in who wants to call themselves a cognitive scientist, that although it isn't politically correct to say it, cognitive science has one core methodology -- computational models of mind. We need to center our educational programs, he says, on this methodology.

Computational and representational models of mind should be at the core, say both Thagard and Schank. Thagard wants to include any discipline that says it is part of cognitive science; Schank wants to make sure we don't lose the emphases on computational and on mind. Are these in opposition? They were presented in opposition at the workshop, but I'm not so sure they contradict each other. Putting computational modelling and representation at the center and focusing on their key issues provides a core that cuts across the disciplines and that uniquely belongs to cognitive science. But a variety of different disciplines have contributed to that core and to the ways we are thinking about the key issues, and their contributions will continue to move that core forward, just as the core ought to inform their investigations. From my own point of view, the contributions and content of any discipline are important to cognitive science to the extent that they provide grist for computationally modelling the mind. Perhaps not a politically correct statement, but not an exclusionary statement either.

The practice of cognitive science education

Our third panel looked at what cognitive science research has to tell us about setting up cognitive science curricula. What do our theories and the things we've learned about learning imply about education? Our two panelists were Roger Schank and Alan Lesgold. Jim Greeno spoke on this topic in an earlier panel, when he discussed the basis for the Stanford Symbolic Systems Program. Alan began by telling us that much cognitive science research shows the importance of group processes to learning. The implication: Make sure our cognitive science programs include many opportunities for collaborative problem solving, collaborative learning, and collaborative assessment.

Roger talked about the role goals play in learning. We learn best when some goal we have gives rise to a need to learn something. He says we should help people learn by putting them in situations in which interesting goals naturally arise. He calls these situations "goal-based scenarios." In a goal-based scenario, one asks the student to carry out some interesting task. In the course of carrying out the task, the need to learn both facts and skills arises. The goal for good education in cognitive science, Roger says, is to develop appropriate goal- based scenarios.

Jim Greeno's comments on the earlier panel mirror the suggestions made by Alan and Roger. Jim presented four principles to follow in educating our students, both at the undergrad and graduate levels:

As much as possible, provide experience doing cognitive science -- in summer internships, senior projects, or as co-ops.
Make things problematic for students to promote reflection. Then help them reflect, both through discussion and by providing appropriate resources. An important implication here is that the curriculum should be organized around problems rather than subjects or disciplines. Rather than addressing learning, for example, ask students to explain how expertise might be acquired in some circumstance. (Roger might amend this by telling us to make it a goal-based scenario, perhaps by asking students to design a system or a set of exercises that would help practicioners of some sort acquire expertise.
Alan would tell us to have them work in groups as they carry out the exercise.) The exploration would touch on issues in learning, memory, understanding, and problem solving; it would address cognitive issues and social issues; and several methodologies might be explored in the process. (In Roger's rendition, not only would these issues and methodologies be explored, but they would be put to use as well. Alan's emphasis on group work would allow the group to tackle a harder problem, and therefore perhaps more interesting and in-depth issues, than one person could do alone.)
Provide students with basic thinking tools that they will be able to employ as they address new problems -- logic, an understanding of principles of computation, computational modelling, empirical methods, and so on.
Prepare students for participation in the community of practice, e.g., by having students organize community activities and having students advise other students.

Roger and Alan were talking about two sides of the same coin. Let's have students learn skills by carrying out an interesting task, says Roger. Let's have them do it by working together, says Alan. Add Jim's comments from earlier in the day, that we should put the students in interesting problematic situations, and we have a strong set of principles for a constructivist approach to cognitive science education.

The culture of cognitive science: The last panel addressed our culture as an interdisciplinary community. How do we get the credit we deserve at tenure time? How can we get collaborations going? What does it take to create a community? The panelists were in agreement. Be aware that interdisciplinary activities can fall through the cracks and make sure they don't. Senior people should educate faculty about the intellectual importance of this interdisciplinary work, making it easier for more junior people to get through. Create opportunities for the community to meet together and work together solving problems. Hiring of interdisciplinary and hybrid types will continue to be a problem in disciplinary cultures. Grasp hard at opportunities for getting around in that culture when those opportunities arise. Most important, allow students to have experiences working in interdisciplinary groups and help them to understand what they are getting out of those encounters.

