About The Author


The influential philosophy professor, John Searle, created the Chinese Room Argument in 1980. It was written to demonstrate a simple point –intelligent behavior does not equate to intelligence. This doesn’t mean A.I. design is impossible, but that a behavioral-based model for intelligence is flawed.

Deep Blue beat Kasparov. Your calculator can outshine Daniel Tammet. Did Deep Blue understand the game of chess? Does your calculator understand mathematics?  Searle’s brilliant scenario illustrates the difference between intelligence and intelligent behavior.


This argument involves the following:

A Room, which contains:

  • A small slot in a wall that’s large enough to pass pieces of paper through
  • An English speaking individual sitting at a desk, who has a massive supply of pencils, erasers and scratch paper

A Chinese speaking individual standing outside of the room. She has two pieces of paper:

  • A story written in Chinese
  • Questions about the story, which are all written in Chinese

Going forward, the English speaking man will be named Mark and the Chinese woman Anne.

Anne slips the Chinese written story through the slot. Mark receives it, and sees a bunch of characters on it that he can’t understand, but what seems to him like Chinese. He doesn’t understand or speak Chinese, but opens the gigantic book and finds instructions written in English, explaining how to manipulate, sort and compare the Chinese characters. Nothing about the meanings of these characters is written.  Only instructions are given: ways to copy, erase and rewrite the characters.

Mark gets to work, and follows the instructions exactly as they are provided in the book.  Sometimes he’s told to write characters on paper, and other times to move and erase characters. Following the instructions without a single original thought, he finally comes to a point where the instruction book tells him to pass his paper back through the little slot. Mark doesn’t even know it, but he successfully produced answers to all of the questions on the paper – solely by following the rules given in the book.

Outside, Anne receives the paper. She reads the answers and confirms their accuracy. Now, you walk into the outside area and strike up a conversation with her.

You: Anne, do you think those answers come from an intelligent mind? Did the person inside understand the story?

Anne: Of course. Actually, the answers were quite perceptive.

And therein lays the problem. Who understood the story? All that occurred was Mark’s sheep-like obedience of instructions from a book written by someone else.

The Analogy

So what does the Chinese Room represent?

  • Mark – the CPU, executing instructions without thought
  • Giant book – software platform giving instructions to the CPU
  • Scratch paper – the computer’s memory
What’s the Point?

It doesn’t matter how perfectly a computer is designed to simulate the intelligence of a human being -because its behavior is a result of aimlessly executing instructions, not understanding. In this case, the means defines the end. You’re reading this sentence, and understanding it without demonstrating behavior of any kind. A system’s behavior doesn’t indicate intelligence or understanding, and a system that behaves intelligently is not necessarily “intelligent.”


2 Responses to The Chinese Room Argument

  1. Ian says:

    Is the software written for a machine not analogous to instinct and learned responses provided for humans by evolution & culture?

    Why does our “intelligent behavior” qualify as intelligence? Would it qualify as such if we were being evaluated by an entity with as much of a processing power advantage over us as we have over computers?

    • Ellz says:

      Some first thoughts: ( a bit jumbled but..)

      First the fact that these languages have been conceived and invented by us not to mention the wall the paper and the pens/pencils and desk were all human inventions. Intelligent use of tools/materials for our own use.

      Secondly a computer would do these things in a split second fashion. A human doing it has to use its eyes and remember the representation of each character as it is learnt.

      This test shows nothing about intelligence other then you cannot expect an artificial intelligence to be smart unless it has been taught. Just like we pass on knowledge to children.

      Mark did well in the test because someone that knew about the language and knew the answers wrote them and made a book. Mark if given time could be taught/programmed to read chinese and mark if he wanted to could’ve wrote a letter asking lee to translate everything in english. Mark in this hypothetical situation cannot be used as a symbol humanity as not every human would’ve done this without questioning in their minds what they were doing. And probably Mark would’ve gone home and read up about chinese and tried to learn something about what he did.

      For a computer to be artificially intelligent, it needs to be given the freedom of choices and for it to believe in its choice.

      For example:

      You ask google if it thinks there is a cure for the common cold.
      It answer only what we have programmed it.

      You ask an artificial intelligence if there is a cure for the common cold.
      It asnwers no. But it thinks about tests and theories because it takes everything it knows and see’s that tests have given cures in the past and decides that if other things can be cured then there must be a way to cure it. It will then simulate and create a list of possibilities of what it can test. And then depending on how powerful its thought process is, it can start testing in its mind.

      Just a random idea of what could be the key for true artificial intelligence, perhaps it lies in the way we all have that black/blue thought mixing place that we all get when we stare into space and think about stuff.

      The ability to make

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>