Fake Avatars and the Duplicates Paradox: An Essay by Extropia DaSilva

In ‘A Tale Of Two Avatars’, Wagner James Au reports on the discovery that there are two ‘Hamlet Aus’ on the social networking site ‘Avatars United’. Like many things to do with life on the screen, a superficial consideration of this discovery leads to a clear-cut and simple conclusion: There is the real Hamlet Au, and then there is a fake Hamlet Au. However, again like so many things to do with life on the screen, this clear-cut and simple conclusion may not hold true in all cases.


Why not? Well, it all centres on what is meant by ‘Fake Avatar’. The purpose of Hamlet Au is simply to provide one more means of getting in touch with Wagner James Au. It is like his telephone number or his email address. Just as a person hopes and expects to speak to Wagner when they ring his number, they also expect to talk to Wagner when they encounter his avatar in SL or any other social networking situation. In fact, it is probably fair to say that hardly anyone communicates TO Hamlet Au, anymore than a person speaks to their mobile phone. No, you speak to the person on the other end of the line THROUGH the phone. Similarly, Wagner James Au is spoken to THROUGH the Hamlet avatar.

Clearly, then, there is a big difference between speaking to Wagner James Au and speaking to some person pretending to be Hamlet Au. That is why it makes sense to talk about ‘real’ and ‘fake’ Hamlet Aus.


But in what circumstances might this not be the case? The answer is when an avatar is not used just as a means of communication with a specific RL person whose identity is generally known, but when the avatar is a roleplayed character. If you define a fake avatar as ‘some person pretending to be [insert name of avatar here]’ you immediately run into trouble when talking about roleplayed characters, for they are, by definition, some RL person pretending to be [insert name of avatar here]. What is the difference between some person pretending to be a character, and some other person pretending to be that character?


At this point, I want to introduce a relatively new phrase: ‘PrimaryBound’. This refers to the belief that every avatar is tied to one specific human. ‘PrimaryBound’ was assumed by Wallace Linden in the blog post ‘Will The Real You Please Stand Up’, when he talked about all the diverse online worlds and social networking sites an avatar might belong to, adding:

“[Here is] what all these online “identities” have in common. At the center of them all, the hub that ties all these personae together, is the very real, non-virtual, analog and offline “you.” Whether the connections are public or not, your Second Life avatar, your World of Warcraft toon, your Facebook profile, your LinkedIn employment history — all of these and more are just different aspects of a single entity: the person reading these words. They are all already connected to each other, via you”.

But we have already seen in the case of Hamlet Au that an avatar with a familiar name may not necessarily always have the same RL individual behind it. For ‘PrimaryBound’ to hold in all cases, the following statement would have to be true:

‘Avatar X is authentic only when that account is owned and/or accessed by one particular RL person’.

But, why should that be the case for a roleplayed character? The assumption here is that, of all the people alive today, there is only one that can convincingly roleplay, say, Extropia DaSilva (by ‘convincingly’ I mean the rest of the online community believes they are interacting with the person the avatar claims to be). Surely, though,  the billions of people alive today should be broken down into more catagories than just  ‘the one who can convincingly act the part of Extropia’, and ‘everybody else who never could pass as Extropia’.

A more realistic way of putting it would be to say, ‘of all the people alive today’:

Some would be excellent at roleplaying Extropia DaSilva.

Some would be very good at roleplaying Extropia DaSilva.

… and so on down a sliding scale towards ’some would be hopeless at roleplaying Extropia DaSilva.

Suppose some person from the ‘very good-excellent’ end of the scale were to login to my account or set up an account under my name in another online world or social networking site, and then they pretend to be me. Would the rest of the online community know this is a fake Extropia? Well, why would they? Here is an avatar called Extropia DaSilva that acts just as Extropia DaSilva is expected to behave. So what conclusion could anyone draw, other than ‘this is Extropia’? Also, remember that ‘somebody logging in to my account and pretending to be me’ is the default situation for a roleplayed character. That is always what is happening. Unlike an avatar used as a tool for communication, where you speak to the RL person THROUGH the avatar, when communicating with a digital person you speak to the digital person aka the character that exists in online spaces. While a person may believe they are speaking to the person behind the character, that is not the case in any meaningful sense because psuedonimity does not allow you to model that person in your mind. Think of Hamlet Au, and you visualize Wagner James Au. But think of Extropia DaSilva and you will visualize… Extropia DaSilva, because that is the only identity you get to know.


