Coming Soon: The Virtuous Cyborg
Technological Cowardice

Lessons from the MUD

AccursedLandsAnonymity and technology mix badly. While you are required in most countries to pass a test of skill with cars, our most dangerous tool, and even the US licenses and records the identity of firearm owners, any fool can appear on Twitter or Facebook with a fictional or falsified identity and act abusively towards the cyborgs they encounter there. However, eliminating anonymity by forcing the use of public identities is a heavy-handed solution that would almost certainly prove insufficient for eliminating the problem, as Brian Green has carefully outlined. But there are lessons that can be learned from earlier digital public spaces that offered anonymity but had less of a problem with abuse, and this can put a different slant on these kinds of problems.

The Multi-User Dungeons, or MUDs, began as spaces for creative adventures, greatly conditioned by the play of the pivotal tabletop role-playing game Dungeons & Dragons. These imaginary worlds were entirely made of databases of text, arranged to create the impression of connected rooms and spaces, within which players could move around and encounter one another. Players would join a MUD using network protocols from the early days of the internet, creating an account with a name that would become their identity in the shared space of the game world. The MUDs would go on to provide the basis for graphical games like EverQuest and World of Warcraft that would achieve tremendous commercial success.

A player coming to a MUD for the first time was likely to have been invited by someone else, and as such was not strictly alone. Nonetheless, players typically entered the text world as individuals, and since players would connect at different times they were often (if not always) alone. Starting players were always unknown to the existing players, so there was always an element of uncertainty about the arrival of someone new. Nonetheless, the community surrounding each MUD, which was typically a few hundred players or so, generally welcomed newcomers, and there was an air of hospitality extended in most MUD communities. Abusive players, then as with in the larger digital spaces today, were the minority, and would quickly come into conflict with the more responsible players who would report them to the administrators, typically entitled Wizards.

The Wizard system provided legislative, judicial, and executive power within the MUD. While the first Wizards would be those who set up the software and provided the hardware to run the game, many MUDs used a democratic system to elect additional Wizards, who worked as a collective to maintain order and community. Legislative acts concerned the code of conduct that applied, and thus set the boundaries of acceptable behaviour – such matters were always resolved by the Wizards working together, and generally involved consulting the wider community of players as well. Judicial and executive power was expressed by taking action against troublemakers – in many MUDs, miscreants could be ‘toaded’, which reduced a character to a powerless amphibian. Wizards would hold tribunals in this regard to determine the justice of any such punishment meted out. Although I have heard of some instances of ‘corrupt Wizards’, my own experiences showed the Wizard system to be highly effective at minimising abuse in MUDs.

While on the surface, MUDs were play spaces, in practice the division between game and communication system blurred. This was especially so because MUDs provided the first networked text communication system that didn’t require manual delivery, like a telegram. As such, many attracted a community of ‘players’ using them solely as a chat interface. These were the original chatrooms, since players would typically congregate in a room of the MUD’s fictional world to engage in conversation. This occasionally caused tension with other members of the community who were using the game differently, but for the most part it was a fact of life in MUDs that some people were there solely to chat, and facilities to do so were expanded in the code for MUDs as the 1990s progressed.

The MUD was the direct precursor to Facebook and Twitter, which descend from earlier copies of the chatroom concept, such as AOL’s offering, which lacked the fictional world but kept the name. Yet abuse in MUDs was comparatively rare, and rapidly resolved by Wizards whenever it occurred. Anonymity may still have fostered abuse, but the systems were in place in MUDs both to guard against it, and to discourage it from happening in the first place. The most effective deterrent against online abuse is community – and the MUDs fostered this far more than the latest digital public spaces.

Thus while a new MUD player might arrive alone and unknown, they were never unguarded – both in the sense of protected from the abuse of others, and watched for signs of conducting abuse. Conversely, a ‘tweep’ (as a user of Twitter is colloquially termed) is alone, unknown, and essentially unguarded – and these are the optimal conditions for abuse to fester. Twitter has an abuse reporting system, but it is distant and bureaucratic, with no community to manage the warnings and reporting, and no community-engaged Wizards to act as magistrates.

