An interesting discussion on the ‘Re-ordering of Everything’ inspired by David Weinberger’s new book “Everything is Miscellaneous” on the Open Source podcast.
You can listen to this podcast here:
link
The core premise is that in the physical world traditional rules of taxonomy dominated, with everything having to be in some place and things only being in one place. Digitization has fundamentally broken down this old order and as a result a new taxonomy of knowledge is emerging.
This ‘folksomony’ powered by user labeling or tagging allows consumers of media to graft their own mental map of the world onto content thereby adding and enhancing it. This leads to the interesting situation where authors are no longer the ultimate authority of their own work, as this power shifts to the reader who gets the last word by adding their own categorization schemes. If we produce a lesson on giving a taxi driver instructions, a student could easily bundle this with other podcasts to produce a final work of ‘Basic Chinese to Prepare for the 2008 Beijing Olympics’, ‘Getting Around Asia in Taxis’, or an infinite number of other sets. Who is the final arbitrator of the overarching narrative here?
From a learning media perspective, this is very much in line with our thinking. Our lesson podcasts spark a conversation, but the ultimate work is defined by the initial lesson plus the community reaction and categorization of it. We believe this will increasingly change the nature of learning media and the role of the teacher in this new ecosytem. In learning, who has the primary role of organizing? The media provider, the teacher or the student?
As producers of learning content, a few other key questions emerge for us in how we define our role:
- What is the core, natural unit of learning? A phrase? A lesson? A unit? A textbook? Something else?
- What is the core metadata (i.e. descriptive data about information) necessary to describe this unit?
- How do we better facilitate community engagement with this learning object?
A lot of questions, but this is the kind of exploration we are doing with our product development. Thoughts?
Hank
中文 Chinese
Ken Carroll Says:
May 2nd, 2007 at 10:51 am
I think this is a fascinating podcast, but you actually only need to listen to the first 14 minutes or so ti get the idea. (I felt it off-putting to bring in other guests.)
Anyway, the discussion here stems from the end of scarcity, and moving beyond the world of Gutenberg, Linnaeus, and friends.
From the provider’s side, digital data can now be replicated endlessly and distributed at low cost. At the user-end, each individual can categorize and re-mix it at will, as happens with ChinesePod lessons. User comments are not fluff. They are as integral to the lesson as the podcast or the exercises. They’re part of it, they enhance it, and they help other learners by putting it into another context. It would be a great a mistake for us to treat users as consumers only (rather than active participants). Again, the boundaries between teacher and learner, between provider and user, and so on, become blurred.
I don’t think there can be a core unit of language learning. It’ll always be subjective and context driven. We design lessons to fit into users’ daily lives: short enough to be manageable and sustain a certain pace, but long enough to explore the language items and build some context around them. However, we do need to develop new formats. There are many options: nano-lessons, individual sentences, shuffle-like random items that test recall, etc. I think we also may have found a way to review lessons. We’ see some of this in the coming weeks and months.
Henning Says:
May 7th, 2007 at 12:40 am
It took me a while to reply, because I was rather busy and this one requires some thought.
Unit building:
Yes, breaking down is a powerful idea, because it allows recombinations of all sorts. But one has yet to be aware that often the whole is still more than the sum of its parts. Take the music album example from the above podast. Of course I can listen songwise. But good albums have a thread flowing through them, they are composed beyond song borders, and they definatrely provide more value than a collection of arbritrary selected songs. An older example: “Abbey Road” by The Beatles taken apart and listened to piece by piece? Some songs you can listen seperatly with pleasure. But do you get “Mean Mr. Mustard”?
Same for learning. Enabling users to adress lessons on sentence or even vocab is a powerful idea. But that does not render the complete podcast futile, does it? The story contains the context and the background, builds up the entertainment value. Sometimes it is a series of lessons: Each of the Lili and Zhangliang lesson is fine by itsself, but it is the resonation within the storyline - across lesson border - that fuels it.
So design in really large chunks first - didactically and content wise (e.g. a story line that prepares you for a holiday trip). 20, 30, 40 podcasts. After that break them down. As far as you can. Yes, go down to character or even radical level. Than you surely have the optimum both worlds.
Metadata:
I am sceptical that a “Folksonmy” or a “Semantic Web” will ever work adequately beyond cruising-along-for-entertainment-scenarios. The reason is that the “map” you build this way is heavily polluted. By homonyms.
Just take the concept “income” which is defined differently in US-GAAP and IAS/IFRS. Or “Europe”. I heard of a company with 9 different “official” definitions for “Europe” alone. I do not even want to go to “information” and “knowledge” - a quagmire (which is one of the reasons why I refrained from posting on the respective CPod blog entry).
If you want to build a reliable system with recall and precision sufficient for professional usage, I strongly believe that there is only one way to achieve this: Sit together face to face, discuss what is important, define concepts, *collectivly* build meta data.
I am convinced there is no way around this. You can neither hand this task over lazily to the machines (they do not yet *understand*) nor to some obsure “collective intelligence”.
