I’ve been thinking a bit more about what the objectives of publication are and what this might suggest about the process. I think that there are three major objectives for scientists.
- Document our empirical findings
- Document the evolution of our ideas
- Efficiently allocate attention to ideas and results which best fit the ideals of the scientific community.
Clearly, the set of objectives changes a little bit if we consider how to publicize results outside of the scientific bubble, but as a starting point, let’s stay within our little community. Looking at the three points above, the third one seems a bit different from the first two.
####Documenting Empirical Findings#### This objective should be self-evidently important. When we conduct research, we absolutely must inform the rest of the field about what we find. If this didn’t happen, every scientist would be an island unto himself or herself, and we really wouldn’t get anywhere.
####Documenting the Evolution of Ideas#### This happens to some degree already, but not enough. If you’re working in a field like psychology, where your theory about underlying psychological constructs, processes, and mechanisms shifts over time (and it should!), it’s important to document how these changes occur, and what leads you to shift your view. Too frequently, theory and ideas contained in published papers (especially if they’re highly cited) are seen as concrete and immutable. I think if the documentation of this process was more transparent, then we wouldn’t fall victim to this problem of viewing ideas as static.
Regardless, in addition to communicating empirical work, the communication of ideas and their evolution not only gives a history of an individual’s work, but allows other researchers to direct their own work in a way which is informed by the work of others. You might or might not buy into the ideas that others have, but you shouldn’t ignore them entirely when conducting your own work.
####Efficiently allocating attention#### Whereas the first two objectives are more about the objectives of the publication(s), this is more about the publication process. One service that journals and conferences both provide is a way of getting people to attend to work which has been deemed ‘good’ by multiple eyeballs. Without this service, we’re dealing with a kind of data-dump situation. It’s difficult enough to stay on top of the increasing quantities of papers published in any given discipline. If the barrier to publication were totally removed, there would need to be some mechanism which ensured that the quality work received the attention it deserved (and that this attention was not driven by other factors, such as the prestige of the affiliated universities).
Journals and conferences, on the whole, seem to be able to accomplish the allocation of attention. I’m not sure they do it perfectly, but they do it well enough.
The former two are a bit trickier. Yes, we publish our empirical findings, but we only publish a few of them - the few that tell the best story. Unfortunately, best != most likely. Rather, best is often the one with the nicest narrative. And I would argue that mostly we do not document the evolution of our ideas.
One possible method of fixing this is to keep the journals (or journal-like tools) around for people to publish periodic updates, but to have folks publish the more day-to-day information in a kind of lab-based wiki. Once you’re familiar with the work that a lab publishes (or, if you see an interesting result in a paper or at a conference and want to know more), you can then go look at the wiki and see how they’re thinking about that topic this week. Ideally, you could also go back to see how that idea materialized and shifted. Furthermore, people should receive credit for the work published in this manner, and the two sources (journals and wikis) should be linked, with each referring to the other for more detail.
I suppose this is similar in form to the lab notebook. A clear downside is that it is easy to be overwhelmed by information. And on a more pragmatic level, there is a lot of science done by people who do not have the technical skills needed to maintain something like this. Not to say that it couldn’t be learned (and probably fairly quickly), but the impetus just isn’t there for them to learn, and technical skills should not be a barrier to intellectual contributions.