Other issues

Which still leaves us with the question: What is it we want students to learn? The workshop simply was not specific about this. The closest answer was Paul Thagard's: computational and representational modelling of mind. There was another suggestion (mine), that it is important for students to understand the implications of what they are doing as well, i.e., how might the world benefit from such an understanding? This is important for practical reasons and intellectual ones. Practically, other disciplines are rewarded with funding for basic research because they have shown their usefulness in the world. We want those rewards as well, let's understand what our usefulness is. (Note that this is not a call to work on applications rather than basic research.) Intellectually, exciting challenges are often what keep the intellectual fires burning. Many cognitive scientists have told us they thought they understood cognition until they tried to make sense of what was going on in the world. Ann Brown's design experiments, for example, came from discovering the need to broaden the space of things we could learn experimentally. She thought she understood a lot about learning until she went into a classroom and tried to understand how and what the students were learning. (3) Understanding real world learning situations well enough to understand how we might improve them, understanding enough about the interactions of people and the world to be able to help them interact better, understanding enough about how people address real-world problems to be able to help them solve problems better. These are only some of the ways in which cognitive science can contribute to society. Each provides us with plenty of intellectual and scientific challenges.

What about students outside of cognitive science? What should they know about cog sci? It was suggested that students outside of cognitive science, especially those who will someday be designing artifacts that will interact with people (e.g., engineers, architects), would benefit by understanding a little bit about how people reason. There was agreement that it probably doesn't matter which pieces of cognitive science these students learn. Rather, exposure that will make them sensitive to taking cognitive issues into account as they design is the important thing.

Why didn't we come up with a list? It's hard to make a list that everyone will agree on, not only because we are a young science and not only because we have factions, but because lists tend to focus on facts that need to be learned, while in the end, what's most important is a good set of skills, deep understanding of some basic set of facts, and thinking skills to take us beyond that initial set. Neil Stillings presented a list that was agreed upon by those attending the workshop he held. But when we looked at it with the eyes of educators on, it was simply too big. We may never get agreement on particular facts that students should know. We can agree, however, that understanding, contributing sound implications for, being able to design and/or build, and being able to interpret and evaluate computational and representational models of mind are the important skills we want our students to walk away with. With these as guiding principles, as they say where I come from, "the rest is commentary."

Is cognitive science multidisciplinary or interdisciplinary? The answer to this was a resounding "both." When a new interdisciplinary field comes into being, it tends to look more multi-disciplinary than interdisciplinary. That is, folks from different disciplines may be interested in the same problems, but for the most part, each discipline attacks those problems separately. Often, at this stage, folks from one discipline borrow the methodology of another discipline, sometimes with deep understanding of how to use it, sometimes not. As the field progresses, research endeavors become more collaborative -- people from different fields do research together, each contributing his or her disciplinary expertise, each trying to appreciate what the other has to bring, and each trying to contribute to the work of the other disciplines. With even more maturity, some researchers develop hybrid styles -- new methods and new formulations of issues that belong to no discipline. Hybrids aren't just borrowing the methodologies of another discipline -- they merge the questions and methods of different disciplines such that their questions and methods are no longer identifiable with any particular discipline. More maturity brings more hybrids and more kinds of hybrids and widespread conceptual change. When conceptual change becomes widespread, and when the community agrees on a set of issues and methods, a new "interdiscipline" is born. This is the creolization I discussed earlier.

Doing cognitive science means working on issues using perspectives, approaches, and methodologies drawn from a variety of disciplines. We do this right now in all the ways described above. A project is multidisciplinary when established methods from more than one discipline are used or when methods from one discipline are used to address issues posed by another discipline. A project is interdisciplinary when it uses new methods or approaches that don't belong to a single discipline but were hybridized somewhere along the way. At least some members of these research groups identify more as cognitive scientists than as disciplinarians and have developed their own hybrid ways of constructing and addressing issues. Cognitive science research is currently of both types.

