On Using Twitter

tl;dr

When I started thinking about writing this, I thought it was going to be a piece about using Twitter to build an effective platform and use it well. But the more I thought about it, the more I realized that learning on Twitter is at least an equally important part of the story. Similarly, I initially conceived of both my Twitter presence and this piece as strictly academic/professional, but I have come to learn that academia and scholarship are inherently political and we are better off recognizing that.

Thus what you’ll find here includes reflections on the affordances of Twitter as a platform (what the software lets you do), thoughts on how to build an audience and communicate effectively, and thoughts on using Twitter both to learn about social justice issues and engage positively.

History/author positionality statement

To help contextualize the remarks that follow, I want to locate myself in social space and sketch my own history with Twitter.

First, social address: I’m a white, cis, female, straight, abled (currently; this can change) US citizen with a PhD, raised by parents with advanced degrees. So, working with and working on understanding a whole lot of privilege. A lot of the advice in this post is similarly meant for people operating from a position of privilege.

I first joined Twitter in July 2010, mostly to see what it was about, and didn’t find it particularly useful until the following summer where there was a conference I couldn’t attend that had an active hashtag. My first post, promoting a tutorial I was going to give at a conference, wasn’t until March 2012.

My Twitter account has always been an expression of my professional identity. Facebook, on the other hand, I use more for personal posts (though people in my field are heavily networked on Facebook, so I have a lot of professional contacts there too). My Facebook account, especially since 2016, has also been very political, in terms of documenting my own political activities and urging others to also take positive political action.

However, over time, I’ve come to realize that I can’t keep my professional identity and professional communications apolitical: basically everything is politicized (even things that seem like they shouldn’t be). Science and academia in general are permeated by power relations that disenfranchise scholars from historically and presently marginalized groups and pretending that that’s not the case/declining to talk about it is just as “political” as speaking out against it and working to overturn it. Furthermore, as tenured faculty, this is an area where I might be particularly well positioned to make a difference.

Learning on Twitter

At some point in the past decade, feeling helpless in response to one of the many episodes of white supremacist violence, I saw the advice to start following and reading Black authors. At the time, this seemed like a hopelessly passive and therefore meaningless action in the face of so much injustice. But it was also something I could do, so I did it — and then kept doing it. A great thing about Twitter is that through following one account you can find other excellent people to follow by noting who they retweet and exploring hashtags they use. An early follow for me was @IBJYONGI, Prof. Chandra Prescod-Weinstein (then a post-doc at my same institution, which is probably how I came across her account). Prof. Prescod-Weinstein is a theoretical physicist and feminist theorist.

It’s absolutely true that my starting to read & follow back in 2012 or 2014 didn’t change anything about the world in 2012 or 2014. But in the time since then, reading and learning from Black scholars, disability rights activists, trans scholars and activists and others has helped me become better at understanding the world around me and being able to articulate what’s wrong when people (some well-meaning, some not) engage in both-sides-ism and other derailing tactics. Being able to articulate what’s going on is useful both for me personally (e.g. when something feels off, but I previously couldn’t put my finger on it) and for my effectiveness in trying to make positive change. The learning I have done on and through Twitter has also helped me engage in better scholarship, especially around the societal impact of NLP.

Here’s a quick rundown of strategies I’ve found to get the most out of Twitter as a place to learn:

