[Please see digital academic workflow video at Screencast of a simple academic workflow. Scroll down just a wee bit to see the whole screen.]

The amount of lost time and lost productivity resulting from inadequate academic workflows affects not only individual researchers but their whole field. While there have not been many studies done on workflows, one survey (University of Minnesota Libraries, 2006) found that 36% of the social science faculty at the University of Minnesota used reference managers compared to only 8% in the humanities. If we can assume that even fewer faculty members are using efficient academic workflows, then the state of research is in need of an upgrade.

The same study (University of Minnesota Libraries, 2006) found that graduate students are 50% more likely than faculty to use reference managers. Thus the concept of an academic apprenticeship (McGowan, 2005) takes on a new meaning, with graduate students able to teach faculty a lesson.

Howison and Goodrun (2005, p. 3) write that the system for managing academic papers “is far from standardized and ranges from manual ‘fumbling’ through to quite sophisticated if users employ citation managers, such as Endnote or the variety of Bibtex management systems. Yet even the most sophisticated users experience difficulty in managing their files.” This, they write, is due not only to industry lagging behind in the area of integrated support for researchers, but due also to the incapacity of PDFs and most other documents to allow metadata to be included within the document itself. When we download a paper, we must download the metadata separately, whereas when we download music, the metadata is included in the music file. This kind of inefficiency, in addition to a lack of available commercial workflows for researchers, has led many of us to create our own workflows, with varied results.

This search for a simple yet effective digital academic workflow is not new. Hundreds of researchers (see e.g. Judson, 2008; Roberts, 2009) have written on the web about their attempts to set up and use a system that will:

1)    Find new articles of interest quickly and download them simply with their metadata

2)    Enable them to annotate, mark up, and take notes digitally, with no need for a paper copy of their documents

3)    Keep all their papers and other research documents in one place

4)    Provide a way to search through those documents in an efficient way

I currently use a very simple system of Google Scholar, BibDesk, Skim (all free) and Spotlight (which comes on all Macs) which is a very limited version of a much more powerful workflow called researchr (Haklev, n.p.) which was developed by Stian Haklev at OISE, University of Toronto. One day I hope to be running researchr (we have experienced some technical difficulties in running my version of researchr reliably (update October 2012 researchr is now working beautifully, thank you Stian!), but in the meantime my current very basic workflow does all of the above simply and efficiently. One advantage of my workflow is that it is so basic that I can run it by myself, with no need for computer expertise, and so can anyone else. It should be noted here that this workflow will only work on a Mac as BibDesk, Skim, and Spotlight are applications that work on Macs only, but similar workflows are possible on any computer OS.

Google Scholar

I chose Google Scholar to do my online searching for articles because it was faster and more efficient than using the University of Toronto’s (where I am a graduate student) online library (Sept. 2011 update: The U/T library system has been completely revamped and it now far faster and more efficient than it used to be). Google Scholar probably has the widest coverage of all online library search engines and provides information on the number of times an article has been cited, which most other libraries don’t (Hull, Pettifer & Kell, 2008). Google Scholar “aims to rank documents the way researchers do, weighing the full text of each document, where it was published, who it was written by, as well as how often and how recently it has been cited in other scholarly literature” (Scholar.google.com, n.p.). The efficiency of looking up a topic and finding the most relevant articles ranked top down is currently unmatched by other online library search engines. However, according to Jasco (2005), a big problem with Google Scholar is that millions of articles have not yet been indexed, so coverage is far from complete. Despite this, I find Google Scholar very useful and will likely use it for many years, along with the University of Toronto online library for backup.

In fact, with Google Scholar I can set my preferences (click on the cog-like spoked wheel to the right of “sign in” in the upper right hand corner of any Scholar page) for the University of Toronto’s Get It! service (any university student should be able to use their own library membership to do the same, just type in the name of your university and click search in the Library links section), which means that as soon as I am logged into the U/T library system, all articles that are available from the U/T library will be identified by a “Get it! U of Toronto” link. I just click on the link and I’m taken to the databases where the article is available. So the U/T online library is nicely synched to Google Scholar now, which is fantastic because the U/T library has access to many articles that are not available online.


BibDesk is a free and open-source reference manager that is under active development by a small group of volunteers (SourceForge.net, n.d.). With BibDesk, your collection of PDFs and other documents is organized by author name, document title, date of publication, etc. so that finding documents is easy. BibDesk automatically imports keywords when it imports the citation data, but you can add your own keywords as well, which is something that I have started to do with painstaking commitment. By adding my own keywords, I can find information I am looking for much faster (please see attached video for a demonstration).

BibDesk can search titles but cannot search the PDFs themselves, which would be a very useful feature, but this is easily done on a Mac using Spotlight, Apple’s powerful search tool. BibDesk also allows automatic generation of references for Latex users, and there are programs (BibDeskToWord, BibFuse) that are supposed to allow for BibDesk to Word full reference generation as well.