What is the role of the disciplines? This question came up over and over again during the day. There was majority agreement, I think, that given the current state of the field, knowing at least one discipline well is important to being a good cognitive scientist. (4) Some thought it necessary so students could get jobs, some thought it necessary so that cognitive scientists would have some set of methodologies to bring to the table; others thought it necessary so that as we create a new discipline, if we do, it will be well-grounded; some thought that deep grounding in something (anything) is important. I think I agree. But are the disciplines central to cognitive science, or do they provide important foundations? My own view: the disciplines provide important foundations. The disciplines, their methods, and their results, must be studied -- they are the intellectual antecedents of our field. Given the current state of cognitive science, competence in cognitive science requires knowing at least one discipline well -- to do really good cognitive science, we may need to have grounding in several disciplines. But more important than knowing any one or several disciplines is knowing how to use disciplinary knowledge and skills to promote the central goal of cognitive science: an understanding of the processes of cognition and the knowledge representations they use that is so crisp it can be represented in computational models. This doesn't belong to any one discipline, nor can disciplinary study by itself prepare anybody for mastery of cognitive science.

Jobs

The issue of jobs came up several times. If we give degrees in cognitive science, the question goes, are we maybe doing our students a disservice, since nobody out there who's doing hiring knows what cognitive science is? There were several points of view on this. Leila Gleitman of University of Pennsylvania says that their cognitive science program, which requires students to get a degree in a discipline and adds cognitive science on top, educates students for a variety of graduate school programs and a variety of types of jobs. Students graduating with a psychology degree may get a job that would normally go to a psychologist, but they might also get a job that traditionally would have gone to a computer scientist, if they have learned computer science as part of their cognitive science program. Jim Greeno of Stanford, says his students have no trouble getting jobs, that those with undergraduate cognitive science degrees can get jobs as "interpreters" in companies with lots of complex information. Cognitive science graduates are broadly educated and have broad perspectives. They know how to speak the variety of disciplinary languages within a company and can help folks from different disciplines communicate with each other. Alan Lesgold says we shouldn't worry, that the jobs are there, but that we need to help educate those doing the hiring so that they understand what a cognitive scientist can provide.

What does my ideal cognitive science program look like? I don't know if my ideal program is an undergraduate program or a graduate one, if it is a degree program, a minor, or a certificate, but I understand more now, I think, about its goals (both social and intellectual), its central core, the role of the disciplines, and its educational activities, both inside and out of the classroom. Its core will be computational modelling of mind. That doesn't mean that disciplines that don't do computational modelling will be neglected or treated as second class in status. Rather, it helps me to frame studies of the issues in those disciplines and to figure out which pieces of which disciplines should be covered.

What is this "computational modeling of mind" that I want everyone to learn? Cognitive science is after an understanding of human cognition that explains the processes and mechanisms by which the mind carries out cognitive tasks, the mental architectures on which these processes are executed, and the kinds of knowledge these processes rely on. The language of information theory and computation allows us to give a computational account of the "algorithms" of the mind, a system architecture, and a representational theory of the "knowledge" and "concepts" in the mind and of the situations and contexts in which the processes operate (much the same way that mathematics allows us to express physics equations and relationships). Programming a model on the computer requires the implementer to give crisp accounts of processes and knowledge structures. But computational modelling (the methodology) is not the only way to get at computational models (the goal). There are other research methodologies that inform on things that computer implementation cannot address. But it is important for those who will follow other methodologies to understand what a computational model is and the level of crispness that we are seeking.

So I will make sure that everybody in my program has enough experience doing computational modelling to understand the detail needed to do such modeling. This deep understanding will eventually guide everybody's research endeavors, whether they implement models on the computer or not. We will study the issues, answers, and methods used and addressed in the disciplines to gain a better understanding of the mind's algorithms and representations, and we will ask what the disciplines contribute to addressing the issues of cognitive science. When we study linguistics, for example, we will aim to understand what our knowledge of language understanding and use tells us about mind's representations, functions, and relationships between functions. We will study not only language, the artifact, but more importantly, how the mind processes language. What is the step-by-step process by which the mind can take a string of words and figure out what they mean? When we take a philosophical view of what scientists do, we will aim, in our explanations and our analyses, to inform about algorithms for discovery and conceptual change, and representational content and structure that support those algorithms. We will study psychology with a nagging question in the back of our minds: what does this tell us about the architecture, functional components, processes, and representations of reasoning and in memory?