• Use tweetdeck (tweetdeck.twitter.com). This interface makes it much, much easier to follow things than the twitter app does. I tend to have one column which is just my “home” (i.e. everything all accounts I follow tweet, which tweetdeck helpfully leaves in reverse chronological order), and then additional columns for specific hashtags I’m following.
• Cultivate a good set of accounts to follow, with an emphasis on people whose lived experience is very different to your own and especially people whose lived experience involves systems of oppression in which you have privilege. (At the same time, I follow relatively few people just for shared areas of research interest. I can usually find relevant tweets about NLP on the #nlproc hashtag and don’t need to also see them in my “home” column.)
• Find some good hashtags to follow: I first found Twitter useful when I was following the conference hashtag for conferences I couldn’t be at (and then for conferences I was at, to see what people were saying about them…). I found excellent people to follow through hashtags like #BlackinSTEM and #WomeninSTEM. Recently, @DrShardeDavis and @smileitsjoy started #BlackInTheIvory, which quickly became a catalyst for eliciting a powerful set of stories clearly documenting the need for systemic change in the academy.
• Develop strategies for learning from conversations that aren’t really “for” you. People use Twitter in many different ways. Sometimes they are intentionally talking to a broad audience. Sometimes they are having a conversation that they don’t mind having overheard. Sometimes they are just thinking aloud. And any given individual might well do all of these. So sometimes (often?) as a reader you’ll come across tweets for which you lack the context. In some cases, the context is there, if you look upthread, or at previous tweets (not threaded). Sometimes it isn’t. If you can’t find the context or still don’t understand, then move along. Keep in mind that you can learn a lot from people on Twitter, but it’s not anyone’s job to teach you.
• Learn to listen to your discomfort. Sometimes I find myself reading something that makes me think, “I respect this person, but would be uncomfortable holding that opinion myself.” In such cases, it’s worthwhile thinking about why I find it uncomfortable. Is it someone speaking from unacknowledged privilege? If so, is this is a good moment to call them in* (either in DMs or in public)? Alternatively, is it uncomfortable for me because it challenges my privilege? If so, that is worth sitting with and trying to learn from.

Engaging on Twitter

Many people use Twitter without posting or without posting frequently, and that is fine. Here are some suggestions for using it actively as well as passively:

Getting started

• Decide what the purpose of your account is. What kinds of things will you tweet about? Who do you want to listen? Note that if you start following professional contacts, some will certainly follow you back! So if you’re posting to Twitter not in your professional identity, you may want separate accounts for professional and personal purposes. At the same time, it’s fine (important, even) to be human in your professional identity. I’d say this goes at least double for people who are established: it’s valuable for junior folks in the field to see the established people as whole people, with excitement and disappointments and lives beyond their work.
• Using hashtags: Hashtags are particularly useful for engaging in conversations and reaching people beyond your own followers. By following hashtags you can work out which ones (if any; this varies considerably by Twitter community and topic) are likely followed by other people and in particular other people whose attention you would like to catch. For natural language processing, the primary hashtag is #NLProc (because #nlp is unfortunately mostly about the weird pseudo-science of “neuro-linguistic programming”). If you engage in interesting, productive conversations on relevant hashtags, you may find yourself gaining followers as well.
• When not to use hashtags: Sometimes you learn a lot from a hashtag and want to draw other people’s attention to it, but don’t really have anything to contribute yourself. This was very much true for me with #BlackInTheIvory. The solution here is to break up the hashtag (e.g. #/BlackInTheIvory) so that your tweet about it doesn’t show up for people who are already reading the hashtag (it doesn’t match the search term) but humans reading your tweet can easily reconstruct what they should search for.
• Liking, retweeting, bookmarking: The discourse functions of liking, retweeting, and quote-tweeting are really interesting (like dissertation-worthy interesting, and I wouldn’t be surprised if there is research on this, but I haven’t searched for it). Sometimes in a Twitter conversation people use ‘like’ to indicate that they’ve read a tweet i.e. to acknowledge the other person’s discourse turn. Some people use ‘like’ as a way to bookmark tweets to come back to and will even put disclaimers in their Twitter bio to the effect of ‘likes and retweets are not endorsements’. I don’t think this is a good idea. If you want to save tweets to read again later, use the bookmark function (unfortunately not available through Tweetdeck, for reasons I don’t understand). If you like or retweet something, not only does it get pushed into your followers’ feeds (yes, Twitter sometimes shows you tweets just because accounts you follow have liked them), people can also go to your profile and see what you’ve liked. Finally, it’s hard to see how retweeting something without even quote tweeting it can mean anything other than: I want this to have more play on Twitter. (More on retweeting below.)
• @ing: When you tag (or ‘@’) someone in a post, you cause a notification to get sent to them and also draw the attention of anyone reading your post to their account as well. So it’s worth asking yourself: Would the person tagged want to see a notification of this tweet? If not, is there some reason for them to see it anyway? It also means it’s kind of like very publicly walking up to someone and tapping them on the shoulder. All of your followers can see that act. Someone who is mentioned is of course not obligated to reply (and keep in mind that not everyone keeps continual tabs on their Twitter notifications!), but the lack of reply is visible, to a subset of Twitter. An additional wrinkle is that the Twitter interface (including Tweetdeck) presently includes everyone mentioned in a tweet in the mentions of all replies to that tweet (and replies to those replies…) unless they are specifically removed as a reply is being composed (then replies to that tweet only include those mentions not removed). Finally, any actions (likes, retweets, replies) taken on tweets an account is mentioned in can send notifications to that account. Accordingly, it can be important to untag people in replies/subsequent tweets in a thread, so as to not overwhelm their notifications.
• Threads: Gone are the days of 140 character tweets, but even 280 can be limiting. If you have something you want to say that doesn’t fit in 280 characters, you can create threads by replying to your own tweets. The Twitter and Tweetdeck interfaces privilege replies by the author of the original tweet, so even if your first tweet gets lots of replies, people can still find the rest of the thread. It’s good to indicate that tweets belong to a thread, by numbering them, or ending each non-final one with > or similar. If you want others to retweet your points, it’s good to try to keep each individual tweet relatively self-contained. If you’re in the middle of a conversation with someone, it’s worth waiting to see if their replies are coming as a short thread so you can reply to the last tweet of that thread … otherwise, things can turn into a somewhat confusing, multi-branching tangle. Finally, if someone replies with a key piece of information that you want others reading your thread to find easily (e.g. an answer to a question your original thread posed), you can highlight it by tweeting a link to their tweet at the end of your thread.