I am relatively happy with BibDesk as finally all my articles are in one place and bedlam has been defeated (Judson, 2008). But I could get neither the BibDeskToWord program nor BibFuse to work. So at present my workflow cannot generate instant references like a program like EndNote (over $100 for the student edition) can. As well, Zotero (free) seems very interesting and useful, so I will check out Zotero and may possibly switch from BibDesk to Zotero in the near future. This workflow is an evolving workflow… which is as it should be. There will always be new and more efficient programs appearing on the market, and I’m keeping my options open.

Image 1. BibDesk Screenshot


Skim is a PDF reader that allows highlighting and annotation. The most useful Skim tool for me is the Note tool, which allows me to write my own notes within the document. In the screenshot below, I have highlighted text referring to a topic I am very interested in – mutual aid – and I have made mutual aid the title of the note. By tagging text in this way, when I do a search using Spotlight, this reference to mutual aid will turn up (even though the words “mutual aid” do not appear in the text) and thus I can find this and other important references to mutual aid quickly and easily. This feature of Skim has ended a problem I’d had for years with not being able to tag text and search those tags.

Image 2. Skim Screenshot

Spotlight is an incredibly powerful search tool that comes installed on all Macs and allows for very efficient searching of documents, webpages, and more. The handy combination of Google Scholar, BibDesk, Skim, and Spotlight has been a huge relief for me, as my workflow used to be terribly disorganized and now it is streamlined, smooth, and highly functional, if basic. I will be developing this workflow in the coming years and I hope that it will continually be getting more efficient and powerful. Please see my basic screencast (link at the top of this page) for a demonstration of this workflow in action.

In closing, I have not yet mentioned anything about Web 2.0 and its capacity for enabling collaboration between researchers. Blogs, wikis, social bookmarking, and other Web 2.0 tools are changing the way we interact, and making international collaboration possible. Once we get our own personal workflows in order, the time saved and the efficiency gained will lead to better and more productive personal research, and international networking with researchers in our fields is leading us to Research 2.0 – networked, collaborative research on a global scale.

The future is looking good for researchers! And the first step is finding a great academic workflow.


Haklev, S. (2011). researchr. Available at http://reganmian.net/wiki/researchr:start

Hull, D., Pettifer, S. & Kell, D. (2008). Defrosting the Digital Library: Bibliographic Tools for the Next Generation Web. PLoS Comput Biol 4(10): e1000204. doi:10.1371/journal.pcbi.1000204

Judson, O. (2008). Defeating bedlam. The New York Times [online version], December 16, 2008. Available at http://opinionator.blogs.nytimes.com/2008/12/16/defeating-bedlam/

Muldrow, J. & Yoder, S. (2009). Out of cite! How reference managers are taking research to the next level. PS: Political Science and Politics 42(1), pp. 167-172.

Roberts, B. (2009). Get started handling academic citations like a pro. Available at http://www.soulphysics.org/2009/12/get-started-handling-academic-citations.html

Scholar.google.com (2011). About Google Scholar. Available at http://scholar.google.com/intl/en/scholar/about.html

University of Minnesota Libraries (2006). A Multi-Dimensional Framework for Academic Support. University of Minnesota. Available at http://www.lib.umn.edu/about/mellon/docs.phtml


In Drive (2009), Daniel H. Pink tells us that we are not motivated by external rewards (like grades, or money) but by a desire for autonomy, mastery, and purpose. Watch the RSA Animate video of his TED talk. It is mastery now that I want to focus on because I have just watched an exceptional TED talk by Salman Khan of Khan Academy.


Khan tells us that a major problem we have in school is that there is no goal of mastery. Students study a topic, a concept, let’s say basic division. They get lectures on it, read a few pages in a textbook, do some homework, do some exercises in class maybe, some more homework, and then a test. No matter how well or how poorly they do on the test, in the following class they will be starting a new topic, let’s say fractions. Now, if you didn’t do too well with division, you’re not likely going to do too well with fractions either. And this just snowballs with time until you get math-phobic kids who maybe wouldn’t have become math-phobic if they had just learned division (or any basic concept) thoroughly the first time.

Image from http://www.crunchyroll.com/group/cr_acadomy_2

Mastery is missing in school, and according to Pink, it’s one of the three things that most motivate all of us, including students. But how can we achieve mastery in a class of 30 or more students? There is a curriculum, there are tests, there is no time for mastery. But Khan says there is. He’s made short videos on 2400 topics from Addition 1 (2+5=?) to Green’s Theorem (is Green’s Theorem hard? I don’t know, I’m not there yet…).

Q: What’s so great about videos?

A: Videos provide time on task. Students can watch the video as many times as they need; rewind, rewatch, review, until they get it right. Videos provide the opportunity for mastery.

Khan suggests watching a video for homework. I’d add that you need to do some exercises right after the video to consolidate what you’ve just learned. If you can’t get the exercises right, watch the video again. Repeat as necessary. Get them all right.