But if it is true that more than one person can convincingly roleplay a specific digital person, PrimaryBound cannot be true in all cases.  But there are still some hurdles to clear out of the way. For one thing, it is no doubt the case that ‘those who are very good-excellent at roleplaying avatar X’ are in a minority compared to ‘those hopeless at roleplaying avatar X’. If you were to grab some random stranger and ask them to log in and pretend to be me, it is a lot more probable that the person you chose would not be able to roleplay that part convincingly. Even if you were lucky enough to have picked someone excellent at roleplaying that part, the fact that hitherto they have not is another problem. They would have little to no memory of that character’s past interactions. Friends would realise something is amiss when ‘I’ do not seem to remember important events from a shared history.

So while in principle ‘PrimaryBound’ is wrong when applied to digital people, in practice it is not, simply because it is so unlikely that you could find a suitable replacement roleplayer. However, certain technological developments might converge on a solution.

The first trend is the growing proliferation of sensors. We already have the technology to record where we have been, what we have seen, what we have heard, what we have said, what we have read/written. There are also plans for future sensors that will be able to infer a person’s emotions, allowing a person to track their changing emotional states as they go through life, compiling extensive psychological profiles.

The second trend is increasing storage capacity and the move to cloud-based applications. We are nearing the point where the storage capacity exceeds one’s ability to fill it. This will make the prospect of having to delete old stuff in order to make room for new obsolete. You could save everything a person sees, hears, reads and says over their whole lifetime.

The third trend is the increasing capability of search software. If this trend should continue, we will reach a point where a person can easily recall anything specific that has been uploaded to the Cloud, be it in text form, or audio, or video.

What these three trends are converging on, are ‘digital memories’ providing total recall. That is, the augmented ability to recall, in vivid detail, every event that happened in our past. How, though, does this help the PrimaryBound problem?

For one thing, it would enable the search for a suitable replacement to be carried out in a systematic way. Here’s how. Take the digital memories of a particular digital person. Everything s/he has seen, heard, written, said, and all  emotional reactions to every event. All those digital memories are then compared to the digital memories of every RL person, searching for those that are the closest match. You can think of it as a kind of automatic casting agent that is searching for people who would be excellent at playing that role.

OK, so let’s assume a person who would be excellent at roleplaying that part is found. What do we do about the fact that he or she has little to no knowledge of personal details about that digital person’s past interactions with others? I guess you know the answer. Not only does the replacement roleplayer gain access to that character’s online accounts, they also get access to its digital memories.

The great thing about this idea is that it divides the task of roleplaying a digital person between the strengths of human and machine intelligence. Humans are likely to remain far better at creative and emotive thinking for the foreseeable future, so a digital person is much more convincing when a human is the pupeteer, rather than a bot. Computers, on the other hand,  are much better at accurate, repetitive thinking, and this will be especially true in the era of digital memories and total recall.

So, a human roleplayer provides the necessary qualities a digital person needs in order for other people to think they are ‘real’, and the Total Recall system would work in the background, supplying the right information at the right time so that, no matter what anybody else asks, the roleplayer is able to make that character respond in a manner that is consistent with what others expect.


One more question remains, and it has to do with motivation. Why would any person want to pretend to be a pre-existing avatar? Surely everybody would want their own avatar or to invent their own digital person? You would think so, but recently some bloggers have reported that their screen names have been adopted by other people. I talked about Hamlet Au already, but there is also Prokofy Neva, who said in his blog entry ‘Fake Avatars’:

‘The Prokofy Neva on Facebook isn’t me, but Tizzers Foxchase of the Woodbury goon squad’.

If Prok had not told me that the Facebook Prokofy is a fake and I tried to communicate with Prokofy Neva on Facebook, would I know it was not really Prok? Imagine that this Tizzers Foxchase says something like ‘Extropia you are a lunatic, fuck off and die, asshat!’ after I say something like ‘Hello, are you really Prokofy?’. In that case, I would assume it really was Prokofy. If I think an avatar is Prokofy, then from my subjective point of view it IS Prokofy, and it remains Prokofy unless I am reliably informed that this is a fake. That would happen either because this Prokofy consistently fails to act like I expect, or if somebody whose word I can trust tells me ‘that is not Prok’. Someone like the person who writes under the name ‘Prokofy Neva’ on the ‘Second Thoughts’ blog, for instance.