When not to tweet

There are several situations in which the best choice is not to tweet (or, to steal a turn of phrase from Fuji-Q, tweet inside your heart).

• When the conversation isn’t “for” you: When you are learning from other people’s tweets, especially ones sharing personal experiences, and especially from accounts you’ve been following for a while, it can feel like you know the people talking. Keep in mind that such relationships are typically one-sided: if they don’t follow you, they don’t similar have a sense that they know you. It can be appropriate to show appreciation by liking a tweet, or in some cases, by retweeting (see below), but in such conversations, replies from non-mutuals are likely unwelcome.
• When you feel defensive: If something you read, either addressed to you or not, makes you feel angry or defensive, it can be valuable to refrain from replying, at least in the short term. Any tweet you compose while angry is more likely to be something you’ll regret later, including cases where it can easily be taken out of context. Also, sometimes if you wait, the problem might go away. A few times I’ve had someone reply to me with something particularly vile only to see them delete it a few minutes or hours later.
• When you might be breaking anonymity policies: My field (#NLProc) is working to maintain fully anonymous review. In such conditions, it’s important to be careful when tweeting about work (your own or others) that is under review. If someone has shared an anonymous preprint with you, tweeting about without revealing the author identity is good. (NB: I’m assuming you’re not the reviewer here! Keep manuscripts you are reviewing confidential.) Tweeting about your own anonymous preprints is not a good idea.
• Snitch tagging: You may come across a conversation where people are discussing someone else and be tempted to reply by tagging the person discussed. This is snitch tagging and it’s frowned upon. Especially if that someone else is more powerful than those discussing, you could be causing trouble for those in the discussion. If the person being discussed is someone you actually know AND you think they would want to be aware of the conversation, you can send them a link to it, without publicly tagging them.
• To trolls: You may attract people whose only goal in tweeting at you (or replying to your tweets) is to troll. In this case, it is frequently not worth replying. You can also mute, soft-block, or block the account that tweeted at you. Muting means you don’t see what they tweet but they can still see your tweets. Soft-blocking is forcing the other account to unfollow you (block and then unblock). They can still see (and reply to) your tweets, but your tweets won’t show up in their timeline unless someone else they follow likes or retweets or replies to them (or they re-follow you). Finally, blocking means they can’t see or reply to your tweets when logged in as that account.

Once you have built a platform

The experience of Twitter changes as your platform grows. It’s very different to be tweeting to 5000 followers instead of 500. It’s always worth being intentional about what you tweet (tweets are public [by default] afterall, and anything can go viral), but even more so once your audience is large. Also, the larger your platform, the more opportunity you have for positive impact.