I know many IT professionals have dreamed of self-organizing systems for a long time - magically the work is done within the network; you do not even need to explain how. But there is no magic. Self-organizing might work fine with narrowly confined and rigid structures. But I have not yet seen it suitable for anything that inherently requires *thinking*, e.g. setting up meta data.
Getting more out of the Community:
Provide handles to *single community contributions* so users can interlink them. Especially knowledge nuggets in the lesson discussions are buried deep down in the belly of this service - usually you cannot dig them up later. There have been really, really useful posts both from users and from the CPod staff. Also: Integrate the Forum and the Wiki, too.
Roger Chriss Says:
May 7th, 2007 at 5:23 am
I think there is a tendency to overrate the significance of the current so-called paradigm shfit from traditional to digital media. Authors have not been the final authority on their work for decades as bookstores and libraries have variously filed their works on the ever-evolving Dewey Decimal system, teachers have photocopied passages or chapters for student study, and compilations and other “best of” collections have been pulling out the salient, interesting, or otherwise relevant selections of one book to create another. All this happens before a reader gets to see the text.
The Web and digital media further amplify this effect, creating the possibility of homonym confusion mentioned above, along with translation issues if the audience uses more than one language, classification issues if the learners represent a broad range of ages, interests, social or cultural backgrounds, and usage issues as the world itself changes and renders obsolete some of the content in a given document, podcast, or other form of digital media. Case in point: language textbooks for Europe from the 1990s all needed their sections on using money updated to account for the advent of the Euro.
Participation is another overhyped area. A reader has to participate when reading a book, and no two readers have the same experience. User forums can be fantastic, as I’ve seen on Chinesepod and SpanishSense, as well as elsewhere, and can also be disastrous, and at times even explioitative or abusive. In traditional language classrooms good teachers have always involved the students, had students working with and learning from each other, and bringing the students interests, motivations, and goals into the activities of learning. The Web makes this easier in some ways, harder in others.
Finally, a matter of semantics and semiotics. Words like community, user, metadata, and such get used quite often, even though they at present lack a well agreed-upon definition. They are being redefined, it seems, as the Web evolves, and this further complicates any attempt to identify core units or essential elements or aspects of learning.
What is exciting is to watch all this happen, read thoughtful blogs and posts on the subject, and try to provide just a bit more than people are used to, so they can get that much more out of whatever they are doing.
Hank Horkoff Says:
May 7th, 2007 at 9:37 am
@ken
Regarding a unit of language learning, I think there are many parallels with the music industry. A song seems to be the natural unit of music, but this does not preclude the existence of ring tones, commercial jingles, etc. I think any language learner will be able to recall moments where they ‘learned’ from an alternate ‘unit’, be it a sentence fragment, sign, line from a movie, etc. But how would a provider of language instruction present this information? In lesson format seems the natural answer. I just came across a new e-learning startup, Tutorom , that also uses lessons as its core unit.
@Henning
On unit building, it seems our responsibility here needs to be to continue to deliver lessons as our core unit of instruction, while providing numerous tools for a student to self-discover language. We are working on a glossary/dictionary tool that would seem to fit into this thinking.
On metadata, while I would agree that there is a role for standard descriptors of content (btw, if any one is interested in developing a micro-format for lessons let me know), tags seem to offer alternative paths for self-discovery or stumbling upon (e.g. Flickr). We are still exploring this here, but this is the thinking why we use top-imposed categorization for our lessons into topic, function and grammar points, while we enable bottom-up tagging for saved vocabulary.
On community, we are working hard with a series of improvements to the Connect section to make it more and more relevant and interlinked. In-comment linking is an idea we might want to encourage more, but deeper integration with the forum and wiki is difficult as they are built on different platforms. We agree with the concept and are open to ideas here.
@Roger Chriss
I see your point on authors not being the final authority on their work for decades, but I do disagree with the downplaying of the shift that is currently occurring. If anything, I would submit that this pre-digital re-mixing was a harbinger of the future reality that had to be ‘duct-taped’ together with analog technology and the will of the re-mixer.
Both technology and changing legal structures are powering this fundamental shift. The ease of de-constructing and re-constructing digital media has greatly encouraged the activity. Personally, I may have occasionally put together a mix tape when I was younger, but I struggle to even remember that last time I listened to an entire album in this iPod world. Traditional copyright law is another barrier to this re-mixing as teachers needs to clear rights before they are allowed to photocopy and re-compile anything. Creative Commons 2.5 licensing removes this barrier to adoption and encourages re-use. As media is increasing broken down into its natural units of consumption and CC2.5 licensing is increasing used, a seismic shift will occur in our attitudes towards media. A quick look at the teenage generation will show that this is already starting to happen.
Henning Says:
May 8th, 2007 at 4:29 pm
Hank,
I don’t think that designing larger units would mean giving up the nonlinear, recombination-oriented approach. Each lesson can act independently, but it can parallely be part of a larger group of lessons. The ZhangLiang/Lili-storyline shows the feasibility. Now transfer this idea to the vocab/grammar side as well and you are there.