In class, the next day, you do some exercises, and if you get them all right, maybe you can help someone else who is having problems with the same concept. The teacher, who doesn’t have to lecture anymore, can use 100% of her class time in 1-on-1 time with students who need help (compare this to the usual 5% Khan says is standard for most classes). Students can work at their own pace, at exactly their level. Khan found that when students are given time on their own to master a concept, they all eventually will, even if it takes a long time, and they later on catch up to the students who caught on faster. Imagine a class where you score 100% on every test before moving on to the next skill, topic, or concept. That’s mastery.

One of the most satisfying feelings I have ever experienced is reading something very difficult and very interesting. It takes a lot of time to get thru the reading, it’s a real commitment, and sometimes you have to re-read passages and take notes, it feels like your brain is working overtime. But when you get it, you really get it, and what a great feeling it is when you know that you have mastered a difficult concept. Those are my favorite articles and books, the ones that challenge me and force my brain to crunch thru hard topics.

I’ve signed up at Khan’s Academy (URL) and there are two reasons that I did this. One is for my own mastery. I’ve always loved math, and science, and I have been away from them for far too long. The other is my son, who, like me, loves math and science. At Khan’s Academy, I can be his coach, and check his progress, and help him when the going gets tough.

And me, I’ll be looking forward to some tough going of my own at a higher level. That’s when the challenge sets in, that’s when learning becomes its own reward.

I can’t wait to get started… 🙂


Pink, D.H. (2009). Drive. New York: Riverhead Books.

In an OISE grad course called Computers in the Curriculum (CTL1606) I was discussion leader last week and decided to use etherpads as a medium for a collaborative discussion to be followed by a collaborative writing session. Here is what happened and my thoughts on it.

1) Scheduling is a problem even when you use a tool like WhenIsGood. Of the 4 students in my group, 2 signed up right away and could have an etherpad meeting with no problems. 2 others signed up later and chose more limited times, one student chose one hour only, and it was that same evening! So there was no time for me to set it up. So that student ended up not participating in the exercise.

To set this up effectively, it would have to be done more in advance, and students could be requested to choose at least 3 times when they were free. And although WhenIsGood is a really great tool, just posting some times and having students write their names under the times would also work, and might work better in that when a student has posted their name, they feel some obligation to keep that time free, which may not be the case when they have posted many free times on WhenIsGood.

Point 1: Scheduling is difficult!

Solution: If I did it again, I’d ask students to organize themselves and choose times together.

2) Even when you type clear instructions, students may have their own ideas about how to do something. I wonder how much this has to do with the fact that it was grad students I was asking to participate in the exercise, and they are often in a position to lead, not follow, as many of them are teachers already.

For example, I posted the following instructions in the etherpad. “Please use the chat box on the lower left to discuss your ideas on these two questions. Please use this white pad to write the document you’ll post later as a note in our class forum. So based on what you discuss in the chat, you write a joint document here based on your chat that answers the questions.” But in both of the etherpad meetings, students didn’t discuss ideas first, they started typing on the whitepad right away. When I inquired about why they had skipped that part, it turned out that one of the students (who participated in both of the etherpad meetings that took place) had used etherpads before in a class but they had only used the whitepad part, not the chat. So that was how she handled this exercise too.

Point 2: Even when directions are clear, students may not follow them in the way you expect them too.

Solution: If there were a next time, I’d join the exercise! This way students would do it the way it was planned and maybe they’d realize the value of first discussing an idea and then writing about it.

3) There were some stability issues with all of the etherpads I used. We used PiratePad for collaboration between discussion leaders prior to Week 4, and our pad that we made is not longer accessible. Likewise of the 2 ietherpads used, one is available now for viewing and one is not. We used TitanPad only once but it’s still there! But one student said there were lag times in the texting. “[T]here were many times I had to delete a response because by the time I typed it into the message box, the topic had changed before I could press return.”

Point: Etherpads are not very stable. This is a big problem.

Solution: Use Google docs next time. (Now I am wondering if unstable etherpads are a conspiracy to force people to use Google docs! It was Google who released the code for etherpads, they must have known they were unstable!)

4. Even tho the students in my group didn’t use the chat, they still had good experiences and wrote collaborative texts. So the exercise was a success from this viewpoint. When I read their co-constructed text, it flows well, with the ideas of two people mixing together*. This is collaboration! So just when students don’t do things the way you were expecting, it’s the end results that matter most.

Point: Students don’t always do what you are expecting them to do.

Solution: Go with the flow! Give over some control of the exercise to the students. As long as the outcome is good, it doesn’t matter if it’s done in a different way from what you were expecting.

*Actually on one of the etherpads, one student posted and another student posted after her. The first student didn’t post onto the second student’s writing. This is not very collaborative. It may be the result of a misunderstanding on the part of the first student as to what collaboration is. But on the second etherpad, the two texts are interchanged, and both students are posting onto what each other had written.