Here we have three different senses of ‘anonymous’, all of which contribute to cyber-disdain, and thus a greater risk of cyber-cruelty. To be alone in a digital public space is to lack a community, and crucially ‘follows’ and ‘friends’ do not mark the authentic social bonds of a community relationship but merely an open communication channel. To be unknown is to be anonymous in the sense of having a concealed identity – a situation that fosters abuse if it is not offset by community relations. Lastly, unguarded marks an invisibility to the systems of justice within a digital public space – a situation worsened by being unknown, and by being alone.

Thus Facebook’s requirement to use conventional identities (to eliminate being unknown) is insufficient to stop abuse, both because its users are mostly alone and unguarded, and also because the size of its membership means that with random encounters, cyborgs are still effectively unknown to each other. This is the fertile soil in which abusive behaviour online grows: as the cybernetic networks increase in scale, community is unsustainable since humans can only sustain viable communities at a scale of hundreds and never at a scale of billions. Two Facebook users, even with public identities, are effectively unknowable to each other – and nothing much can solve this problem short of managing encounters in a way that most would find intolerable. Guarding against problematic behaviour is more tractable when there is a village-scale community to engage, respond, and react – while at planetary-scale even robot-assisted magistrates are rendered impotent by the sheer scope of the network.

Anonymity is the root of online abuse, but there are at least three senses of this term that matter. We tend to focus on unknown anonymity, and thus miss the importance of alone anonymity and unguarded anonymity. My emphasis on being alone may seem misplaced. For instance, in his discussion of the problems of anonymity, Brian reports that “people in groups are more likely to transgress.” I agree with this claim, even though this may seem to run counter to my suggestion that alone anonymity is a key part of the problem. However, Brian’s point concerns ‘mob mentality’, and a mob is not a community in any relevant sense. Indeed, precisely what makes a mob dangerous is that people are alone together when they are a part of it – and this anonymity of the crowd (which also operates fairly innocently in audiences for musicians and so forth) becomes dangerous when the people concerned are also unknown and unguarded, as happens all the time in our digital public spaces.

When Sherry Turkle raises concerns about the way we are alone together online, she is not talking about the mob as such, but her work also emphasises this same concern: the undermining of authentic community by the current design features of the new communication systems. Yet different designs will produce different results. It is notable that blogs, which assign executive power to the blog owner (and thus are close to unguarded), and are ambiguous on the question of being unknown (since it is the blog owner’s choice how they identify) still manage to be less of a locus of abuse than the large-scale digital public spaces since bloggers are never alone. Forums tolerate contributions that are alone and unknown because they are not unguarded, thanks to the presence of moderators who can work effectively because the scale of the network of contributors is manageable. When a moderator ‘knows that such-and-such is a troublemaker’, they mean that particular cyborg is not anonymous in the sense of being unguarded. Different solutions to fostering cyber-respect (and minimising cyber-cruelty) hinge upon the different senses of anonymity.

What does not work – indeed, cannot work – is expecting our robots to plug the gap caused by scaling networks beyond human capacity to form a viable community. Abuse will remain endemic on Facebook and Twitter for as long as their cyborg participants can be functionally alone, effectively unknown, and inadequately guarded. If there are solutions to this problem, and it is not clear that there are, the most important lessons to learn are those revealed by the stories of the MUDs, the pioneering digital public spaces, the first cyborg communities of their kind.

With grateful thanks to Peter Crowther, both for feedback on this piece, and for running UglyMUG, a game that changed both his life and my own. Also, my thanks to Brian Green for his outstanding summary of the relationship between privacy and anonymity, which this piece responds to.


Feed You can follow this conversation by subscribing to the comment feed for this post.

Your comment about the US is generally not true: there are many ways to legally own a firearm without licensing. Identity is only checked on *some* sales. Details vary from state to state.

Automatic weapons, concealed weapons, and the like are licensed and restricted; the federal government controls automatic weapons, but concealed weapons are handled at the state level.

