The Chinese Tent

Welcome to the Chinese Tent…

Imagine a large tent, and in it, a person who speaks only English. We’ll call this person the respondent. Outside the tent is a person who speaks only Chinese. We’ll call this person the questioner. The questioner writes a message in Chinese and passes it to the respondent through an opening at the front of tent. The respondent receives the message, then using an infinitely complex look-up system, matches the message on the card with a corresponding card containing a response in Chinese. The respondent delivers the pre-written response back to the questioner.

From the perspective of the Chinese questioner, the English respondent has read the message and understood it, demonstrating intelligence. However, has the respondent actually responded intelligently? Remember, the respondent does not understand a word of Chinese – he has simply matched up a pre-defined response to the received message.

This is not intelligence, but an elaborate forgery. The Chinese Tent exposes the folly of thinking that computers can actually be intelligent. Computers can simulate all sorts of reality; this does not make the reality they simulate real.

When a computer runs a virtual stock market exercise and makes millions of virtual dollars, the user does not become wealthy. When a computer runs a weather simulation that has 40cm of rain falling in one day, the user doesn’t get wet. Yet when a computer simulates intelligence, we all rush to believe it’s actual intelligence, and not a pale imitation.

It all comes down to definitions, something technical communicators are very fussy about. So how do you define intelligence? Here’s one standard dictionary definition:

“the capacity for learning, reasoning, understanding, and similar forms of mental activity; aptitude in grasping truths, relationships, facts, meanings”

    The key word here is understanding, which implies consciousness; the knowledge that one exists. This is something that computers simply don’t have. 

    However, if you define intelligence as the ability to give a meaningful response to a question, then computers could indeed be intelligent. A recent example of this was Watson, a super-computer developed by IBM. Watson was a contestant on the Jeopardy game show, beat his human opponents and won a million dollars. 

    Depending on the definition of intelligence you choose, Watson is intelligent or simply another version of the Chinese Tent. It is able to analyze questions and supply responses, but if you were to ask it how it felt about winning against humans, it would have no response.

    What, then, is an “intelligent” document? It is one which understands the information requirements of the user and then seamlessly delivers them. Now, there are help systems that allow a question to be entered and then try to give a relevant response, but these are very complex to set up and have mixed results. A far simpler solution is to create a complete and meaningful index, one which anticipates all the strange and wonderful ways a user might look up a topic.

    An index also has the advantage of revealing potential gaps in your content or the index itself.  For example, you may have the following index entries:


    Where is the entry for editing files? Doh!
    A complete and meaningful index is an intelligent index, and helps to make your document intelligent. Of course, once a user finds the topic they are after, it has to be intelligently written. This means that the topic should answer questions, and not raise them. Questions such as:
    • What is this object, thing or concept?
    • What is it used for? Why would I use it?
    • How do I perform a specific task?
    • What are specific things I need to be aware of?
    • What things or tasks are related?

    An effective document, then, is one giant answer book. An intelligent document gives the user the answers they need, without the user even realizing the effort they made to find them.


    One thought on “The Chinese Tent

    1. Margaret

      I think indexing is a forgotten skill that can serve several purposes. Many of the newer documentation management tools, such as DITA, XML, and many kinds of single sourcing depend on tagging and markup schemes that enable you to assess what content you already have and what will be needed to complete any given project. How can anyone expect to master any of these tools if they have never learned how to create an intelligent index? SharePoint and other new enterprise content management tools won't work without an overall organization that makes sense to the users.
      Without indexing skills, it will all go back to GIGO–Garbage in, garbage out!

    Comments are closed.