LinkLanguage is a special type of language, that you get when you use wiki.
Link-language is made possible by:
LinkLanguage permits fluency in link-making.
You can think about all the WikiWords we put together on a wiki, that CommunityMembers get good at “speaking,” and call that the “LinkLanguage.”
That said, we have some particular things to say about LinkLanguage, and so we want a more precise definition. Also, beware! This phrase appears to have some WordMagic to it. Make sure you understand what we really mean by the phrase, before using it, lest you say things that the links betray!
A plausible abstract definition:
Let’s explain what we mean by some of these words:
| “signs” | On most Wiki, the “signs” are CamelCase words. You use a CamelCase word, you automatically get a link. On MediaWiki, the signs are phrases surrounded (in the source code for the page) by double-brackets. Conceivably, on some medium, (perhaps an SvgWiki?,) the signs could even be icons or images, pulled (perhaps) from a toolbox, looked up by either name or tag, and automatically linked. |
| “in common” | Several people have to be using the signs, in order for the LinkLanguage to be a “language” of interest, at least by this definition. |
| “fluently” | The signs have to be available for fluent use by the people using the signs. If a wiki page name is “Foo/Bar/Subpage/Baz”, the page name is likely not part of the link language. |
| “automated linking” | Linking must be either fully automatic (like CamelCase on OddMuse,) or semi-automatic (like “you have to enclose it in brackets for it to link” on MediaWiki.) |
| “more information…” | “More information” is a page that describes what the sign (word, phrase, icon?, …) means, what it refers to. |
If you miss even one of these elements, you no longer have what we care to call a “LinkLanguage.”
So, for example: “no computer, no LinkLanguage.” (Unless you have some other magical way to automate linking.)
Suppose you had a large projection screen, in your living room. And suppose that whenever your friends and you got together to talk, you did it in your living room. And while you talk, you’ve keyed it so that certain phrases always trigger certain pages (about what the phrases mean) onto the large projector display.
You would be speaking “LinkLanguage.”
Now, let’s say you had a wiki, but all the CamelCase phrases links to some page based on statistical correlations, (as per PlainLink,) or something like that.
If the CamelCase phrases don’t usually link to pages that tell more information about what the concept means, then you don’t end up with what we’re calling a LinkLanguage. (But if it does, then we do call it LinkLanguage.)
Let’s say- you start manually linking phrases of your own to URL’s.
[http://example.net/ blah blah blah blah blah]
You keep writing it, putting it in, by hand. Is it LinkLanguage?
Probably not; It’s not very automatic, and it’s also not in common. The idea is that this is a language people are communicating with, that exists in common. When you do it by hand, it makes it very hard for other people to pick up the language. So, “no,” that’s not LinkLanguage.
Question for consideration: Are the CreativeCommons licenses a LinkLanguage? (Don’t go fishing for a “correct” answer; We don’t really care if it is, or is not. We’re just stretching imagination here.)
Let’s get right into it:
(A) People can look up what you meant. When you refer to something not only by name, but also, simultaneously, by link, you make the meaning of what we’re talking about much more explicit.
So for example: We say, “CommunityExpectations,” or “ContentAndCommunity,” or “WordMagic,” or “LinkLanguage.” Those who know us quickly know what we’re talking about. And those who don’t, can go look it up, by following the links. HyperTextApproximatesConversation.
Since wiki make linking and page creation so easy, it makes it easy for people to make a shared vocabulary that is more precise than usual. Anyone who wants to understand a thing can just follow it up.
(B) “ArgumentRouting?” is an interesting concept. This is more in line with HyperTextApproximatesConversation. The idea is that if someone reads a page, and then they disagree with it, they can “zoom in” on particular things that they disagree with- deeper link language.
The idea is that our LinkLanguage forms a “shape,” out of all of it’s links, that we can call an ArgumentPyramid. You trace things down to the source of conflict, the source of DeepDisagreement, and then argue there.
Then again, there may not be a source point- ArgumentsAreHolographic?, after all- Arguments can manifest themselves equally in several pages, simultaneously. That said, if LinkLanguage even just brought you to the realm of 3 specific pages, rather than the 20 pages that include links to them, then we have seen an example of argument routing.
(C) “Present complex arguments and perspectives.” Still more, on how HyperTextApproximatesConversation. (Wow!)