Hi Tom:
I thought that even on sales where no identity was checked prior to purchase, that details were still entered into a register. I'll be sure to modify the text for this in the manuscript of the book that this material is heading towards.

Thanks for your clarification here!

PS: If you let me know your surname, I'll include you in the acknowledgements of the book too.

A thoughtful response, Chris! Let me give you a few points to consider.

Early MUDs weren't exactly the paradise you mention. It a frequent that a Wizard would let the power and authority go to their head and would become abusive in their own right. I saw games collapse as one bad apple got Wizard powers and used them to exact personal revenge on people they thought had wronged them before. Or, when someone just got bored and wanted to listen to other people scream in impotent rage.

I think early MUDs were also helped by two other factors: scale and relative homogeneity. As you somewhat mention, MUDs were more village size whereas new social media, particularly the big names, are on the scale of continents. People sometimes talk about Facebook participation as a significant percent of the entire world population; it's hard to have a few elders keep the peace at that scale!

Plus, there was also a relatively homogeneous population in early MUDs: most people connected via universities, so most people had that in common. Most MUD players were male, middle-class, white, educated, and probably very nerdy. People who didn't fit these criteria learned to blend in or be ostracized. This helped people make certain assumptions about the person on the other end of the connection that we simply cannot make today when interacting on social media. Unfortunately, humans have always been kinda terrible at allowing for such diversity.

I'll also disagree about your assertions about people being "alone" and "unknown" and how that affects them. I am not convinced that being alone causes people to abuse others; humans tend to be largely social animals, so they often look to make connections rather than simply lash out at others. I still think my comment about "mob mentality" is accurate, and I think you miss an important point: the mob is still a sort of community, just not a healthy or functional one. A lone abuser throwing insults from a Twitter "egg" account is easy enough to ignore; you can block a single bad actor just fine. It's when there's a concentrated effort to harass someone with multiple anonymous accounts that this becomes a big problem. And this concentrated effort usually comes from some other community of people who have identified the victim as deserving the harassment. It's this organization that makes the abuse intolerable, not just a few lone voices striking out. And it's this ritual of striking out at enemies that reinforces one's membership in this community, even if it's a terrible community.

I also heartily disagree with the statement, "[Being] effectively unknown to each other [...] is the fertile soil in which abusive behaviour online grows...." While I agree that being unknown leads to people frequently not being overtly friendly to each other, people generally don't look for conflict with others they don't know. Usually there's a reason why people attack others, even if it's not immediately apparent to observers and not something we personally agree with.

I think the biggest issue we're looking at is scale, followed by cultural assumptions. Even back in the MMO days, us developers talked about how a lot of solutions wouldn't scale very well as we wanted bigger and bigger games. And, I think we're still struggling with the concept of a "cyborg" as a culture, and people need to change their thinking in how we interact with others in medium that doesn't foster the same cues we've used for a very long time.

Anyway, thanks for your response, Chris! I love the Cybervirtues series. Please don't hesitate to get in touch if you want to talk more.

Hey Brian,
We seem to be agreed about the problems of scale, and most of our disagreements come in the small print. Whether the mob is a community depends upon what is meant by community: if we mean 'a group of humans acting together', that obviously fits a mob. If we mean 'a group of humans connected by personal relationships', the mob is clearly excluded. That seems to explain our differences on that front.

The question of what proportion of online abuse is traceable to mobs versus lone individuals is actually incredibly difficult to quantify for obvious reasons. But I can at least online my impressions here. Firstly, that co-ordinated mob behaviour is a substantial proportion of abusive incidents, but far from a majority. A huge volume of abusive text online is 'drive-by comments', and this is not attributable to any kind of mob unless we're collecting people, say, by political views, which would seem to be a tenuous criteria. I think you are underestimating the volume of lone wolf drive-by abuse... the Pew research into online abuse suggests the majority cases (64%) are coming from strangers or unidentifiable cyborgs, for instance:

Now that doesn't necessarily exclude your mob scenario, but the evidence I've found when researching this doesn't support your claim that it is mostly mob actors. Of course, it doesn't categorically exclude it either. We might have to wait for further research to emerge - or it might be that these two scenarios will never be amenable to being distinguished by researchers. The evidence that we have confirms only that both kinds of behaviour - mob and lone-actor - occur. Nobody is able to take this much further.