Whether a person would expose somebody else pretending to be their online personae or not is probably down to the level of seperation between the 1st and 2nd life self. In his book ‘The Making Of Second Life’, Wagner James Au occasionally talked about Hamlet in the third person as if he were a person in his own right. But I think it is fair to say that there is very little distinction between Wagner/Hamlet. As for Prokofy, there is one obvious difference between ‘him’ and the person who owns ‘his’ account. I will not spell it out here, but instead leave it up to the individual to do a brief bit of investigation via a Google search along the lines of ‘who is Prokofy’ if they do not know what I am talking about already, and care. Does this difference mean Prokofy is not just another means of getting in touch with a specific RL person, but is instead a character whose beliefs and behaviour may not be shared by the person roleplaying him? Personally, I do not know what the case may be but since Prokofy has never declared himself to be a digital person or roleplayed character, perhaps we should take that as proof that the RL person sees little to no distinction between the 1st and 2nd life selves.

But what about a person like my primary, who insists (not that everyone believes it) that ‘Extropia DaSilva and I are not one and the same person. She has her identity, I have mine, and the two are kept largely separate’. What would my primary do if somebody else was pretending to be ‘me’? Probably, she would not say anything, and instead let others decide for themselves if this really is Extropia DaSilva or not. After all, since that is a roleplayed character she cannot claim ‘Extropia DaSilva is a fake if somebody is only pretending to be that person’, because there is ALWAYS somebody pretending to be Extropia. Does that mean I am always fake? No. It all comes down to authenticity of performance. I am real so long as other people believe they are interacting with ‘the’ Extropia DaSilva. It does not matter who is doing the roleplaying. It only matters that they do a convincing job.


One last thing before the end.

What if a human being was not a necessary part of the system that enables a digital person to ‘live’? One futuristic possibility is ‘whole brain emulation’ or ‘mind uploading’. This is the idea that we will one day be able to scan a living brain in sufficient detail to capture the processes that give rise to memories, personal identity, self, consciousness. Those processes are then reverse engineered into a mathematical model running on a suitable neural computer. In simplistic terms, a person’s mind is copied. But what identity should we attribute to the copy?

Traditionally there are two schools of thought. One insists that the copy is a continuation of the same self. ‘I’ have been transferred from one brain/body to another. The other school of thought insists ‘the copy is not you’. In other words, mind uploading creates a different person. Both schools of thought have published long tracts defending their point of view, but I will not go into details here.

Instead, I will suggest a third way that exists between these two extremes. Imagine we copy the mind of Wagner James Au. Is the copy really him? Or really somebody else? Or, is it an identity that is sort of Wagner James Au and sort of somebody else at the same time? It so happens that such an identity already exists: ‘Hamlet Au’. Recall, for instance, his occasional reference to Hamlet Au in the third person, as though he were a person in his own right. If you were to non-destructively scan Wagner’s mind and then the software emulation ‘woke up’ and said ‘I am Hamlet Au’, everyone, including Wagner, would probably be convinced that this was the truth. But if the upload said ‘I am Wagner James Au’, the original would probably be faced with all kinds of existential ‘he cannot be me because I am me’ kinds of questions.

Perhaps, then, the identities people project out onto online spaces, which are always ‘kind of me but at the same time kind of not me’ (even if the person claims to be 100% the same in RL or 100% different), will turn out to provide a comfortable solution to the ‘duplicates paradox’. Perhaps the promise of mind uploading is not immortality, but rather the creation of ‘mind children’, an ‘other’ that shares your personal history, skills, and everything that makes you ‘you’, and who will take your life experiences and develop them further, diverging from your ’self’ and becoming more of a person in their own right as time goes by.


It seems likely that this will be a situation we will not face for a long time to come, if ever. Perhaps it is just not possible to upload a mind. Perhaps no technology, no matter how sophisticated, is capable of consciousness. Perhaps, even if it is possible in principle, achieving this goal will turn out to be too complex for humans to solve?  But, whatever the case may be, Vernor Vinge’s prophetic remark ‘the sense of self itself is in for rough times’ is already showing signs of coming true. We see it manifested in every concern that a claimed identity may be ‘fake’ and in the ways roleplaying and digital personhood thwart us in our quest for a quick and easy solution to the question ‘is that person who they claim to be’?

Print Friendly, PDF & Email
%d bloggers like this: