A critique of the MLA citation format for tweets.

mla-tweet-citations.md

Why the MLA's Format for Citing Tweets Misses the Mark

A year ago, the Modern Language Association (MLA) announced its preferred format for citing tweets:

Lastname, Firstname (TwitterHandle). "Tweet message in its entirety." Date of tweet. Tweet. (Extrapolated from an example at http://www.mla.org/style/handbook_faq/cite_a_tweet)

Earlier this week, for whatever reason, many academics on Twitter were passing this info along as if it were brand-new. Many were excited that the option to cite a tweet as a tweet exists at all, and it is heartening that the MLA has attempted to keep up, citation-wise, with the far more rapid pace of technological experimentation.

A few scholars critiqued the MLA's decision because it did not include any space in the citation for the tweet's URL--something that the MLA used to require of electronic sources but has since changed its mind. Now, instead of including a source URL, the MLA only asks that it be identified as a "Web," rather than a "Print," text. This change is presumably to avoid the inevitable link rot that will occur with many websites as well as to circumvent the problem of multiple-line URLs incorrectly included in citations of articles located via electronic databases (which might include session-specific keys and be entirely impractical to type from a print bibliography into a browser).

Where the MLA has really failed, however, is not in overlooking the place for an optional URL but in fundamentally misunderstanding the nature of electronic communication in general and Twitter in particular.

Mistake 1: The 'Real Name' Problem

One of the MLA's mistakes in tweet citations--and in electronic-source citations more broadly--is in assuming that there is a genuine first and last name identifiable in connection with a given text. (This issue is connected to the much older conversation on web-based ethos and identity, but I will not rehash it here.) That is, does "Person A" on a given web forum actually possess any recognizable first and last name--assuming he or she has provided it--and, if so, is it a legitimate name (or is it an issue of alias or impersonation)?

In other words, why does the first and last name matter more than the medium- or resource-specific username or handle by which one may be more commonly known? When I communicate on Twitter, my username @brockoleur is how I "sign" tweets and how other users communicate directly with me. That I have put my real name into my profile is incidental.

There are many users who have not done so, or who put in obviously false (usually humorous) monikers in the name fields instead. Am I, as a scholar, supposed to override that desire for self-identification and refer to such a user by his or her real name--if I know it--or do I lend credibility to the potentially-false name provided by taking each user's information at face value? This set of concerns could be avoided by leaning towards the latter approach by citing the username rather than the name, which could still be mentioned when desired as an optional qualifier. For example, if I were, for some reason, to cite myself, I could say:

Kevin Brock, through his Twitter handle @brockoleur, has argued that ...

Because the information is available, it's been included--but if it were not, I would simply refer to the Twitter username. Note: I recognize that a Twitter handle can be changed, just as the information in a name field can be changed. However, because it is the preferable way to refer or respond to someone through Twitter as a medium, I suggest the username as the primary field or data point for authorship.

Mistake 2: The Absolute Reference Point

All web texts have as their Achilles heel the potential for link rot, the "dead end" that occurs when a website goes away for whatever reason. This is, as noted above, one of the likely reasons as to why the MLA changed its tune regarding the inclusion of URLs in its web-source citations.

However, there is a way in which many journals provide absolute reference pointers for their articles: the Digital Object Identifier, or DOI. An article with a DOI can be located in various databases thanks to a method independent of any one website.

While tweets remain reliant on the Twitter server to be tracked down and do not possed individual DOIs, there does exist a reference ID number, which functions very similarly to a DOI, by which a specific tweet can be located--it's directly in the URL for the tweet. For example:

https://twitter.com/brockoleur/status/365557713360535552

If the MLA doesn't want to make URLs required for only some web texts (such as tweets), it could, at the very least, provide a space for the tweet ID to be included. When combined with the other Twitter URL components, a reader could quickly and easily access the cited text.

Sure, this is a step or two of work more than one might want. An optional URL field could also save some typing here. In either case, there's more attention paid to the medium's qualities than currently is offered by the MLA's official format.

Final Thoughts

In the end, what the MLA could easily do is extend its current format for citing tweets in a way that:

1. offers only a bit more work on the part of the citing scholar (by adding in a tweet reference ID or the entire URL for a tweet)
2. identifies the user account as the primary form of author identification, especially as the "real name" attached to a username may or may not reflect the author's real identity

Will this actually occur? It's doubtful to happen any time soon, at least. However, it's important to recognize not just that a particular type of text differs from other texts ("Print" vs. "Web" vs. "Tweet" and so on) but that each possesses qualities absent from the others. If we want to treat these different texts--especially those of emerging media--with respect and as legitimate documents worthy of criticism, we need to pay close attention to what those qualities are and not merely graft the citation formats for other media onto these newer and distinct types of text.

I recognize the MLA's FAQ entry says that if a username but not a real name is known, use just the username. However, what I'm arguing is that the username--whether or not the real name is included--should be the preferable "primary" name field, with the author's real(?) name a secondary data point.

I agree that the Reference ID should be included. Without it, how does one view the tweet in conversation? Go to the user's page and scroll backwards? I think this current citation structure decontextualizes the tweet, treating it much like a single print text. Instead, the tweet should be seen as an utterance in a longer conversation.

This is another instance where the hyperlink is the best form of citation.

(Also: I too am confused as to why this MLA tweet citation has sprung up a year later. I'm thinking that the genealogy of this particular resurgence would make for an interesting blog post.)

Tim, great point. While it's possible to view a single tweet critically, as a stand-alone text, chances are high that each tweet is a point along a conversational trajectory--which needs to be acknowledged and discussed as well.

And what, of course, about the tweet that only comes to a reader's attention through a retweet? Does the reader-turned-citer give credit to the original author or should we consider another horrible print-based example that uses "retweeted by" as a kind of "reprint" or "indirectly quoted in" reference (as the latter is used for MLA citations of such texts)?

Great points, Kevin. Why do you think a style been created for a specific for-profit company rather than a more generalizable style? Is there another example of such an extremely narrow focus in the MLA style guide? Can I apply this style to posts from Plurk? Sina Weibo? If we are making Twitter a special case, what about the Library of Congress archive? How could that influence the style (which is presumably concerned with making posts easy to locate in the future)?

The proprietary considerations are definitely significant (as is separating a 'Twitter post' from a 'Facebook update' and so on).

What also of the inevitable need to cite (1) a user's body of tweets (his or her broad timeline)? What language is used there? (2) a massive body of tweets as part of a large data set--something Ryan Omizo has commented on in regards to my critique on Twitter?

I totally support the idea of adding in the unique tweet reference ID, or complete url as a more distinct identifier. Regarding this being more cumbersome for the citing scholar, with citation generators, such as the one I created (http://tweet2cite.com), making the process extremely easy, any added difficulty is quickly mitigated.