Perhaps I can move us closer to common ground here, however, by pointing out that in drive-by comment abuse, the fact that others have also hurled abuse makes further abuse more likely to occur (in principle - again, research here is near impossible). Now you could call that a mob, and I might be tempted to do so, but it doesn't change the fact that the victim in these cases is unknown to everyone in the mob. Remember, I am marking with 'unknown' the absence of a personal relationship. And again, the Pew research favours this understanding. This perspective does not contradict your view regarding mob mentality, of course, but I stand by my assessment of the role of unknown (or effectively unknown) anonymity as an exacerbater of the problems.

Regarding 'alone', most of my evidence comes from other kinds of abuse (not online abuse), in which lone actors are by far the dominant player. For instance, a study reported here puts the figure at 70% for lone actor harassment:

Of course, this might not translate into the online context... but my sense of the situation is that this is the case. However, this does tie into your claim about humans being social creatures... people who conduct harassment or abuse alone are usually alone in life - they have no strong social ties to anyone else. This is the opposite pole to mob mentality: not that they are being empowered into acting by peer effects, but rather the absence of the moral representation that comes from having personal relationships sending someone down a dark road.

I think we are more aligned here than otherwise, and our disagreements are ones of definition and emphasis. Certainly, we are agreed that scale is the most important aspect of the problem.

Many thanks for your pushback! I will need to email you shortly to check you're happy with the way you are included in the book manuscript, since I have made you more of a direct character in that. :) I'll be sure to edit the text to incorporate some of the points you raise here before sending it on, though.

With gratitude for your support and your involvement,



I think we do largely agree, but as you say there are some finer points we don't quite match with. But, that's fine! Two smart people discussing differences in perspective is a great way to work through things.

I think that the one thing you're missing is about the connections we make online. I think that connections made purely online are different (not necessarily better or worse) than the ones we make offline, in-person. So, someone can go to a website, meet people and make online friends, who then decide that person X is "the enemy" and must be shown the error of their ways. The person goes with their online friends to harass person X.

In my experiences running (relatively smaller) online communities, it was rarely the single person who had much of an impact on others. It was a small group that emboldened each other who caused the real problems. They would support and feed on each other, which perpetuated the abusive behavior even if the victim didn't react initially.

I got your email, and I'll take a look sometime soon. Thanks again for the wonderful discussions. I really look forward to your posts!

Have fun,


Hey Brian,
It's interesting that our experience of online abuses are so orthogonal: you encountered groups, I have only encountered singletons. I would be tempted to say that this was a product of US versus UK influences, but all except one of the MUDs I played were in the States. It might, however, reflect your longer time in the MMO space, since I stopped playing in about 1995, and you just kept going.

I'm over in the States right now and was with Mike Sellers, and you came up in conversation. I told him that you were responsible for the continued running of Meridian 59 - I hope I got that right! :)

All the very best,



It could be a time thing. As the internet grew in popularity and people recognized the power of community, we saw more communities forming and therefore more community-focused harassment campaigns against individuals. I think it's easier to make connections with random people now than it was in the early and mid-90s.

Ah! I haven't talked to Mike in a long while. As for M59, I'm not currently involved with it at all. But, I did keep it going from 2000-2009. :)

Take care and safe travels in the States,


Verify your Comment

Previewing your Comment

This is only a preview. Your comment has not yet been posted.

Your comment could not be posted. Error type:
Your comment has been posted. Post another comment

The letters and numbers you entered did not match the image. Please try again.

As a final step before posting your comment, enter the letters and numbers you see in the image below. This prevents automated programs from posting comments.

Having trouble reading this image? View an alternate.


Post a comment

Your Information

(Name is required. Email address will not be displayed with the comment.)