Students in my program will begin with a course in which they learn to ask questions about mind and about computation, and in which they begin to learn where the answers might come from and how to use those answers to put together computational accounts. They will continue with courses in which they learn about computation and computational modelling -- the whole variety of ways of doing it. They will build a series of computational models, each of which will require them to ask different questions about mind and to explore the literatures and methods of a variety of disciplines. They will analyze their models in computational terms to understand where the complexity lies. They will analyze the models in cognitive terms to determine their validity, scope, and the new issues the models generate. They will study other methodologies as well, in an effort to learn how to use a range of empirical, modeling, and other methods to draw implications for and provide constraints on a computational theory of mind.

Students will take a variety of topical courses, each centered on an interesting problem. My memory course, for example, will ask why I can remember so little of my childhood, and why it's so hard to remember all the museums or restaurants I've been to, but easy to remember the visit to the 3-star restaurant in France. We'll address memory issues from a whole range of perspectives -- what do psychological studies say? what does linguistics suggest? what do the several types of computational paradigms contribute? And so on. Project work, group work, and reflection on how we came up with answers or why the answers don't seem complete will all play key roles throughout my curriculum.

I'll assign my students real-world problems to solve. Some problems will be basic research problems; others will ask them to apply what we know about cognition to the development of aiding systems, learning environments, or educational materials or curricula. For example, I may ask students to run experiments, build a program, or use some other methodologies collaboratively to come up with a computational theory of reconstructive memory, of creative problem solving, of understanding science fiction stories, or of design. Or I may ask them to design an interactive system that will help students better learn some physics concept or to design a system to help design practicioners do a better job. These problems will indeed be real-world, collected from sites in the community, and students will have access to those sites and the people in those sites as they work on their project. I'll also ask them to address hard intellectual issues, for example, to articulate their point of view on modularity. The curriculum will be largely student-centered. In the course of solving problems, they will become interested in certain issues, for example, and request that certain speakers be invited to address them.

What about grounding in some discipline? What about doing it in a way that gives the faculty time for other things besides teaching? What about doing it in a way that allows undergraduate students to take courses outside of cognitive science? And so on? No answers yet to those questions. I think we're left with a lot of good ideas -- an underconstrained problem that we still need to operationalize and constrain more. I welcome additional discussion about how to do that.

Is there agreement on all of this? Well,....., no. Many disagree with the notion of computational modeling being central to cognitive science, though I suspect that is because the notion of computational modeling hasn't been articulated well enough yet for everybody to understand what some of us mean by it (perhaps the explanation above will help). There is much disagreement about whether cognitive science will or should creolize and what it will look like when and if it does. I am hopeful that it will, and I see much evidence that it is happening -- not yet among the masses, but there are plenty of hybrid methodologies and approaches emerging. There's a lot of discussion still needed about what belongs in the core, even about what counts as cognitive science. There's disagreement about whether run-of-the-mill disciplinary people should be considered cognitive scientists. There's disagreement about what experiences our students should have and the forms of our programs. There's disagreement about the form cognitive modelling should take. And so on. Will we ever agree? Maybe, maybe not. Do we need to? Not now, I think. Most important are that we create cognitive science educational experiences for our students that give them deep understanding of some things (especially the kinds of issues we address and the form of the answers we are ultimately looking for), that we try the different ways of doing it, and that we talk about it and share our views and experiences and analyses of results. I look forward to continued discussion.

Acknowledgements

Thanks to Mike Byrne, Angel Cabrera, Ashwin Ram, and others, who helped me remember and understand several issues and discussions. Angel helped me understand creolization. Ashwin helped me articulate what a computational model is. I may even have stolen some of the words they used as they reminded me.

(1)Gary Dell gives another, and perhaps the original, view on Cognitive Science and creolization in an article on the Cognitive Science program at University of Illinois in an earlier issue of this newsletter.
(2)Presented in this issue.
(3)Journal of the Learning Sciences, Volume 2, Number 2.
(4)I think many people would also agree that if cognitive science ever creolizes, then the need to master a contributing discipline might vanish. The more contentious issue is whether cognitive science should or ever will creolize.

Back to the Cogsci Home Page.

For more information:

Human contact point