Really hoping that proliferation of instances will lead, in time, to divergence in modeling (at the software level) the parameters of "community" and sociality.

GNU Social/Mastodon shares features with SMTP, IRC, Usenet (surprised not to have seen that mentioned by others), with BBS networks before that, and even such antiquated systems as postal newspaper exchanges. In this history, "social media" giants are distinct only in their resort to centralization.

The organizing model is unchanged.


What GNU Social/Mastodon may afford is a decoupling of network application from network effect. In particular, Mastodon builds on the network effects set down by GNU Social, in the same way that gmail and MS Outlook build on the network effects of SMTP.

This means that the network application instances can diverge, in how they model and render the underlying network, how they integrate and leverage tools and tooling, so to present different use cases for participants.

My fervent hope is that some of those divergences fundamentally question the model of network "spaces", this itself a derivative of the commonly held notion of a "public sphere" (per my reference above to postal newspaper exchanges).

My suggestion here is that the episteme, the condition of possibility, if you will, that informs our discourse on what and how so called "social networks" can enact relationships between human operators of computing devices is deeply anchored in an earlier epoch.

One defining property of our prevailing model of "spaces" is that of embodied membership. An individual is either "in" a space or not. A "space" is understood as a container within which some things are and other things are not.

The fraughtness of this model is already evident to anyone who ever described idiomatically as "having a foot in both camps".

Nonetheless, this spatialization of membership if familiar and easily reduces to a logic of truth values, a point we shall return to.

Yet our world is full of cases where embodied membership is an inadequate explanatory device. Consider a person born to parents of different socially constructed racial categories. Is this person a member of one category or the other? Both? This is quite too simplistic a question.

Indeed, how their memberships manifest on a census questionnaire may differ from how they manifest at a family reunion from how they manifest during a workplace interaction. It's not a universal "am" or "am not".

We can't save embodied membership as an explanatory device here—by resorting to the kludge of "feet in both camps"—because biraciality (or indeed, race, qua race) doesn't play out the same way being, say, both a member of the varsity football team and a member of glee club does.

There are no sign up sheets that definitively determines across all social encounters that one "is" a member of one social construct or the other or both. Membership, here, can't be reduced to values of "T" and "F".

This is a conceptual frame that many white people (and, especially, white men) may struggle with. There may not be sign up sheets, but there are hegemonic social conventions that define certain memberships as givens.

But conventions are always anchored in a model. And the model, here, is simply inadequate, no matter how well it may serve some. It is inadequate in society, and are likewise inadequate as a basis for social media technology. Yet here we are, using the same fundamental model.

Quite simply, our model of "spaces"—the basis around which our social technology is organized—insofar as spaces manifest through embodied membership, is inadequate to describe the fraughtness of social reality.

What if we, instead, conceived membership not as an expression of a state machine, but as a bundle of affinities and exceptions, which play out within and between interrelated membership bundles? Provisionally, let's refer to this conception of bundles as "intersectional membership".

What if our social technologies were built from a model (or models) where such intersectional membership was a fundamental property?

Mastodon, thus far, is not modeled on intersectional membership. Indeed, stated aims and intentions of the community notwithstanding, both the implementation and the ethos of Mastodon would seem to proliferate embodied membership within the context of the same commonly accepted "spaces" model as previous social media technologies.

William Shatner is or is not on Mastodon. I am or am not on a given instance. You are or are not someone I follow. I am or am not someone you follow. Individuals of certain political leanings are or are not encouraged to join. At some point, instances and individuals will or will not be listed on a whitelist, or a blacklist, or a greylist.

All easily reducible to truth values. All as easily worked out as is the assertion "You have paid your annual membership dues to the Audubon Society."

Now some will interject, at this point, that this is a function of social networks being run on computers. Computers are binary, goes the common sense argument, so of course our computerized representations, our models, will be binary.

Only, that same mechanism, binary representation, is used every day to represent floating point numbers of arbitrary precision and to represent so-called "enumerations" of terms that reduce to numerical representation only as an implementation detail.

It isn't that our computers are binary, but that our model—a model that worked just as well for the Massachusetts Audubon Society in 1896 as it did for Twitter in 2006—is founded on a property of embodied membership.

This property of embodied membership, in turn, depends in its modern expression on a representational system formalized by George Boole in the 19th century, although the model has been with us informally since well before then. Ingroup/outgroup long predates symbolic algebra.

So the question is, how might we begin to rethink the model of "spaces" that we so take for granted?

Can we imagine models of sociality with properties of membership that don't reduce in the same way as does embodied membership, to mere truth values? Can we conceive a non-spatial social media?

How would we go about designing social technology grounded in such alternate models? How would our tools/languages need to change to support this? How might the use cases of such applications play out?

(I have definite ideas on these points, but I'm posting these questions more to encourage others to start thinking critically about how we model the social, rather than to offer my own proposed solutions.)

@beadsland But there is a fundamental difference between email and GNU social. GNU social allows for completely different kinds of social networks, with some common features (that's my understanding at least). Every consumer-facing email application I've seen hasn't strayed too far from what everyone knows as 'email', even if the protocol can be useful in other ways.

@dgdas9 There are two points here. First, in what way are the social networks in one instance totally different from the social networks in another? I would suggest both fundamentally build on the same underlying model of universal addressability, for instance.