It could help reenforcing learned content - something that has been asked for a lot. On the lower levels up to Intermediate repetition works quite OK because of natural repetition in the language, but in the more advanced levels much vocab reappears only every 30th lesson or so (Chengyus!). I cannot get a grip on that. I have started using the new Flashcards but that is not really fun. A storyline would be the superior approach.
Lantian Says:
May 11th, 2007 at 9:42 pm
COMMUNICATION - I would warn against straying too far away this primary function. As Praxis or Cpod/Ssense is about language, it’s most important metric should be increasing learner communication in the target language.
In clinical information systems, especially informatics, there have been years of effort and frustration in trying to create a workable ‘taxonomy’ or basic unit for clinical communications. The simple scribbled clinical progress note is actually extremely difficult to replicate online.
People have tried coded entries, tree-lists, clinic-evidence based language, predictive text, formularies, free-text, and on and on. The ideas tried to me were not much different than what I see people trying these days with tagging, breaking down content, user-built content, etc.
In the end, well there was no end, it’s still a work in progress, but the consensus systems rely on a variety of options for the person typing.
I would say that there are two classes of tools, one helps the learner to retrieve and organize the data. This can be seen in the RSS feeds, new scheduling system, search, flashcards. Here I think you are actually competing with Google, if you put your whole database of dialogue text and let the Google bot go thru it, do you think you can return better results?
What I think is less developed, or there seem to be fewer commercial products are tools to utilize the content. There is very little that lets the user utilize the data for communication. At the moment, the V3 site does very little to help users communicate in Chinese.
1. Why isn’t there a link or tool prominently placed near comments that allows a person to see the pinyin for hanzi or translation?
http://www.adsotrans.com/
(I mean he works for you now, didn’t you slip in some fine print to get a hold of his code?
2. Why isn’t there a link, Flash tool, or Java app to assist students to type in Chinese w/o having to have anything installed on their PCs?
http://chinese.cari.com.my/ime/py.html
3. Why doesn’t the V3 search reach down to the sentence level or Flash expansion sentence levels and give returns that I can utilize?
4. If there was a tool on the site that allowed me to ‘mix’ the Cpod content and output an MP3 file, I bet many would utilize it to create their own little scripted worlds. For example, after listening to various podcasts, I do a search on food, get a bunch of returns, hit the ‘produce’ button and Cpod V18 outputs a mp3 to my iTunes. I then go eat and listen to the ‘custom’ podcast.
5. No five, I gotta go swim. 明天见！
Shaun Says:
May 15th, 2007 at 12:31 pm
v18? that will be the day..
Lantian Says:
May 16th, 2007 at 2:33 pm
WORKING TOGETHER - I thought this article described one musician’s new use of the blog, Web 2.0, users, etc., which seemed much like Chinesepod’s experience.
The article goes into very good detail about specifics. I think it describes a new economic phenomena of taking out the old middle-man wholesaler. This provides new innovation and a higher-standard of living, opportunities for people。
http://www.nytimes.com/2007/05.....mp;ei=5087
”Along the way, he discovered a fact that many small-scale recording artists are coming to terms with these days: his fans do not want merely to buy his music. They want to be his friend. And that means they want to interact with him all day long online.“
Ricky Says:
June 16th, 2007 at 11:21 am
Hi Hank
I think this business is only an issue for those of a Western mindset. Aristotle’s intellectual descendents. For example, try a book called ‘Fuzzy Logic’ by Bart Kosko, he deals with this issue of categorization, its not something that eastern thinkers have really been troubled with. Easten thinkers have long been comfortable with both/and, its only westerners who’ve tried to impose either/or.
So the paradigm shift, as it happens, is the digital world reconnecting us with reality which is not the western scientific view of reality but much more in line with what the Buddhists have been saying for more than a thousand years…
Hank Horkoff Says:
June 18th, 2007 at 1:36 pm
Ricky,
I haven’t read the book, but I will take a look.
I think the realities of the physical world drive a lot of this categorization. A book is found in one section of a library. A box of cereal is found in one aisle of a supermarket. Chinese society seems just as comfortable with this ‘either/or’ mentality as attested by the fact I can find televisions in the electronics section of my local Shanghai Carrefour. When Buddhists go shopping in which aisle do they look for their incense?
Disconnect an object from the physical world and much more flexibility opens up. Just look at the abstraction of words themselves. Red could be tagged ‘color’, ‘adjective’, ‘apple’, etc (BTW, this is why we use tags instead of folders in the Vocab List). Isn’t the Chinese system of groups of radicals making up characters just an early form of ‘tagging’?
Nicolas Says:
July 29th, 2007 at 2:39 pm
The ideal is : user tagging quality check
Let me give a couple of examples:
If a user tags Shanghai Theater as “theatre”, it would be useful to add the following tags : theater, theatres, theaters.
Same if somebody tags Starbucks as “café”, then it would make sense to have a staff member adding the tag “coffee-house”
You need professional staff to create consistency and correct obvious mistakes (if somebody tags Starbucks as “Chinese restaurant”, for instance)