ߔߎ߯ߟߍ߫ ߞߊ-ߖߊ߬ߣߏ߬ߟߌ߲߯ߗߌ (Pule kaJanolintji) ngumcwaningi wolimi lwesiNtu nezabathwa, isazi somlando nemikhuba yokusetshenziswa kwezinkulumo zaseNingizimu, futhi umthuthukisi wezinhlelo zokuloba zezwekazi, njengesiBheqe Sohlamvu.
Pule is a scholar of linguistics and a cultural historian of speech practices of Azania/!Naremâb, promoting writing systems of the continent, like N’Ko or Ditema tsa Dinoko script. Their MA research describes South African cryptolects - secret forms of speech.
This conversation took place between i0 xen0 and Pule, who met in Zurich in 2022 through the School of Commons residency, where they were both fellows. All images used with permission from i0 xen0 and Ubungxenye.
1—THE PROXIMAL & THE
The proximal archive is that which is close at hand, the grand narrative stuff, the more hegemonic networks of information; and the distal archive is where one can find traces of information that can lead into deeper kinds of knowledge, more marginalized or minoritized kinds of knowledge. For example, there is indigenous knowledge online, to some degree; or there are, at least, traces of it. [Anticipating the archive means] anticipating what the proximal archive will be, [and] making an educated guess of what the distal archive might have traces of, that the algorithms could pick up on, and then maybe finding some gems and hidden troves that might suddenly pop up. It’s like a treasure hunt, in a way, you know, but a treasure hunt without a map.
I can give an example. I work with the principle of using the Southern African writing system, Ditema tsa Dinoko, to imagine a commentary on a particular issue that I can generate an image about. Working with the idea of decentralizing home, the idea of Africa around the world, or indigenous knowledge systems in Africa, imagining if they exist around the world, and in a thought experiment. Traditional mural art in southern Africa has its symbolic forms, there are geometric symbolisms and color symbolisms that are embedded in mural art. And it’s a tradition of an ephemeral mural art, because it’s often done with mud on the wall, so it’s lost after the first rain, it disappears. And this kind of trace, or mark, of someone’s actual hands on a house, and imagining that kind of symbolism, and that kind of practice, say, for example, in the houses around Lake Zurich.
So I knew I was going to be coming to Zurich, and then I said to myself: Okay, imagine if there were traditional Sesotho houses with mural art around. And so the process then, of guessing, is to see whether words like “Litema” might be picked up by the algorithm. I was using Midjourney, which is one of these image generating tools, and it works through discord. And you just put in a list of terms: genre things, style things, terms like “4k” or “hyper-realistic,” terms like that.
What, if I wanted to see those houses around Lake Zurich, would I need to tell this algorithm?
And so you sort of just have general ends, like “African mural art,” you know, things that are in the proximal archive - terms that you wouldn’t really like to call in; like getting the thing without liking the terms. It’s like a means to an end. Then also just trying some terms that are maybe slightly in the distal archive, some terms in an African language, that kind of thing.
2—MELTING & SYMBOLISM
What it often does with Midjourney is that it creates something that is slightly off center, one image melting into another in a way that I find quite satisfying. I don’t like the ones that are very well-defined. Rather, you’re not quite sure what’s going on; like a décalage, a mismatch in the orientation of the images or something. So the houses have that, on the sides of Lake Zurich, with an impressionistic version of African symbolism, something where the colors are flowing and melting into each other, something with the flavor, but not exactly. And then to reinforce that with the actual symbols, I put it in the sky above the houses. So the process is to generate the image, and edit it. When you put the words in the sky, then it will reflect on the lake. It’s sort of a commentary about what is the archive? And what is the possibility of imagining using that archive, with the AI as an imaginarium tool.
This connects to our project in the School of Commons, because it’s this idea of words as nodes in a network. When we work with those Wikipedia-visualizing tools, it puts your mind in that way of understanding how you can move perhaps more accurately, or more imaginatively, guessing what are the constellations of nodes that might be there, and then drawing out the image through that. Besides working with images, I’ve also been working with text recently, through OpenAI ChatGPT, which is like a discourse engine: it’s going to talk to you and give you information. And there they claim that it’s in all languages, but obviously, it’s in all languages that are in the proximal archive, it’s gonna have very little of the distal archive, especially when it comes to language. So I was making it write poetry: you put in a command, asking it to, say, “write me a poem in Zulu language”.
In the generated isiZulu language example, it actually did quite well. It wrote a poem that was fairly sensible and actually was quite evocative; pretty short little verse, a couple of lines; which tells us something about the archive. Now, Duolingo has isiZulu on it, and beyond that there is an online presence of that language. Of course, for some reason it “decides” also to give an English translation, which itself is really another poem, not really a translation. But then when I try Sesotho, it will do something just like saying, “I I I I I, Me Me Me Me Me, Mother Mother Mother Mother Mother” in the English “translation”, and the “Sesotho” text that precedes is just iterations of grammatical elements tenuously related to the sense of this. When I saw that it was like, that’s not a bad Dadaist poem! And it really molds into that whole idea of dada, because it is about chance. With the materials that exist, what is its chance? Just a flip through a dictionary, and they hit on a word. Then I tried different languages and the languages of South Africa, so it’s located for me. I start with guessing what’s in the proximal archive, so it will be Zulu; and then go to Xitsonga forms which are more of minority languages, and then go further and say Khoekhoegowab, at 300,000 speakers or so, and which has a high prevalence of click consonants in the language (The name “Khoekhoegowab” proximally refers to the standardised language that is meant to accomodate for ǂNūkhoegowab, Haiǁomgowab, and Namagowab - so it is the most proximal one can get when one really wants, say, Namagowab). And then I try even a seriously marginalised and endangered language, like Nǀuuki, which like those mentioned above, is also a language that has symbols in the writing of it that are not so easy to work with in terms of computing, because you have to actually find special Unicode symbols - the click consonant symbols ǀ ǃ ǁ ǂ ʘ. Symbols like ǀ and ǂ resemble keyboard-ubiquitous characters like / (a slash) or | (a pipe), or # (a hash). Often when people are writing in a language like Namagowab, which has these symbols, they will just substitute with a hash or a forward slash, and of course, those things cause problems online, because they are operators in HTML. So that contributes to systemic exclusion of that kind of archive. That archive is itself fraught.
For example, when I put in Khoekhoegowab - the name of the standardized language - Afrikaans words came out! It was quite a poetic means of an automated commentary on what the archive says, about what the social status quo is, with regards to culture - language and culture. In that case, I didn’t expect what would come out, because I’m not so used to ChatGPT. I didn’t expect anything that came out, and it was quite illuminating, because also, the words that it selected were quite potent. So, I mean, it goes, “I I I I I I, mother, mother, mother, mother, mother, mother” and then it will give another word that was seemingly random. But that word, of course, in a poem form, suddenly becomes quite potent, you know, because you find the meaning in it.
The machine is not finding meaning, it’s just finding the pattern in the characters. It is an artificial approximation of a meaning-making process, which can be much different to what we might have imagined otherwise. So I can see it as a very sharp tool to do different things that you might have thought you could do.
Some of your work involves Wikipedia visualization tools; in the two School of Commons events, you invited us, the other participants, to try to guess a way to connect conceptual nodes using a Wikipedia visualization tool. So we would try to make a bridge between two ideas, connecting them based on what we think might be mentioned and linked in the first paragraph of a Wikipedia article. So in that exercise, you invited us into that anticipatory engagement of the archive. What do you find that people get from that, or hope that people get from that?
When you are explaining a topic that maybe a minority of people have engaged, most people have at least a surface level grasp of it - they might have heard a term here and there or have a general idea of it, that you have to guess the extent of; so you are already trying to anticipate the archive in the people. It’s a way of explaining things, it’s very connected to teaching and learning, and just ordinary conversation. You’re anticipating what someone’s knowledge network is going to be. Sometimes you make a mistake, but if you’re shrewd about it, and if you have a broad social and cultural base (if you are diverse in the way you interact), you begin to learn a broader network of how ordinary references are connected. We all do that, but especially when you’re honing in on a kind of niche topic, you know, you learn the network that, probabilistically, might exist between your conversation partners. Not for sure, though. You can make a mistake, and someone can surprise you, or it can do something different in the conversation.
I like to do it quite polemically. I think I have often done that with a polemic, where you put it in terms that might be kind of controversial, and it is likely that people have a certain idea about it, and then you’re going to show how that concept is slightly differently arranged or can be slightly differently arranged, or even dramatically differently; and what is this conceptual network of what people assume in comparison to that of your intervention that brings in a differently arranged network. And for me, that’s an exciting process; it’s an intervention in that path of guesswork, using the unexpectedness of a connection, as an intervention.
I can give a brief example. I often like to bring up the idea of the “Afrikaner” identity, and the Afrikaans language in Southern Africa, you know, I mean, a lot of people around the world are familiar with Afrikaans, Afrikaans language, and maybe even Afrikaner identity. But people have, in a generalized way, an awareness that it’s “a daughter of Dutch”, or related to Dutch, or the result of Dutch colonization. But then I’ll make this intervention to say the simple fact that “Afrikaner” is a word that literally means African. The historical fact is that it was never related to anything Dutch, it just has a Dutch linguistic form, that is, it has a sort of Dutch grammatical base and vocabulary, but culturally and in terms of the politics of it - the historical politics of it - it came from indigenous and slave communities. I will drop a very little node, and then I will drop the next node, and the next node will be: did you know that Afrikaans was originally written in Arabic script? Because most of the slave communities were Islamic communities. Then I might get to the point of saying that it wasn’t “the Dutch” that colonised the Cape (proximally “the Cape of Good Hope” in colonial terms, and distally, “!Naremâb” in indigenous terms), but rather, the world’s first corporation, the VOC (Vereenigde Oost-Indische Compagnie “United East-India Company”), which is still the largest company to ever exist, worth around 7 trillion dollars in today’s terms, adjusting for inflation - many times larger any national entity like the Dutch crown, and in this way, point it out as the forerunner of all corporate colonialism that still happens today. So I am anticipating the proximal archive, and I’m making a little detour, or a shortcut, into a distal archive that can be illuminating.
As I explain this to you, I didn’t really think of this while I’m doing it. It’s just intuitive. But now that I’m explaining it, I see that I actually do have kind of an agenda here. And it’s an agenda of re-centralizing a distal archive, which has traces of what would sometimes be called “hidden histories” that have been covered over. It’s a process that happens in ordinary discourse, you know, in conversation or in written discourse, where you know these networks already exist in people’s minds. What exists in people’s minds relates to what exists online. What is the archive mentally, psychically with people? And then, in what way does it connect? Where does it disconnect from? What is online?
What really stood out to me when I first did that exercise with you in the summer was that although I already knew that Wikipedia is a specific knowledge system; it’s not the universal knowledge system, it’s a knowledge system. But it was striking how extremely obvious that was in the exercise, when we actually put the nodes on the screen and tried to guess what they connect to. It takes the apparent universality of something like Wikipedia, down to the realm of something that’s much more relatable, and that helps to reveal a multiplicity of knowledge systems. Is that also part of your agenda?
Yeah, absolutely. I mean there’s the mythology, and then there’s the actual data. And mythology can be data, but mythology is not spoken. It’s an unspoken system that’s behind the scenes in the way we interpret the world - personal mythology; things that are never really surface, but they’re a tacit assumption that we work with. Whereas when you visualize it in the actual online archive, then you can start to see that there can be a mismatch between what the mythology is about something, and what actually is the data about it. Mythology is a subjective system, especially because it’s not often brought to light; people don’t have to state their own mythology, it’s very subjective. This [practice] of network ontology visualisation objectivizes the conceptual networks, and it’s also a feedback process of how that mythology interprets those connections in the data, and how that data is constructed out of mythologies that might be there, although there may not be straightforward links between things. Often, it’s an incomplete picture, hazy, pixelated, or like the file is corrupted. It’s got a lot of noise in it. In that noise there’s a degree of entropy that doesn’t exist in personal mythologies. Personal mythologies are quite well structured, I think, they retain that structure mentally; whereas when it actually goes into “real” archive, especially through digital means, entropy is there.
At one point, when you were sharing the IXam traditional mythologies with us, you started to use some AI-generated images as illustrations. You used multiple images for the different kinds of mythological concepts and mythological beings, and when you showed them, I remember you said “it looks like this, and it also looks like this, and it also looks like this.” [This use of imagery was an element of a storytelling experience that had a lot of other elements, including song and movement, which created a deeply moving container, in which we then encountered the images.] I thought it was really magical, the way that you had done it. How did you arrive at using image generation in that way?
Thank you so much. I think, in general, the process of art making is a process of selection. You’re making a choice at every step as to what you’re going to keep; either which direction or which material you have to keep, or what form you’re going to keep and what you’re not. I mean, on social media right now, people are complaining about AI saying “this is terrible, the end of art”; it’s a debate that suddenly hit a meme level of engagement. My comment always is, you know, this is a tool, like any other tool that has been created over time. When photography was invented, you probably would find exactly the same dialogue happening, as with each new tool, an outcry from certain sectors that sort of circulates and becomes the zeitgeist of it as being a terrible sin.
In the process of selecting those images, I wanted to get, for each of these mythological figures, something that could signify or represent them. In the IXam cosmological narratives, there are three couples; each of those has a hard material element, and an ephemeral element; and then each of those has a daughter, which is a phenomenological element - an experience. And so I wanted to have a very clear way of showing that system, because I think that’s quite an elegant interpretive system, that’s in the mythology. For example, the rain is meant to be an animal, a livestock type animal, but also wild. So it might be a hippopotamus, it might be a rhinoceros, it might be maybe a cow, a bull - a bovine; any kind of large animal that could be seen as something that could be herded, you know, because the mythology says that the rain is herded by the dead. The diviner speaks to the dead and the dead herd the rain towards the people, so that the rain can follow, in the system of rain-making.
It was a process of putting into the AI generator on to, first of all, the setting: it would be a Karoo landscape (Karoo is a biome in southern Africa, with the word likely originally being !ʼÃukarob “Dryland”). I add the prompt, “a cloud that is a rhinoceros.” Then I get back to this idea of making a selection; it gives you the number of options, and I make 8 or 12 different versions of that. And then you select what is interesting. I selected three images of the rhinoceros turning into a cloud, which I thought made a kind of sequence, you know, where one was much more more cloud-like which had a cloud with a sort of animal form. It was a way of seeing into the cloud, the animal. And then one which was much more animal-like, but not rhinoceros, like some animal, which perhaps had a strange horn that has a tree-like structure. Then the last one was more or rhinoceros-like. For me, that was an interesting trio, because it’s different lenses on that idea, three very interesting lenses into that concept. And what that does for me as a practice, and what I hope it does for those who see it, is to also realize that this concept, within quite a specific interpretative system, has a very great diversity of possibility in that interpretive system. There’s an infinity of ways that concept can manifest, it is not this very clear thing of “a rhinoceros looks like this.”
It’s the metaphoric possibility, in the visual metaphors that happen by chance. We are going to find the meaning, we’re always going to find the pattern, and we quite often find patterns where there are none. I often like to talk about apophenia, which is this idea of finding patterns where there are no patterns - we’re going to find the pattern, no matter what you do, if you framed it (and even if you haven’t) someone is going to find something in it.
4—”DON’T BLUFF YOURSELF
INTO THINKING YOU’RE THE
ORIGINATOR OF STUFF”
Are memes an archive?
Memes for me are oral history, so clearly. In African historical practice, you know, we always have to take into account oral history, because the written histories have very short lives. We’ve got very few written histories that go back any way. There are, of course, but they are very metaphorical. They look not like straightforwardly descriptive histories in a Western sense; they require a high degree of interpretation in order to get into that archive. And that goes towards the principle of orality, that oral history is the most important way that we understand history in many places in the world, and definitely in southern Africa. We interpret history through a principle of orality.
And for me, memes are exactly that. It’s like, you have to know that thing. And that thing is known because of conversations between people, because people are saying these things in a very discursive way. It’s all about a conversation, somebody’s responding to something else. It’s not like a long response; it’s very pointed. We all know conversations which are just a series of memes that actually develop the theme. And the other thing that I find very interesting, working a lot on Facebook groups is that you have, in niche groups on Facebook, up to five or ten community memes that are canon of that community. Somebody who suddenly joins a group without knowing the history, that might go back a number of years, doesn’t know that meme.
For example, there is a group on Facebook called “Entomemeology”, it’s a shitposting group for entomology. There’s a meme on there, which is all about “brown recluse”. It came out of a joke that in the United States or in North America, people have this stereotype that everyone who doesn’t know about spiders thinks everything is a brown recluse, because it was known to be something dangerous. It’s now an old meme in that group to just call everything a brown recluse, or everything a variation of “recluse”, and it goes into all its diverse manifestations. Someone coming into that group can be quite baffled by what’s going on, it takes quite a while to learn the oral history of the group. And it’s interesting, because it’s all taking place in a written and visual medium, but it’s oral history. It’s not a written history. I think it also calls into question, what really is the written history?
I think so many popular ideas over the millennia are just memes. It’s only now that we sort of identify this terminology and say “this is a meme,” but I think this is the global practice of orality, as a communal archive. It’s a communal archive that has pockets of specificity and pockets of generality. Sometimes things that are quite specific - there’s a number of memes that are very famous that might come from a regionally specific soap opera in some part of the global south, for example. Like the meme “Am I a Joke to You?” is the character Tau from Generations, a soapie which itself is a meme reference for any South African; or there’s the famous meme where the person is trying to figure out the maths problem, and that’s from the Brazilian soapie, Senhora do Destino. But typically only South Africans would know that the person in the meme is Tau, and have the whole background of that character as well; and typically only Brazilians would know that it’s Nazaré Confusa and not just “Math Lady”. There’s sort of like a gap: some people know the original reference, and some people don’t, so there are levels of depth of the orality. That can happen in textual discourse, in long forms of prose and in the literature archive, but I think it is much more potent and much more salient as the organizing principle in a communal archive, which is a meme archive.
In terms of memes as a communal archive, does that archive also have a proximal and a distal component to it?
It’s a different organization system, really, because it doesn’t have linear form. By nature, it’s not going to be linear, it has this radial form with different levels of Depth. When you’re telling a narrative in a linear way, which a lot of literature (especially in nonfiction literature) is, there is a linear structure. linear structure. There is a very small minority of nonfiction that jumps around and proceeds in a rhizomic way or whatever. Because of the nature of orality in memes, I think it’s a difference. Probably, if you visualize it, it would look totally different, like a different organization system in the network ontology diagram. When a meme has become famous, there is a general level of understanding of that meme, but then a specificity of whether you know this actual soap opera actor, but that’s besides the point of the meme; It’s just extra information; an extra level of meaning that is there. There is more to investigate, but less that is hidden.
I think a lot of linear narrativizing, either intentionally or just because of the structure of telling stories like that, hides a lot. It’s not just like there’s a lot of info to get through, but a lot is actually hidden under concepts that are unquestioned; hidden under placeholders that are unquestioned (I also think this is not what this kind of narrativising portrays itself as - it presumes itself transparent). With memes, it’s much more of an active communicative process and less of an archival process. It’s less of an intentional archive; literature is an intentional archive, whereas memes are more of a de facto archive that develops in a more organic way.
If memes are not an intentional archive, they don’t really do that abstraction work. Abstraction is the mechanism of hiding things behind concepts. They don’t do the conceptual collapse of deciding, “oh, all this stuff goes into this one node.”
I think a very important thing about memes is that there’s a low priority on authorship. Authorship is irrelevant to them, but there’s such a high priority on it in much of literature. It’s about “who’s the author?” And being able to cite to the author. Whereas, you’ll hardly ever know who made that meme; if you researched a lot, you might find out. But it wouldn’t even matter, it’s beside the point. It’s not about who made it. It’s the communal point of reference, it has no authorial authority. I think that a very important, strong point about meme culture is that authorship is clearly not a high priority. It’s not part of the terms of the system.
I don’t like authorship, which is often connected to ownership, and also authorship I find to be a kind of dishonesty because I don’t think you can create anything original or anything that is your personal voice. I also think that the idea of a personal voice is a myth. Everything, one can say, is a communal practice… everything is a communal practice; all mark-making is communal; any sign-making is communal, because the system of communication requires both the sender and the receiver. Any communication is by nature communal. There can never be a singular author of something.