Second, consider how you can *do* email in one application versus another. Gmail categories and Outlook calendering offer markedly different experiences of message flow. Granted, underlying models were not challenged, yet they diverge.

@beadsland Let's pick a few social networks, such as Pinterest, Twitter, and Facebook. I don't know of the constraints of GNU Social, but I'd make an uninformed guess that it'd be able to 'copy' all of these platforms in an acceptable way. And, while these networks build on the fundamentals of human interaction, they are very different in nature.

Email is not. The core product, the paradigm of email communication, is fundamentally the same across apps, though I'm not very knowledgeable here.

@dgdas9 I agree, all three networks you name could be emulated over GNU Social, just as Twitter already has somewhat been in the form of Mastodon. Facebook would be the most divergent, as it would require degrees of separation to be incorporated into the model, but it could be done.

The reason such emulation would be possible is that these three platforms are, at their core, very similar in nature. Not that they build on fundamentals of human interaction, but on a shared model of sociality.

@dgdas9 Likewise, the core of email is not a "paradigm of email communication". Rather, SMTP via TCP/IP recapitulates models that were first developed by pre-industrial societies to successfully transport letters and packages across geographic distances.

Email was not a new paradigm, it is merely a new implementation of pre-existing network and transport models—only abstracted and formalized by OSI and CCITT (i.e., Open Systems Interconnection), very late in the game, historically speaking.

@beadsland It was a revelation to me when itunes used tagging to organize music, rather than the rigid folder structure I'd had with my collection in winamp.

Breaking identity down into an ambiguous, ever-shifting collection of 'tags', rather than a rigid codified hierarchy, would go a long way towards this end.

How does that look, in social media?

@mykola Yes, yes, YES!!! This is a beautiful example! Tagging as a divergence from folder structure! Wait, you mean a song can be in more than one category at a time? And not simply as a hierarchical breakdown of Artist > Album > Song? This is so obvious, and yet for years our mp3s went without.

And yet, it wasn't entirely a new idea. It's really no different in that respect from the "subject" categories of card catalog systems in physical libraries. The model existed already.

@mykola As for how this looks in social media: We already have a tags system in the form of hashtags, but that's static, fixed in amber with the authorship of a post. "Lists" are sometimes used hackily as second-hand tags.

What if instead we had folksonomic tagging, where individuals could not only "boost" but mark posts with their own tags, and even second the tags of others? Each post might accrue a tag cloud as group commentary: a collectively aggregated marginalia.

@mykola As for tagging identity (rather than posts), as you suggest, that's even a more promising avenue to explore.

Imagine a network model where profile tag clouds interact, such that resonances between clouds can aid in discovery of like-minded individuals, not only directly (X has three tags in common with you), but indirectly through folksonomic marginalia. We might even imagine discovery through tunneling of resonant tags across degrees of separation.

@beadsland @mungojelly I am too busy wearing my coding hat to put on my theory hat just now, but will follow up on this when my various fires are out. Thanks for linking me!

@beadsland As a programmer and (aspiring) social media architect, I'm very interested in this, but I'm a long way from having any actionable understanding of this.
I'm just gonna throw a few fish at a wall and see what sticks;
If I built a social server, that was capable of communicating many different protocols, and the user could for any content choose to share over any subset of them, would that be closer to a non-boolean network?
Or would it just be a more complex one?

@zatnosk The wall may get a bit stinky, but we can always hose it down later.

Your proposed scenario sounds something like Friendica. This would be a further proliferation of booleans, just as federated instances within a single Mastodon network proliferate booleans.

You don't make a fruit salad by mixing a wider variety of vegetables in a bigger bowl.

@beadsland Hm. Ok. I suspected as much.
What if we (and this is a fish I've thrown before somewhere else) had "sharded identities"? A sharded identity would be multiple accounts on different servers representing (or even on same servers) that would all point to the same identity. Just like we have different personas in different contexts, we would have different shards in different social circles.
Then content could be shared by me as any subset of my shards.

@zatnosk *ducks as another fish slaps against the wall*

This scenario sounds like a structural inversion of (yet, nonetheless functionally identical to) Google+ Circles. Another variation that begins from an axiom of sociality as containers.

To approach this, we need to think less Jean-Paul Sartre, more Annemarie Mol.

@beadsland Hm, you're right, it is partially inspired by Diaspora*s aspects / Google+s circles.
I'll admit I haven't read either of those directly, but only Sartre am I somewhat acquainted to.

@zatnosk Your appeal to social personas put me mind of Sartre. It's a terribly flawed philosophical frame, IMHO, that pretty perniciously permeates common consciousness, and as such will be little help in thinking alternative models.

@beadsland Then I might need new philosophical frames to come up with non-boolean concepts that can be modelled as a social web.

@beadsland do you have any suggestions of philosophical frameworks that I might be familiar with and are usable for inception of new fish?

Show more

@beadsland these sharded identities can be publicly presented or just reserved for those I choose to show them to. So my public profile would display a mix of my publicly visible shards, and after connecting with someone I can allow them to see more of my shards (and they can of course opt-out of seeing them if they want - a partial mute).

@zatnosk *scrapes some fish scales from the wall*

What you've done here is proliferate booleans in the same way dropping a piece of glass on a hard surface proliferates pieces of glass.

Sign in to participate in the conversation

Server run by the main developers of the project 🐘 It is not focused on any particular niche interest - everyone is welcome as long as you follow our code of conduct!