5. Finally, collaboration occurs on many levels. In our discussion forum, we collaborate by posting our ideas and commenting on others’. In the etherpad exercises, students collaborated by co-constructing a written text together. In an online chat, ideas flow between people, and negotiation of meaning is done in realtime (like When ideas have sex, TED talk). My friend Stian would say that these are granularities of collaboration. And I would agree with that. Collaboration happens on many levels! I think that collaboration on all of these levels would be a good experience for anyone hoping to use collaborative exercises in their own classes.

Final Comment: Well, in taking risks, we learn! That was one of the ideas we discussed this week. I took a risk by using a tool that looked promising, but I didn’t know that stability would be such a problem, or scheduling. Now I know! And I’ll be chatting and using etherpads again. I may use a Google Doc plus chat as well to collaborate.

Update (August 2012): I no longer use etherpads. I use Google docs only because of their stability and Google docs’ many useful tools. More on Google docs later.

(Image from Collaboration Techniques) 1

Situating this post: I am on a flight from Hong Kong to Tokyo, listening to Astor Piazzolla.

The CSCL 2011 conference (see 2 below for a brief description of CSCL) was fantastic, with tech luminaries for keynote speakers (Erik Duval, Roy Pea, and Ed Chi, now a research scientist at Google) and an overwhelming array of presentations at a very high level of research that really boggled my mind.

One interesting part of this particular conference is that presenters were told to make a 5 or 10 minute presentation (depending on whether it was a short paper presentation or a long paper) and then we all broke up into smaller groups to talk with whichever presenter had caught our interest the most. This put a lot of pressure on the presenters as presentations delivered in the usual type of conference are usually 30 or 40 minutes long (in my experience), so presenters had to really compact their description and their findings into a tiny slot. But the collaboration afterwards was fantastic! In fact, I think it benefits the presenters themselves at least as much as the participants, as they got so much feedback (mostly positive, but some constructive negative feedback too) on their research.

In a typical conference, there might be a few questions after a presentation, but little actual dialogue, and I’ve often seen Q&A sessions where the question asked was not really answered by the presenter. In our collaborative discussions, topics of interest were discussed in depth, and I have to say that although I wish the presentations themselves could have been longer (they really were squashed 3), this represents a new and very participatory way for researchers to engage in dialogue. I learned at least as much from the collaborative after-sessions as I did from the presentations. So this heightened my awareness of the importance of collaboration in learning, which can be done in both online and f2f environments. My own personal experience in collaboration has been minimal. This needs to change (I’m working on ideas on how to change it).

This brings up the point that we as teachers need to be involved in collaboration before we can understand the value of collaboration and share it with our students. In a recent study by Austin, Smyth, Rickard, Quirk-Bolt & Metcalf (2010), researchers found that the level of engagement of teachers in collaborative learning had a positive effect on the success of collaboration of their students. This is very important! The less engaged in collaboration the teacher was, the less successful was the collaboration of her students. Therefore if we hope to reap the benefits of collaborative learning for our students, we had better get collaborating ourselves! An interesting parallel is with technology. It seems likely that the better our own personal experiences with technology are, the more likely we will be to use them in class successfully. So let’s get playing with a variety of technologies, and let’s get collaborating! Let’s collaborate on our learning experiences with technology! 🙂


Austin, R., Smyth, J., Rickard, A., Quirk-Bolt, N., & Metcalfe, N. (2010).  Collaborative digital learning in schools: teacher perception and effectiveness.  Technology, Pedagogy and Education, 19(3), 327-343.

1 I read recently that the use of images helps us to better internalize and retain information so I am going to be using images in posts whenever possible. In the physical version of Wired (July, 2011), I read, “When information is presented verbally, a person will retain about 10% of the message 3 days later; add a picture and retention soars to 65% (John Medina, cited by Carmine Gallo, p. 108). I wonder what the retention rate for written text is?

2 Computer-Supported Collaborative Learning, thus there was a focus on collaborative learning (CL) and how computers and technologies support and enhance CL.

3 All papers were given to us in pdf format on a flash drive on the first day of the conference, so even with the very short presentation time, we could get the whole paper that very evening if we wanted to (this was also a huge improvement over waiting for months for the proceedings to eventually be published.)

In a recent blog post (Ode to Piaget), I tried to explain the source of my frustration with a quote from Stahl, Koschmann & Suthers (2006, p. 9), who write, “CSCL locates learning in meaning negotiation carried out in the social world rather than in individuals’ heads.” I wrote that in the cognitive constructivist view of learning, the locus of cognition is indeed in individuals’ heads and to ignore this is unlikely to produce anything of value in understanding how any kind of cognition works, group cognition included. Individuals have brains and cognitive models which represent and structure our individual knowledge. These structures are constantly changing as we learn new things. Assimilation and accommodation are constantly occurring, and any theory of cognition, including group cognition, must acknowledge these cognitive processes. However, events are also occurring at a group level that cannot be explained by investigation at the individual level (Gureckis & Goldstone, 2006; Thompson & Fine, 1999). It seems that there are two levels of activity, one at the individual level, and one at the group level. Doesn’t an account of group cognition need to take both of these levels of activity into account?

