Kinds of openness in cognitive science

The response to the first few days of open source cognitive science has been gratifying. There are fascinating problems and technical challenges in this area. It is a treat to be able to think about them together with interested colleagues.

Cognitive science could benefit by more fully adopting some of the existing forms of openness:

  1. Open access, where publications are made available on the web without charge (cognitive scientist Stevan Harnad is a champion, with CogPrints being cognitive science’s answer to the Physics pre-print arXiv).

  2. Open courseware is a movement that invites educators to make their course-material directly available on the web. (An example is Tutus Vilis’ completely Flash animated course, The Physiology of the Senses—Transformations for Perception and Action)

  3. Open source hardware and software, where plans for apparatus and tools for analysis are made freely and openly available (Praat, for example, is a GNU GPL licensed signal analysis package which can be used for analyzing speech, generating auditory stimuli, and doing speech synthesis).

  4. Open stimuli, where stimulus sets or corpi are made available for use in replications or new experiments (See the Psychological Image Collection at Stirling (PICS (, or the Irvine Phonotactic Online Dictionary (Vaden, Hickok, and Halpin (2005))).

  5. Open workflows, in which researchers can freely share chains of experimentation, analysis, and visualization. (Talkoot is a collaborative workflow initiative for Earth Science. A quick search of the workflow custom search engine on Google reveals no workflows in the cognitive sciences).

  6. Open data, where individual researchers release their datasets, either as the data is collected, upon publication, or after a suitable embargo period. (An example with a rich dataset would be The Linguistic Atlas of the Iberian Peninsula (ALPI)).

  7. Open model repositories, where computational models from published papers can be centralized. (ISPOC is an open modelling initiative which includes cognitive models).

  8. Open research, where open lab notebooks are used to describe ongoing details of a particular strain of research.

Cognitive science can include up to six areas of research (psychology, neuroscience, linguistics, computer science, philosophy, and sometimes anthropology). One could visualize an 8×6 matrix serving as a preliminary grid for exploration.

Did I miss anything? Other great examples? Your thoughts are welcome.


Boersma, Paul & Weenink, David (2009). Praat: doing phonetics by computer (Version 5.1.11) [Computer program]. Retrieved August 3rd, 2009, from

Vaden, K.I., Hickok, G.S., & Halpin, H.R. (2005). Irvine Phonotactic Online Dictionary, Version 1.3. [Data file]. Available from


12 responses to “Kinds of openness in cognitive science

  1. I’m not completely systematic about it, but I like to do a lot of my reading in the open. Much of what I read for my book project I tag “tfos” at delicious. If someone wants to follow my reading, they can take a good go at it that way. Unfortunately, this omits stuff I find offline, including most books. I’ve only gradually become systematic about it, so the record is far from complete; it’s more of an ongoing experiment than anything else.

    • I’ve been trying to think a little more about how to make my paper offline reading available. One of the reasons I’m adding references at the bottom of each of these blog posts is to make a record of them, so I can later take all of these references and add them to a downloadable BibTeX bibliography, or something similar. Apart from being open source, BibTeX writes text files, so it should presumably work fine with a versioning system.

  2. My lab looked at using Zotero or Mandeley for sharing readings. Unfortunately Zotero’s collaborative functions are still quite immature, and Mandeley is quite restrictive on sharing (and proprietary). For now were stuck with emailing PDF’s around, but hopefully Zotero will mature into a collaborative database of some kind. There are lot’s of link-sharing type sites that can be used for this, but they often don’t provide very good literature database functions for writing papers, and who wants to keep two databases.

    • One of the things I’ve been trying is putting up a shared directory of papers on the internal lab server. This is very simple, but by carefully labelling the filename, they sort nicely, and I can find papers quickly. I can then link to these from a intranet wiki. My hope is that I can also link to them from a BibTeX database.

      The nice thing about the server approach is that it doesn’t matter where I’m working – I always have access to the papers I’m reading or might want to refer to.

      Many university libraries now offer internal access to papers that it would be really handy if you could link to the library’s copy from your bibliography program. If this was reasonably transparent, it could solve some of the sharing problems.

    • I’m hoping Mendeley gets better at sharing (and I’m sure it will, they’ve been very good at responding to feedback thus far) because its aggregation/suggestion features already have me hooked.

  3. I’ve been struggling with some of these issues as well, and would love to see something like SciRate get used for cog sci/HCI publications. I’ve started putting what I can on (workshop papers, for example). Is there a reason for having a separate cogsci repository? Why not just put everything into one big pile?

  4. By the way, SciRate should be able to handle CogPrints through the OAI feed.

  5. For archiving papers w/ bibliographic links, you might be interested in gPapers. It’s modelled on Papers (which is gorgeous), but a lot rougher around the edges. I might try contributing some code this year to help bring gPapers up to snuff. It’d be a great tool…

  6. Free Software is one key element, open access to data, coursework and the like another one, but don’t forget the importance of open formats.


    • Merwok,

      Open formats — yes! — a very good addition. And presumably open formats become more important the more complex the digital object. It would be nice if stimulus archives used PNG, but it would make little difference to accessibility. But for data sets that are more complicated than can be easily written to CSV, open formats become crucial. I suspect that any of the forms of openness listed above could benefit from open formats.

      • I don’t know about this stimulus archive, but the use of Flash for open coursework rings a bell.


        • Merwok,
          Good point! The use of Flash for open courseware would be a great example of a lack of open standards for complex objects. There is no easy way to take a Flash animation and tweak it to fit a new pedagogical situation. Assuming that a courseware author wanted to enable and encourage remix-ability as well as re-use, does anyone have any sense of what good open alternatives to Flash might be? The Wikipedia article on Flash has a section on open standard alternatives, citing SMIL, but I don’t know whether open authoring tools are available for SMIL that would make easy re-mixability straightforward.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s