Not only does LinkLanguage link to a page, but that page links to pages, which link to still more pages, which participate in a complete HyperText. Anyone who is wondering why you believe a certain way can dig deep into the argument by clicking into the LinkLanguage. They can then either be convinced, or be unconvinced. If they are unconvinced, and wish to discuss the issue further, they can locate the true source of disagreement (“ArgumentRouting?,” again), saving a lot of time spent figuring out what the real issue is.
This is true in normal human speech and writing as well. Depending on what words a person uses (“Did they say Economic Justice? Or did they say Class War?”), the person is linking to different meanings, arguments, and even forums and authors for debate.
But this is far more true in HyperText. The difference is that in HyperText, you can see exactly what is on the other side of the link.
We seem to have the need to communicate “messages” (or rather, “documents,” in the language of DocumentsVsMessages,) that are larger than we are capable of telling.
We believe in the power of reason and dialog, being children (at least somewhat, perhaps both metaphorically and literally,) of TheEnlightenment?. But it can take a freaking long time to communicate a whole WorldView to someone, and to tailor it to the particulars of the person’s prior knowledge, experience, role, individuality, tastes, etc., etc.,.
There are fundamental limits on what one person can speak to a mass! Undoubtedly, there are difference in people’s understandings, interests, and motivations.
LinkLanguage helps people to communicate enormously complicated messages. People can skip what they already understand, jump to what they don’t understand, aquaint themselves with important basic ideas, more quickly lock terms with what they already understand, but are just used in different terms.
HyperTextApproximatesConversation. We are constructing an artificial ArtificialIntelligence. That is, “it’s not an AI, it just acts like one.” You can “ask it questions,” by clicking on the interesting terms (“What do you mean by that?”) and then seeing what it says.
On a wiki, when you need an actual answer, you can just attach the question to the appropriate page, and, like in AmazonsMechanicalTurk?, (“ArtificialArtificialIntelligence?,”) you will get a response from a human, continuing the “discussion.”
LinkLanguage is clearly not perfect.
Consider:
These weaknesses in LinkLanguage bring up an interesting concept, and that is: “TrustedLinkLanguage” - or “LinkLanguage that you can trust.”
“Do I trust those people’s LinkLanguage, and why?”
KnowledgeIsBasedInTrust. If an individual can’t trust their own “knowledge,” then they basically don’t have it. When the HiveMind can’t trust it’s own knowledge, then it can’t be very smart, either! So there are implications for CollectiveIntelligence: It is founded on the basis of trust. (see also: InterCommunityCooperation.)
The ability to introduce new terms to the vocabulary and to “define” them on separate pages, linking nearly every use of the term with a page about it is a peculiarity of wikis. It doesn’t exist on weblogs, traditional websites, or sites produced by a traditional content management system. Traditionally, wikis have used the CamelCase syntax to link to pages. The resulting page names are obvious names – not just some part of a sentence that has been highlighted and links to some other page. These links are recognizable links to wiki pages and facilitate the development of a Link Language. The notation is important; it affords a Link Language. See WhatIsAffordance.
Noble and Biddle 1 have used semiotics (the study of signs in society, treating communication as an exchange of signs) to explain some issues arround the naming of design patterns in object-oriented programming. In a first step, they say that signs map a signifier such as the world “purple” to the signified (the color purple). As such, the world “purple” can be replaced by any other signifier, as long as the community understands the mapping (eg. in other languages other signifiers are used). This is how we use our Link Language.
Since wiki pages are rarely written like the patterns in a PatternLangue? as Alexander envisioned it, the set of wiki pages usually is not a pattern language. A pattern is not just a piece of text and some discussion. A pattern is, according to Noble and Biddle, the mapping between the name and the pattern description, and the pattern description is itself another sign, mapping solution to intent. Patterns in architecture and software design are tried and true solutions to specific problems, usually containing an explanation or some sort of hidden wisdom about the world around us. Such pages comprise but a small subset of ordinary wiki pages.
Ordinary wiki pages are only first-order signs: They map the name to some text. As local usage of concepts grows, new and powerful associations are formed. By invoking the page name (and thus linking to the page on a wiki), the page names turn into elements of our Link Language.
Contributors: LionKimbro, AlexSchroeder, SunirShah, HeatherJames, piranha, MurrayAltheim, DavidCary
wow. that is cool. i enjoyed reading that very much. i like these nitty-gritty kinds of discussions on how wiki works.
your idea of needing some kind of HardWiki as the basis of an InterWiki association sounds like the need for a dictionary in a spoken language. You consult it for accuracy and clarity, but there is some flexibility. language changes over time, and the cultural resistance (for example between generations) is actually a quality necc. for clear communication.
it makes sense that a hard wiki could work… but i think that if you have this kind of thing universally applicable, the quality of it would dilute. but in a WikiSphere? (what’s the correct term?, heh) of a few sympathetic wikis, a language could be developed and cross-checked and tested. i guess like what goes on here + nearlinks?
Let me extend the HeatherJames analogy between WikiWords and English words. If you don’t know what a WikiWord means, you click on it and read the discussion there. If you don’t know what an English word means in printed text, you look it up in a dictionary. In a conversation, you can also just ask the person what he meant by that word – occasionally he means something quite different from the dictionary definition.
The meaning of English words change over time – ever look up “computer” in a pre-1920 dictionary ?
It’s technically possible to freeze a few web pages defining some term once and for all. But perhaps there are other alternatives we could explore.
We accept the fact that no one really talks in Shakespearean English any more. Perhaps we could accept that this is not the “final” version of that wiki page.
What do we do when English words change ? Why can’t we do the same thing when a WikiName changes meaning (i.e., someone changes what’s on that page) ? (Perhaps I should talk about this over at ContentSwizzling).
wanted to link to something, but been a little nervous about doing so, because it might change?
Yes, that happens all the time to people who maintain a static web page – they copy-and-paste a quote onto their own page, because that other web server might change – or worse, go completely offline.
But with local wiki pages, I might as well just make a link. If the wiki web server goes offline, both pages will go offline. If someone edits that other page, he’ll probably edit this one as well.
When you refer to a concept 20 times in your text, you don’t need 20 links in your text. Too many links make your text hard to read: AvoidClutterLinks.
We use our Link Language when we believe the person reading will have some benefit from knowing what is on the other end. Careless linking of the same words actually holds different meaning than the same words by themselves. The combination of words into a link means something than the words just by themselves. LinkLanguage routes arguments. If you route people into arguments and explanations that have no bearing on what you’re actually saying – you’ve made a bad link.
Another problem we face is the JargonProblem: If you have very specifc ideas that need a name, choose a name for the concept and put it on a page. This is good, because we can accumulate ideas and thoughts. At the same time, however, we need to be aware that communication requires shared symbols. If we all develop our own WalledGarden, our own private set of symbols to use, then our communication will break down. We need to reach out to other readers and write for them, to convince them of the merit of our pages. Only then will the pagename pass from the private sphere into the public sphere.
Similarly, to believe that writing some text on a page defines the pagename such that you can use the pagename without ambiguity is to misunderstand how language works. A Link Language does not do away with misunderstanding, it doesn’t automatically provide all the context you need, etc. A Link Language provides you with some context. It can acts as a topical focus for people interested in further discussion, and it provides the means to make some context explicit by writing it down on the page. As seen above, however, the terms can be used in a certain community even if the underlying page is missing or contains something different. In that case the situation can be improved by improving the page, but it would be foolish to believe that all ambiguities and subtleties of natural language can be resolved this way.
This relates to the PlatonicCategories page I started but haven’t had much time lately to enlarge upon. Words are symbols that mean different things in different contexts. Dictionaries don’t define words, they merely attempt to capture common usage. To pretend that they platonically defined words would be to give them more authority than they have. It is we in the end (the public at large) that are the arbiters of meaning; lexicographers are simply trying to create ongoing snapshots of what we say. As anyone whose language is not considered “proper” can attest, lexicography is like any other science in being as intelligently grounded, or as ignorant, biased and racist as are its practitioners. The “King’s English” is that of the king, not that of the serfs. Why is the king’s language considered more correct than the serfs?
The idea that using a WikiWord somehow removes all ambiguity and all of the problems associated in the use of language ignores the reality of language. The relationship between a single word and its meanings comes down to mapping contexts, and wiki don’t currently (to my knowledge) have the ability to establish and maintain the enormous complexity of contextualized language in the use of a WikiWord. WikiWords cannot escape the bind of their meaning being defined in usage. The WikiWords themselves are used in all sorts of contexts (like real words), so their meanings are (like real words) bound up in their usage, not solely in what’s at the other end of the link, which, after all, is akin to the lexicographer’s definition: just a snapshot of some of the uses of the word.
It seems that what is needed is either a resolution that the system is broken, but broken in the same way that language itself is broken, or to try to create some methodology for establishing and maintaining contexts. The latter is something that is being dealt with in the field of Knowledge Representation, with conferences, journals, etc. devoted to the subject. Perhaps I can try to dig up some references…
Why is it that bloggers don’t have a Link Language in 2004? The system doesn’t afford it. See WhatIsAffordance and SocialAffordance. It is not easy to map terms to resources (easy page creation), it is not easy to use the terms in context (easy linking), and it is not easy to collaborate on the texts (collaborative editing). Wikis do afford it, however.
One possible way out might be an IntComm:LocalNameServers?. Such an external service might at least facilitate the mapping process and leverage the vocabulary existing on other wikis. This would make the easy page creation and collaborative editing arguments obsolete, since page creation and collaborative editing will probably happen on wikis. The bloggers really just need mapping and integration into their software.
If the bloggers build their Link Language by talking in their blogs, and then posting the agreed upon definitions to their wiki, or whatever HyperText they end up with, then we would have another easy path to the end of the PublicRefineryProcess.
Note: Presently bloggers invent words, and teach each other the meanings of the words. And they even attach URL’s to them. But it’s hard to do so. Only a very few terms are built and consistently linked – they become the “famous papers” of the blog world. Essays like “Worse is Better” or “Emergent Democracy” could be considered part of their rudimentary link language.
In a way a Link Language works via memes. If you create high-impact pages, other people will remember the page, and its name, and reuse it, linking to it, and thus telling others about it. By creating new pages, you’re often suggesting that other people use it as part of their Link Language.
Think of discussions about Racism. The discussions I’ve had about Racism tend to take about an hour or two before we can really get into it. Why? Because every word is completely ambiguous. And talking about why the words are ambiguous takes a lot of wind out of us. It takes time to seperate supremicist (“white power”) from discriminatory (“affirmative action.”) It takes time to discuss the history of the civil rights movement. It takes time to explain, and to be clear. And all across the way, emotions raise, and people get upset, because we have massive IntelligenceFailure, and can’t communicate what we really mean faster than we jump to conclusions, anticipating the meaning behind the other person’s words. Live communication is a massive game of synchronizing meanings, fraught with conceptual danger, especially in these emotion-laden issues.
But when it’s all HyperText, it’s easy. “What does Racism mean?” Well, lets just click on it. Ah, I see- it is a word that has a history. The page suggests differentiating “supremicist” vs. “discriminatory” when talking about racism. Racism historically refered to supremecist racism, but now many people link Racism to discrimination on race lines in general. So, we need to be more specific when we talk about what we approve of and don’t approve of. Fortunately, we have a huge LinkLanguage helping us be unambiguous, and to present our arguments. SupremecistRacism and DiscriminatoryRacism.
A link language goes beyond a mere dictionary, because it maps terms to definitions and explanations, opinions, and discussion. – It goes far beyond simple definitions in that it uses more links to “dig deeper” into the terms used, without being as specific in scope as a PatternLanguage.
The closest thing to HyperText LinkLanguage in our non-HyperText languages is the way that we can embed meanings in a sign. For example, words have root words that mean things. Chinese characters have sub-characters that mean things. Those meanings contribute to the understanding of the word.
PatternLanguagesAreDoubleSpeak
PublicWebJune2004Map CategoryNameSpace
linklanguage link links linking camelcase wikisyntax language
This page is majorly due for rewrite. It needs to be PlainTalk.
And I want to include this link:
Apparently, waterfall was described in a paper that intended to refute waterfall. ..!
I’d want to check the actual paper, though, in order to make damn sure we’re not repeating the mistake..!
The relevance is this: When we’re using LinkLanguage, we have to be careful that the other side really means what we think it means.
Words change in meaning over time, and that can happen to LinkLanguage as well. If we are going to use a phrase in a way that it doesn’t mean, we should make sure it doesn’t link.
Not sure how this fits with the PlainLink concept; PlainLink would probably link it anyways.
At which point we have to ask ourselves: What is the point of the link? Is it to provide relevant information to the user, or is to explain what we mean as an author? Are we producing a corpus (a body of related text,) or a HyperText (as an intentional work?)
> lest you say things that the links betray! > means what we think it means. > PlainLink would probably link it anyways.
The PlainLink specification shall “do the right thing” for plain text, but current implementations are broken in a variety of ways.
I just saw this quote:
(Emphasis mine.)
I feel similarly about LinkLanguage.
I love the idea of a [[link_language?]].
Language are the neurons flashing in the [[hive_mind?]].
An example of how I’d imagine a link language using [[local_names?]].
This page is drawing argument about something that the page is not about: CamelCase vs. CleanLink? / (insert name of the other proposal here- where all text automatically links however possible.)
This page needs a rewrite, as well; Likely a few connected pages as well.
Amber & I just completed a huge move, and don’t have Internet at home, so our schedules are completely haywire right now; I literally do not know if I have the 12 hours to work on the pages. But I think I do.
Where else on the wiki have we disussed CamelCase vs. CleanLink? vs. other linking systems? What was the name of the one championed by PatrickAnderson?
In the current write-up, I think there’s too much of a mathematical approach / feel to it.
I think what’s really important is how it is like spoken language, weaving together different discussions, different sequences of time, into a coherence.
There is a pragmatic component (how it works, mechanically and socially,) and there is a living component (how it interacts with living systems.)
The practical effect is the transcendence that results; LinkLanguage (hypertext emergence at the level of spoken discourse) is literally (yes, I used the word correctly) a whole new level of existence and ordering relations.
The question of CamelCase vs. otherwise is to the side.
Thoughts on the plans?
It’s not just feel. It uses the annoying mathematical tradition of presenting you with requirements and results immediately, without telling why the requirements are required or how the results result from them. I would start with replacing the formal definition with just a description of how it does, something like:
I’m still reading, re-reading and analyzing this page…
Thanks.
It works basically like a communication variant of Wikipedia:Subroutine (or function) in programming.
I have been referring to this page a lot lately in http://socialmediaclassroom.com/community lots of people there new to wiki, trying to figure out useful ways to use wiki in education contexts.
Very good Radomir. Lemme refine:
People writing in a wiki often put important concepts or terms on separate pages in order to refer to them. If the name of such a page is repeatedly used by many to point to an idea described or discussed on it, the page has become a part of the [[link_language?]] of the respective wiki-community.
Leut’, die in ein wiki schreiben, lagern oft wichtige Konzepte oder Begriffe auf eigene Seiten aus, um auf sie verweisen zu können. Wird der Name einer solchen Seite wiederholt von vielen benutzt, um eine Idee aufzuführen, die dort beschrieben oder diskutiert wird, so ist die Seite Teil der link-Sprache der entsprechenden wiki-Gemeinschaft geworden.
See [[kabo-wiki-hive_-_whoRthey-wiki_-_one_big_local_name?]] for the current choice.
Sam, I can’t seem to be able to access that page, it says “Access denied. You are not authorized to access this page.” I think it makes it very hard to refer to
Mattis, do you remember which page had this WikiWord vs. rest of the world discussion? Somehow I don’t like hyphens…
Humans link language to definitions even when only spoken.
Written language works without CamelCase, []‘s, _‘s or -‘s.
We name each term carefully to define it within each context.
A listener or reader links each term to a definition that may not coincide with the author’s intent.
TheSheep, I hear you. It is for right now a closed site online. Yet, it’s purpose is more important than whether it is open or or not. I’ll advocate that the discussions and resources be opened up, as they should be referenceable.
TheProfessor, it’s a while ago I had be improvising about naming all grammatically possible versions of a page’s title on the page editable titles. It should be all possible with the [[local_names?]] now, I guess - I haven’t been experimenting yet though.
TheSheep, I pretty much liked the hyphens right away. Spaces and hyphens are enough. The combination “space - hyphen - space” clearly distincts sections. No “/” is required, that’s my feeling. When uri’s were made there were different restrictions to what restrictions we face now (see kabo-wiki-hive - local-names-sand-wiki - talk, 2008-10-06) for an easy-to-read global file system. The huge semantic advantage now is that we can use spaces. Admitted, utf-8 instead of ascii is also pretty cool.
http://www.communitywiki.org/cw/LinkLanguage
web - community-wiki org - cw - link language
Sorry, I don’t remember the page about WikiWord vs. rest of the world discussion.
Current Plans for the page:
It is getting more and more difficult for me to understand wiki and our developments here secularly; I see evolution.
Proposed “description of how it does”:
RadomirDopieralski: People writing a wiki will often put important concepts and terms on separate pages, in order to refer to them all over the wiki. Soon everyone start using the names of these pages to call ideas the page describes or discusses. When this happens, the page has become a part of LinkLanguage of that community.
MattisManzel: People writing in a wiki often put important concepts or terms on separate pages in order to refer to them. If the name of such a page is repeatedly used by many to point to an idea described or discussed on it, the page has become a part of the link language of the respective wiki-community.
SamRose: It works basically like a communication variant of Wikipedia:Subroutine (or function) in programming.
Referenced by:
Placeholder:
Cost estimates:
So I’m guessing this reconstruction is a major project, costing around 2+(10*2)+20 tops (42 hours,) or 2+(5*1)+10 hours min (17 hours.)
Sam, I still remember that I owe you for lionwiki 1 CSS chart diagram. Can I make taking this on clear my name with you?
I have only just yesterday received dialup at home. (We moved.)
And HansWobbe, I still have outstanding obligation to you as well. I’d really like to do this, though, and have everyone’s blessing. Hans, I think this would help to answer the question: “Why Wiki?”
Take care, Lion Kimbro
Lion: hell yeah, you don’t owe me anything. I learned CSS in about a week, no sweat, and with some early help and pointers from TheSheep (who pointed me to W3C CSS School. http://www.w3schools.com/css/ )
Lion… Please don’t worry about any “obligations” to me. The contributions you made were quite helpful and I would rather you felt free to re-engage in whatever way you may care to contribute, in the future.
OK!
Here’s my plan, after going through the page contents:
1. Definition (and making use of the RadomirDopieralski MattisManzel definition)
2. major ideals: fluency, commonality, automation (aided by NamingForLanguage?, WikiCoherence? (invoking: TrustedLinkLanguage, and possibly SharedIdeals??) and ComputerAssistedLinking?)
3. effects: precision (LinkLanguageIsMorePrecise?,) leading to ArgumentRouting? and, most importantly, the delivery / evolution / emergence of a ComplexMessage or complex discussion
4. EvolutionOfHumanity? – a description of why this is worthwhile at all; TheoryBuilding, EvolutionarySpirituality, the ArrowOfComplexity?, ThePainCreatesTheOrgan?
5. Correcting Misconceptions: PrecisionDoesntBanishParadox (connects with ParadoxOfExpression and MindTheGap and PlatonicCategories)
6. see also: CamelCaseVs? Others (CamelCase, PlainLink, LinkSyntax?)
I’m also thinking that I’ll make “LinkLanguageSmallPages,” to help define small terms that come up, or places where “a full page would be really nice, but we’re cutting it short for now.”
What are some ideas that can improve this plan?
I feel pretty good about this plan, though I am wondering about order: Should it be pyramid (def, major ideals, effects, evolution of humanity – build-up, from less controversial to more) or reverse pyramid (evolution of humanity, effects, ideals, definition – the order of newspaper journalists)?
I favor the pyramid generally, but people do often tell me, “get to the point.”
Next step, if there’s no discussion, is to start working out pages.
Define external redirect: TheEnlightenment NamingForLanguage kabo-wiki-hive - whoRthey-wiki - one big local name SvgWiki ArgumentsAreHolographic CleanLink SharedIdeals WikiApproximatesAutomaticConversation CamelCaseVs TheTimes WikiCoherence local names PatternLangue LocalNameServers hive mind link language WikiSphere LinkSyntax ComputerAssistedLinking ArrowOfComplexity ArtificialArtificialIntelligence EvolutionOfHumanity AmazonsMechanicalTurk mind the gap LinkLanguageIsMorePrecise ArgumentRouting ThePainCreatesTheOrgan