I needed to do some more reading on these topics, as I was finding Stahl et al (2006) were not explaining group cognition in a way that was useful to me. I found several articles that have been of immense help, and so this blog post is to summarize the main points of one main article (I had hoped to include more in this post, but instead will do so in later posts) and integrate them into my own thinking.

The first and extremely helpful and insightful paper has clarified the issue of cognitive and socio-cultural perspectives on learning in an objective way. The article titled, “Reconsidering group cognition: From conceptual confusion to a boundary area between cognitive and socio-cultural perspectives?” by Akkerman, Van den Bossche, Admiraal, Gijselaers, Segers, Simons & Kirschner (2007) is a review of all the major studies of empirical studies of group cognition. The articles they review run from pre-1994 to 2002, so more recent studies of group cognition have not been included, but the 22 studies out of a total of 167 represent well the body of research that has been done on group cognition until 2002.

Akkerman et al (2007, p. 42) explain in detail the two “socio-genic views” (Valsiner & Van der Veer, 2000): cognitive and socio-cultural. I will include only a brief extract here:

“[T]here is the axiomatic preference for fusion (of person and the social environment) or inclusive separation (i.e., the person is viewed as distinguished from the environment, yet interdependent with it) bases for socio-genetic models” (Valsiner & Van der Veer, 2000, p. 6). Whereas the former socio-genetic view perceives the person and the social environment as one whole and relies on terms like participation and adaptation, the latter socio-genetic view perceives the person and the social environment as separate units that are related to each other and uses terms such as internalization and externalization.

The former, fusion, represents the socio-cultural perspective on learning. This has also been associated with the participation metaphor (Sfard, 1998) and a focus on learning within social interaction. The inclusive separation model represents the cognitive perspective, aligned with the acquisition metaphor (Sfard, 1998), with its focus on learning within the mind. Both of these perspectives are represented well in the empirical studies of group cognition reviewed by Akkerman et al (2007). In their review, 11 studies represented cognitive perspectives and 5 studies represented socio-cultural perspectives. 4 represented both, and 3 of these are considered to be “boundary crossing studies” as they “were able to pursue a mixed discourse of both cognitive as well as socio-cultural perspectives on mind throughout the whole study” (p. 48).

Here are summaries of the reviews made by Akkerman et al (2007). Please see tables 1 and 4 below for a more detailed perspective.

Cognitive perspectives: Studies reviewed by Akkerman et al (2007) refer to the knowledge of the individual team members as it is constructed in mental models. In this perspective, it is assumed that individual people “organize knowledge into structured, meaningful patterns and store them in their memory” (p. 45). Group cognition is conceptualized by these cognitive studies as “a similarity or overlap between individual mental structures” (p. 49). Group cognition occurs at the group level but relies heavily on the individual as entity. There is a focus on “similarity” in the methodologies of all the cognitive studies, “with the underlying hypothesis that the more similarity in the identified knowledge structures of the individuals, the better the team functions” (p. 49).1

Socio-cultural perspectives: All studies focused solely on group cognition and did not determine the cognition of individual members. Instead they focused on the whole of contributions made during the interactions. The question was “how group participants acted on, and thereby defined the specific domain or object of activity” (Akkerman et al, 2007, p. 47). Socio-cultural perspectives as described by Akkerman et al (2007, p. 50) perceive group cognition as something “constituted by the group as an entity in itself. As such, cognition resides in the active mind, as a phenomenon situated in the group interaction. The concept [of] group cognition is then defined as a process of coordination of participants’ contributions in joint activity (Matusov, 1996)”. In this view, Akkerman et al (2007, p. 50) write, “using similar modes of thinking is argued to lead to reproductive processes, with the danger of group think narrowness, while disruptions resulting from different views of participants and socio-cultural subgroups are perceived as offering potential for productive, creative processes and group development2 (Homan, 2001; Matusov, 1996)”.

In addition to cognitive and socio-cultural perspectives, Akkerman et al (2007) discovered a third group of studies which they label “boundary crossing” literature because of the integration of both cognitive and socio-cultural perspectives throughout the study. They focus both on “stable mental model forms with information stored in memory and mental model states that are active and situated in the group collaboration” (p. 51). In these studies, shared information is considered to be distributed. “What they were able to do was to decompose subsystems within the whole complex system of a team and reveal their inherent properties, while simultaneously revealing the intrinsic nature of the complex system itself, separately from the subsystems. On the one hand, they identified individual subjectivities in terms of stable cognitive maps, and on the other hand they identified group processes in terms of the individual mental states situated in the interaction. So in a very precise and clear way they integrated in their focus the intrinsic nature of decomposable subsystems and the intrinsic nature of the whole system.” (Akkerman et al, 2007, p. 53).