• How to lift up other voices: Even on Twitter, where text-based interaction would seem to make demographic factors less salient, still there is evidence that female twitter users from academia and tech have fewer followers than male academics/tech workers, and presumably similar things can be said about Twitter users marginalized along other dimensions. Once you have a platform, you can use it to do something about this by amplifying the voices of others. The primary way of doing this is through retweets.
• Retweet thoughtfully and respectfully: First, before you retweet, determine if the author of the tweet really wants a larger audience to see it. If you’re unsure, you can contact the author and ask. Second, if you are retweeting to amplify, first do a plain retweet, i.e. share the author’s own tweet as is, without using it as a springboard to say something else. Then, in addition, if you think there’s something you could say that would help contextualize the tweet you are retweeting for your audience (and thereby bring it more readership), you can also do a quote-tweet and provide that context. Keep in mind that quote tweeting both shows greater engagement and moves the conversation over to your audience (away from the original poster’s), if people engage with your quote tweet. Finally, be prepared to help out with trolls that your retweet brings to the original author, if such help is needed & wanted.
• Clean-up trolls & sealions thoughtfully & respectfully: It is an unfortunate fact of Twitter that the platform attracts lots of trolls, and furthermore that some accounts have large numbers of troll followers who will someones launch coordinated attacks. There are also Twitter users who don’t think of themselves as trolls (i.e. who don’t conceptualize what they’re doing as making trouble for trouble’s sake) but who like to argue odious opinions ad nauseam. Sometimes this takes the form of sealioning. In my experience, it is also much easier to handle trolls and bigots when I am not the specific target of the bigotry. When I am, and someone else joins to take up my side, it is a tremendous relief. And so I do my best to offer similar support to others when I can. However, sometimes the target of the troll would rather just ignore them (“Don’t feed the trolls”) so it’s worth checking first whether they actually would want you to engage and, should they say yes, untagging the target from your replies so the conversation doesn’t keep showing up in their mentions.

Who has time for this?

Engaging seriously with social justice issues, on Twitter or in other ways, can take a lot of time (even setting aside the time sink of trolls and sealions). And so you might ask: Who has time for this?

Nobody. The answer is that nobody has time for this. But the sad truth is that the ability to opt-in or opt-out of these discussions is only afforded to some of us. If you have privilege, and you’ve been wondering how to do actually use it for good, this one concrete way: Take up some of this work so that it doesn’t keep falling on the same people who are furthermore often the direct targets of the oppression and therefore not as well positioned to push back against it (though better positioned to understand it).

Postscript: Affordances of the software

Finally, I’d like to end with a few notes about the affordances of the software that didn’t fit above but are worth keeping in mind as you start using Twitter:

• Most of what is sent out over Twitter is publicly readable and searchable, but what any given user actually sees is highly variable. Thus something that is very salient to you (because you were tagged and got a notification) might be invisible to most (if the person who tagged you doesn’t have many followers). Similarly, you might get the sense that “all of Twitter” is talking about something, when in fact that’s only true of some relatively small community on the platform.
• Sometimes you’ll see people sharing a screen shots of a tweet instead of quote tweeting. Screenshots differ from quote tweets in that they avoid sending a notification to the person whose tweet is being shared and also persist even if the original tweet is deleted. Thus sharing a screenshot of a tweet is not a friendly thing to do.
• Speaking of screenshots, Twitter allows you to add alt text to any images you share. This is an important practice to keep your tweets accessible to people who use screen readers. I’ve also been told that ascii-art memes (e.g. “In this house…”) are disasters for screen readers. If you want to play along with such a meme, the best practice for accessibility is to share it as a screen shot with alt text indicating the type of the meme and the text you added.
• In the Tweetdeck interface at least, conversations with muted participants (as in accounts I’ve muted) get very confusing because Tweetdeck displays them as if the muted account’s tweets don’t exist. This makes it look like tweets that reply to the muted account’s tweets are actually replying to something further up thread. Tweets are short, and so tweeted replies usually rely heavily on context. When the interface hides or changes that context, the message conveyed can change drastically. I’ve been tripped up by this a few times. If someone is saying something that seems out of character or like it’s replying to the wrong tweet, it can be worth double-checking if there’s something missing from the conversation by viewing it in an incognito window (i.e. not through your own account).

*On calling in vs. calling out: Calling out is making a public statement about a way in which someone else has done something unjust. It can be important: leaving e.g. racist statements or other racist acts unchallenged conveys tacit support. But calling someone out rarely leads to a change of behavior on their part. Calling in, by contrast, is an invitation to the person to acknowledge their mistake and make amends. In my experience, approaching someone with a question can function as an attempt at calling in with an easy pivot to calling out if they double down: “Wow, that makes it sound like you believe X. Is that really what you mean to convey?” For another take on calling in vs. calling out, see https://everydayfeminism.com/2015/01/guide-to-calling-in/