The kind of work we’ve been doing in Johannesburg, with the VR space at SodaWorld Studios where we have a VR and in-the-flesh crossover event. We’re trying to make a DAO, a Digital Autonomous Organization. It’s not in practice yet, but the idea is that we would have a social token system in the community. The community is basically peopleto see a live music venue with multiple art forms succeed, and everyone who participates is assumed to be sort of behind that common mission; we all contribute to things taking place in that space. You can earn social tokens in that system even just by showing up to the event, because by showing up to the event, you’re making the event. So it’s not seen as about you showing up to the event and you are now going to pay to receive the art that somebody else has created, and they’re the author and you’re the audience. Rather, we are all creating the thing mutually, and this is a way of being honest about it and saying: “You can contribute to this by attending it”; somebody else is going to contribute by bringing chairs; someone else is going to be working at the bar. Those are all different things that make the event, and you earn tokens according to your contribution. Then, everyone owns any profits that come from the products of the event, and all the things around the event, according to their contributions. It’s a socialist way of organizing that system of communication, of art-making as communication, and cultural events as communication. That feels to me just more honest about how the thing is happening.
There is an older outlaw (Bra Maxi500 of Sobantu/Sophiatown) that I interviewed in my research, he said to me: “you know, you mustn’t bluff yourself into thinking you’ve come up with something original. History repeats itself.” He was very clear about that - don’t bluff yourself into thinking you’re the originator of stuff. I think of everything that was a very wise statement by someone who’s really, really seen a lot.
Meme-making is just an ordinary process of social discourse. Intentional archiving is not the baseline of how interaction happens - it has a different intentionality, but it is majoritized as what the archive is. Whereas really, like, the majority of archives are these unintentional archives that are oral. That’s the majority of all archive; intentional archive is only the tip of the iceberg. And it is so clear with Wikipedia - there’s so much unintentional archive; we have a mental map of what the archive is, and it’s just not like that in practice.
I think any act of intentional archiving is imposing an agenda without necessarily inscribing the agenda into the object itself. So that part remains hidden from view.
The structure of the archive is not stated, and it’s not declared. It’s collectively assumed, and because it’s collectively assumed, not everyone has the same idea of what it is. Not everyone has the same idea of what the archive is, but we all have a sort of mutual agreement vaguely as to what it is, but it’s unspoken, it’s undeclared. I think the intentionality of it is an assumption that everyone has that the archive is intentionally created and has authorship. But the majority of the archive is not authored, or it’s collectively authored, or it’s authored by its own software.
I was thinking about authorship recently with respect to taking responsibility, and asserting power. For example, corporate archives, and documentation within organizations. Who is taking responsibility for the way something is done? Where can I point to, to have proof that I am allowed to do something?
I think being allowed to do it also goes hand in hand with what become the dominant terms of interpretation. I was doing research into legal documents from the British colonial period in South Africa, and the narrative that is dominant of how Identity is in South Africa, for example. You can find a trace of it - the foundation of it in these legal documents, because the legal documents got perpetually reproduced and it’s the status quo, so it becomes just by the nature of it being legal, and it having to be enforced in legal terms - that becomes the framework for the majority narrative of how identity manifests (for example, what being example, what being “Native” means, as a colonial imaginary). Everyone starts to believe that narrative because it’s the status quo. There’s a general assumption that that’s what exists, to see certain identity terms coming up in the legal documents over the 1800s. Some identity terms were actually introduced in that legal document, which now are considered set in stone, but were clearly created in that period through a legal inscription. We could find the authors, but it’s also not really relevant who the authors are; the only relevance of it is that it has a juridical function. It’s legal. That’s what makes it have power.
I’m still looking at this statement, “more to investigate, less that is hidden,” and in that example, although we can investigate the authorship, it just doesn’t matter.
It’s like an abstract author, the author is The Official. Doesn’t matter who it is. All you’re doing is uplifting the terms of the authorship in the archive, it’s never uplifting an author as an individual. And I think that’s maybe why people get so competitive, because it’s really not about them. It’s just about the points they have, in terms of what is legitimized.