These boundary crossing studies would seem to be what I have been looking for: an integrated system that allows for two intersecting levels of activity, one at the individual level and one at the group level. However, while these studies do incorporate cognitive perspectives, they do not fully represent the socio-cultural perspective because they treat cognition as decomposable. Cognition in these studies is treated as happening at two levels and is seen as separable and analyzable at the two levels, whereas this kind of separation or decomposing is anathema to the socio-cultural perspective. The socio-cultural perspective does recognize the individual mind, but sees it as non-extractable from its situated perspective. “Socio-cultural perspectives do not deny the existence of an individuals’ mind, nor its agency, but they understand this mind as situated in the participation processes in systems of socially organized activity that are themselves evolving” (Akkerman et al, 2007, p. 43).

When we look closer at the ontological assumptions underlying the two perspectives, we get to the root of the problem. According to Packer and Goicoechea (2000, cited in Akkerman, 2007, p. 54), “Cognitive perspectives imply a dualist ontological approach, in which construction is viewed only as a cognitive activity in which subjectivity structures and shapes data that comes from a distinct and separate objective world. As opposed to that, socio-cultural perspectives bring forward a non-dualistic ontological approach, in which subjectivities and the objects themselves are constructed and mutually define each other…. Cognitive perspectives seem to assume that things (e.g., individuals or mental models) can exist independently, although it can change by its relations. Socio-cultural perspectives seem to assume an ontology in which things only exist in relation to other things (individuals are participants; minds are situated in social action).”

Akkerman et al (2007) conclude their review with their vital question, “Is it advisable to strive for a coherent theory on group cognition, or is it better to have the two perspectives, as the differences are insurmountable?” One answer to this question is that the two perspectives could possibly be reconciled through dialogical engagement. In this view, each of the two perspectives reaches towards the other and tries to incorporate its views into its own views. This was done to some extent by the boundary-crossing studies identified in this review. As well, “Valsiner and Van der Veer (2000) and Hermans and Kempen (1993) … [consider] the individual as a dialogical system by itself. Cognition is then defined as an individual property, but the individual itself is an inherently social entity, constituted through its social relations with others” (Akkerman et al, 2007, p. 55). While these studies represent attempts to reconcile the two perspectives, the underlying problems of ontology and locus of learning remain unsolved.

The other two possible answers are essentially yes, they can be reconciled, and no, they cannot, and it is this latter answer that Akkerman et al agree with, and so do I. The two perspectives cannot be reconciled because of the very different ontological views and where each places the locus of learning. However, I have not yet read Valsiner and Van der Veer (2000) nor Hermans and Kempen (1993), and to me these offer the most hope of reconciliation. I will read these in the near future and hope for new insight. But the central questions keep returning. Can we ignore the disparate ontologies underlying the two perspectives? How can we combine two perspectives into one when they are so fundamentally different?

Finally, we can ask here the critical socio-genetic question posed by Valsiner and Van der Veer (2000, p. 6), and asked by Akkerman et al as well (p. 56): “How to construe persons as being social without abandoning their obvious personal autonomy, separateness from any social unit (group, crowd, community), while being members of such units?”

So now my frustration with the statement made by Stahl, Koschmann & Suthers (2006, p. 9), “CSCL locates learning in meaning negotiation carried out in the social world rather than in individuals’ heads” has been explained. I am a cognitivist who situates learning within the heads of individuals, who are themselves situated in society and various groups. I see cognition happening at two levels, one at the individual level and one at the group level. I see these as necessarily connected but separable for the purposes of analyzing what is happening at each level. In contrast, Stahl, Koschmann & Suthers (2006) have a socio-cultural perspective and as such see cognition as inseparable from the context in which it is happening. Our fundamental views of group cognition differ, and now at least I understand the differences more clearly.

I hope this summary is useful to anyone who is trying to understand group cognition and the various views of it. If you have any articles to share that would help to illuminate this discussion further, please do post a comment below. This is a mighty adventure for me and intensely interesting. Here we are in late spring of 2011, considering whether the two main perspectives of learning can be reconciled into one theory. In my previous post, I wrote that cognitive constructivism and social constructivism are mutually compatible. I wonder now if I will have to revisit that post and do some rethinking. I may also have to do some rethinking of my future in CSCL research if indeed CSCL requires a socio-cultural perspective, as Stahl, Koschmann & Suthers (2006) indicate above.

In any case, thank you for reading and please do leave a comment below if this kind of exploration is interesting to you.


Akkerman, S., Van den Bossche, P., Admiraal, W., Gijselaers, W., Segers, M., Simons, R.-J., & Kirschner (2007). Reconsidering group cognition: From conceptual confusion to a boundary area between cognitive and socio-cultural perspectives? Educational Research Review, 2(1), 39-63. Available at http://igitur-archive.library.uu.nl/ivlos/2007-0507-201709/akkerman%20-%20reconsidering%20group%20cognition.pdf

Banks, A. P., & Millward, L. J. (2000). Running shared mental models as a distributed cognitive process. British Journal of Psychology, 91, 513–531.

