found drama

Recently, Nicholas Zakas wrote a blog post about switching from “embedded links” to footnotes. His decision was largely in response to some data cited in Nicholas Carr’s book The Shallows ¹; of those data, Carr writes:

The test subjects who read the pages linearly actually scored considerably higher on a subsequent comprehension test than those who clicked back and forth between the pages. The links got in the way of learning, the researchers concluded.

Zakas summarizes and concludes:

In each study, participants were given the same material to read with and without links, and those who read the material without links always tested higher on comprehension exams later. This immediately made me think that I was doing my readers a disservice by including links directly in the text.

And while I admire that he put some serious thought into this question, I also respectfully disagree ². And here’s why.

First, let’s consider the footnote and the endnote ³. Depending on the format of the text, the footnote might be a spartan, strictly utilitarian mechanism for citations and references (i.e., legal documents, academic or scientific papers), or it might be as vivid and lavish as the host text itself (i.e., Danielewski’s House of Leaves, just about anything by David Foster Wallace). In other words, if there are already some governing rules related to the format your text may take (academic/legal/scientific/technical documents, I’m looking at you), then there are probably some hard-and-fast rules about how you may use footnotes; but if your text’s format is freeform (literature, I’m looking at you), then how you use footnotes is almost entirely up to you.

But this is all about “non-interactive” text. The good ol’ fashioned printed page. Books and fliers and magazines and newspapers and pamphlets. “Dead trees”, as the kids say. You’ve got some physical boundaries to work with, and one of those physical boundaries is that you cannot turn the page from one book into the cited material ⁴.

A brief personal survey ⁵ of footnotes as used on the web appears to me to down into two camps which (for the sake of being colorful), I’ll summarize as “the Daring Fireball camp” and “the Wikipedia camp”.

“The Daring Fireball camp” uses footnotes for asides–for supplementary information, for tangents and punchlines to embedded jokes, for material that might otherwise go into a parenthetical but is perhaps “too much” for a parenthetical. Maybe that material is too long to sneak in (“like this”) or perhaps to include it inline would be too disruptive ⁶. Do these footnotes explicitly exclude citations? Certainly not, but the principal purpose of the footnotes is not to provide a place for the references. The underlying philosophy here seems to be: “We already have a grammar for references on the web, and they’re called links.”

Mr. Zakas would fall into “the Wikipedia camp”, employing footnotes for the important citations (online or otherwise). When you follow the superscript or bracketed number to its mate at the bottom of the text, you will find a reference and probably also a link to the referenced material. It’s utilitarian, and it makes sense; this pattern is grounded in those ages-old writing traditions we cited above–the “dead tree” tradition. One could easily argue that this is a “more correct” approach due to the fact that it has a longer-standing tradition to back it. In some ways, I agree that the footnotes-for-citations approach is more appropriate; from a strictly stylistic perspective, and given that Zakas (in this case) is writing on more academic/scholarly interests, then they make the online document appear more like its printed brethren. But there seems to be a bit of a catch in doing it this way–and that is that hardly anyone seems to cite references this way on the web.

Does that make “the Wikipedia camp’s” footnote method wrong? Not in the slightest, but we need to ask about what the problem is that the footnotes are trying to solve. And I believe our answer is two-fold.

First, what is the problem that the footnotes are supposed to solve? It comes back to that quote from Carr that “links got in the way of learning”. How? In what way did links get in the way of learning? Having not yet had the opportunity to review the originally cited article, I can only speculate–but the (perhaps obvious) speculation is that the impairment comes from the disruption. Links break the linearity of a given text, creating recognizably actionable opportunities for the reader to “move beyond” the text, onward to whatever it references. If a reader interprets the link as being an urgent call to action, then the reader is going to click the link, follow it, experience a context shift (i.e., navigating from one page/site to an entirely different page/site), and then have to decide whether to consume that new content (i.e., read that page) or navigate back or dive further down the rabbit hole ⁷; in other words, depending on the reader, you could lose them on the very first sentence. But that sentiment sounds like an SEO specialist (“You don’t want to lose eyeballs! Think about that bounce rate!”), not some scholar who is concerned about whether you will fully comprehend the point he is trying to make.

And that brings us back around to our question: how is it that links are getting in the way of learning? Is it the context switch? Or is it the disruption? And by “disruption”, I mean the fact that we are creating any opportunity for the reader to decide: “Do I plow ahead in the text? or do I follow this bifurcation?” If the problem is a major context switch (i.e., navigating from one site to another) then yes, switching from links to footnotes could solve the problem. But if the problem is the bifurcations, then footnotes in an online text are arguably more disruptive than links. Why? Because footnotes are non-normative on the web and by using them, you are potentially increasing the cognitive burden on the reader. Their consumption patterns goes from: read → see link → follow or continue? to read → see footnote → what is that? → follow or continue? ⁸. In a nutshell, the argument that cautions against the footnote is that readers don’t know what to do with them, they don’t know what to expect ⁹.

Granted, a scholarly/academic article discussing the idiosyncrasies of some up-and-coming API is different (and has a different audience) than some tech journalist’s blog post with a raging link-bait headline. So you would expect that some reading the former would “know better”–that he or she would know what to do with a footnote (or at least be able to figure it out quickly enough). But if that’s the case, and we are expecting more of them, why not just leave the links as links–aren’t those readers also disciplined enough to leave the links alone until they need them? and/or to circle back on them later? and/or back-track and re-read as necessary? Carr would say that the question is irrelevant–that you can’t blame the audience, nor can you hold them accountable for how they read what they read.

But that’s hopelessly bleak and deterministic.

We’ll skip the sidebar about how all audiences are not created equal and stick to the main point: that even if Carr’s conclusion is right–that “links got in the way of learning”–that we need to be sure that we are looking deeply on the matter. A link is merely an apparatus for navigation and citation, it’s not some demon dust that causes cognitive impairment just by walking into the room. Something else is at work there.

My theory is that it’s the disruption (any disruption) of the text’s linearity that’s the source of the observed “impairment”; and that you just have to have faith that curious and motivated readers will finish and “get it”. The answer is in that ignored first half of the observed quote:

The test subjects who read the pages linearly actually scored considerably higher on a subsequent comprehension test than those who clicked back and forth between the pages.

But I could be wrong.

I wish Zakas good luck with his footnote experiment; and I’ll keep using mine for asides.

1. I have previously reviewed The Shallows on this blog.[↩]
2. And that being despite my gratuitous love of the footnote.[↩]
3. Which for our purposes will just be “footnotes” from here on out.[↩]
4. Not unless it’s one hell of an omnibus.[↩]
5. And by “brief personal survey”, I mean that I poked around to a few sites that I visit somewhat regularly (and a few not-so-regularly) looking to see who is using footnotes. I assure you it was at least the beginning of “scientific”.[↩]
6. Or perhaps the whole point is to disrupt the reader. Infinite Jest readers, you know what I’m talking about.[↩]
7. And/or give up on that reading session all together.[↩]
8. And depending on the implementation of the footnote, it may also require some manual scrolling and/or scanning instead of clicking. But that gets us into a weird sort of “meta” area within our argument. And/but that is a digression not worth diving into at the moment.[↩]
9. This is where maybe things get a little muddy: because you could easily argue that the “grammar” around links is already pretty vague because you don’t know what’s on the other end–the cited text? an Amazon affiliate link? that Rick Astley video? But putting the aside, the important bit (for me at least) surrounds the question “what do I do with this?” and not “what’s on the other end of this?”[↩]

About Rob Friesel

2 Responses to on footnotes, links, and cognition