Derry, S. J., DuRussel, L. A., & O’Donnell, A. M. (1998). Individual and distributed cognitions in interdisciplinary teamwork: A developing case study and emerging theory. Educational Psychology Review, 10, 25–56. Available at http://www.wceruw.org/archive/nise/Publications/Research_Monographs/DERRYDUR/DerryDurALL.pdf

Gureckis, T.M., and Goldstone, R.L. (2006). Thinking in Groups. Pragmatics and Cognition, 14: 293–311. Reprinted as Gureckis, T.M. and Goldstone, R.L. (2008) Thinking in Groups. In Cognition Distributed: How Cognitive Technology Extends our Minds, Edited by Dror, I.E. and Harnad, S., John Benjamins Publishing Company.

Hermans, H. J. M., & Kempen, H. J. G. (1993). The dialogical self: Meaning as movement. San Diego, CA: Academic Press.

Homan, T. (2001). Teamleren. Theorie en facilitatie. [Teamlearning. Theory and facilitation]. Schoonhoven, the Netherlands: Academic Services.

Matusov, E. (1996). Intersubjectivity without agreement. Mind, Culture and Activity, 3, 25–45.

Packer, M. J., & Goicoechea, J. (2000). Sociocultural and constructivist theories of learning: Ontology, not just epistemology. Educational Psychologist, 35, 227–241. Available at http://compsci.duq.edu/~packer/Pubs/PDFs/2000%20EdPsy.pdf

Stahl, G., Koschmann, T., & Suthers, D. (2006). Computer-supported collaborative learning. In R. K. Sawyer (Ed.), Cambridge handbook of the learning sciences (pp. 409–426). Cambridge, UK: Cambridge University Press.

Sfard, A. (1998). On two metaphors for learning and the danger of choosing just one. Educational Researcher, 27, 4–13.

Thompson, L., & Fine, G. A. (1999). Socially shared cognition, affect, and behavior: A review and integration. Personality and Social Psychology Review, 3, 278–302.

Valsiner, J., & Van der Veer, R. (2000). The social mind. Construction of the idea. Cambridge, UK: Cambridge University Press.

Yoo, Y., & Kanawattanachai, P. (2001). Developments of transactive memory systems and collective mind in virtual teams. The International Journal of Organizational Analysis, 9, 187–208. Available at http://iims.uthscsa.edu/sites/iims-drupal/files/Virtual-Teams-3-Other.pdf

1 This hypothesis contrasts with Gureckis and Goldstone (2006) who found that less information-sharing benefited the group in small world networks.

2 This is more in line with Gureckis and Goldstone (2006), and my own way of thinking. Diversity within groups is an asset. I had planned to incorporate Gureckis and Goldstone (2006) into this summary but was unable to because of number of issues raised by Akkerman et al. I’ll return to Gureckis and Goldstone in the near future.

3 Please note that I have read only Akkerman et al (2007), Gureckis and Goldstone (2006), and Stahl, Koschmann, & Suthers (2006) thus far. All other references were referred to within the text of Akkerman et al (2007).

Highly condensed, with all the main points, and graphics too!

Just below is a summary of Zhang, J., Scardamalia, S., Reeve, R., & Messina, R. (2009). Designs for Collective Cognitive Responsibility in Knowledge-Building CommunitiesJournal of the Learning Sciences (18), 1, 7-44


Stahl, Koschmann & Suthers (2006, p. 9) write, “CSCL locates learning in meaning negotiation carried out in the social world rather than in individuals’ heads”. I think that this statement is… misguided. If learning is happening anywhere, it is happening within individuals’ heads. Individuals are definitely a part of the social world, but to point out the social aspect of learning at the expense of individual learning is nothing short of a dismissal of constructivism.

There are two main forms of constructivism, and they are mutually compatible. (Note: I wrote this post before reading Akkerman et al (2007) and that has changed my way of thinking somewhat. For an update, please read this post Cognitive and socio-cultural perspectives on group cognition: Can the two be reconciled?)

The first to appear was cognitive constructivism, as theorized by Jean Piaget1. Piaget was the first constructivist; he came up with the idea of constructivism, and he named it and described it.  At the time it was simply “constructivism” and Piaget said, “I am a constructivist. I think that knowledge is a matter of constant, new construction, by its interaction with reality, and that it is not pre-formed. There is a continuous creativity” (Piaget, n.d., in Papert, 2000, p. 2). According to Piaget, the two processes that make up the construction of knowledge are assimilation and accommodation. Assimilation is the absorption of new information and experiences by one’s existing cognitive structures. In this case, the new knowledge “fits” the existing structures so there is no need for modification of those structures. However, when the new knowledge does not “fit” the existing structures, the structures themselves can be altered to accommodate the new information. (Another possibility is that the new information will be ignored.) The final result of assimilation and accommodation is equilibrium, which lasts until we again experience cognitive conflict. According to Piaget, this system of assimilation and accommodation explains human learning from infant to adult.

We may ask where the new knowledge comes from. It comes from others – thru books, thru interactions with other people, thru searches on the Internet. It comes from interacting with society, and while Piaget believed the locus of construction to be in the individual, Vygotsky believed it was in society. Vygotskian constructivism is therefore known as sociocultural constructivism.

What do we find when we examine the two theories? The most crucial difference comes in the differences in weight that the two theorists placed on the role of society in learning. Piaget’s model of the social aspects of development link successful social experience with intellectual development, and referred to the importance of effective collaboration between pupils and the influence of culture and communication in learning (Smith, 1989, cited in Wood & Bennett, 1997). However, Piagetian constructivists (cognitive constructivists) focus on the individual as the cognizing subject. Individuals actively construct their own knowledge on the basis of new experiences interacting with existing cognitive structures.

For Vygotskian constructivists, the social world has primacy over the individual. Social constructivists see society as the bearer of cultural heritage without which the development of mind is impossible (Cole & Wertsch, 2004). Vygotsky saw learning as necessarily taking place within society and puts the emphasis on social interaction as the means by which learners construct knowledge.

However, both theorists did focus on the position of the individual learning in society. Piaget saw construction as taking place within the individual, who later tests his knowledge for viability by interacting with others. For Vygotsky, it occurs in the reverse order – the learning itself takes place in a social context and is then internalized by the individual (Dockett & Perry, 1996).

How, then, are we to resolve the issue of whether it is society or the individual in which the locus of learning resides? I say that it resides in the individual’s head, and that we as individuals are a part of society, and the societal impact on our learning cannot be ignored. Neither can we ignore the fact that when we walk away from a discussion from which we have learned something, either by assimilation or by accommodation, that the learning we have accomplished resides in our head. The constant comparing and contrasting of new information with our own knowledge takes place in our head. While the construction of the meaning took place with another individual, and may not have occurred without that interaction, the actual change that takes place is in the head of the individual.

So I’d like to point out that ignoring the construction of meaning in the heads of individuals is about as blind as ignoring the fact that one does a great deal of one’s learning with others. I see them as two sides of the same coin; the individual in society. You cannot take the individual out of society, and you cannot take learning out of the head of the individual.

So while we in CSCL are exploring how individuals learn in groups, I think that we must remember that groups are indeed made of individuals, and that while the construction of knowledge (i.e. learning) takes place because of and during interactions with others, that learning is quite definitely taking place within the head of the individual. We must look at both aspects of learning: the individual and the social. Social systems are made of nodes connected to each other. We cannot focus on the connections at the expense of the nodes; they are integral, and cannot be separated from each other.


Cole, M. & Wertsch (2004). Beyond the Individual-Social Antimony in Discussions of Piaget and Vygotsky.  Retrieved from http://www.massey.ac.nz/~alock//virtual/colevyg.htm

Dockett, S. & Perry, B. (1996). Young Children’s Construction of Knowledge, Australian Journal of Early Childhood 21(4), p. 6-11.

Papert, S. (1999). Child Psychologist Jean Piaget. Time Special Edition 2000. Retrieved from http://www.ordiecole.com/piaget2.html (audio translation provided courtesy of the University of Geneva’s Archives: Jean Piaget).

Smith, L. (1989). Changing Perspectives in Developmental Psychology. In C. DesForges (Ed.), Early Childhood Education, Monograph Series No. 4, Edinburgh: Scottish Academic Press.

Stahl, G., Koschmann, T., & Suthers, D. (2006). Computer-supported collaborative learning: An historical perspective. In R. K. Sawyer (Ed.), Cambridge handbook of the learning sciences (pp. 409-426). Cambridge, UK: Cambridge University Press. Retrieved from http://GerryStahl.net/cscl/CSCL_English.pdf

Wood, E. & Bennett, N. (1998). Teachers’ Theories of Play: Constructivist or Social Constructivist? Early Child Development and Care 140 p. 17-30.

1 Piaget’s work is impossible to summarize because of the incredible volume of it. Go to Google Scholar and type in Piaget. You don’t need to type in his first name; there was only one Piaget. You’ll notice that the first page is entirely books, articles, monographs, etc. by Piaget. On the first page of Google Scholar only, he has been cited 39,861 times. Click on page 5, and 10, and 15. Click on page 20. Really, go do it now. I was amazed when I found that there is almost no end to Piaget. Has anyone else in history produced as much as this one man from Neuchatel, Switzerland? Has anyone else in history ever been cited this much? You may note that Vygotsky has an impressive 29,980 citations on his first page of Google Scholar. And we must remember that Vygotsky died at the very young age of 37. So perhaps it is unfair to compare them. And since they both wrote about the most important theory of learning ever explained, I won’t compare them more than is necessary, for they together make the whole that is known as constructivism. But Piaget came up with it first.