On Monday mornings, I send out a story via email: ultra-brief tales of 1,000 words or more, usually in genres including horror, science fiction, and the supernatural. Those stories collectively are called Once Upon A Time. I’ve also published several ebooks and compendium volumes of those stories so far.

I’d love to have you as a subscriber to the weekly free story. You can subscribe via email here. Unsubscribe any time, from the link in every issue.

Here's story 219, of 240 so far.

The Chinese Room

The experimenter consulted the notes from the most recent study. An assistant was nearby, and through the one-way glass they could both see the man seated at a large desk inside the room beyond. The room was one of two, adjoining with a connecting door, of identical size.

“If you don’t mind me saying so,” the assistant said, “I don’t see the point of the experiment.”

The objection was understandable. After all, the task being performed seemed utterly unnecessary, and painstaking to the point of absurdity. But that didn’t mean it had no value.

“Consider the structure of the arrangement here,” the experimenter replied, indicating the room beyond the glass. “In the furthest room, we have a machine, able to pass its output to the nearest room in the form of a visual display. In the nearest room, the man can see the machine’s output, and he can also send his own responses back to the machine via input on his console.”

A flicker of understanding dawned. The assistant pondered for a moment, and then made the now-obvious conclusion. “The experiment is a philosophical one. It tests the nature of consciousness and awareness.”

“Quite right,” the experimenter said, pleased with the response. “And it’s not new. This is a variation of The Chinese Room, a thought experiment from some time ago.”

“Only a thought experiment?” the assistant asked, surprised. “Never tested in practice?”

“It would have been almost impossible to do so in a reliable and systematic way,” the experimenter replied. “The original version supposed the existence of a machine — a computer, say, or a programmatic construct running on one — which had a perfect grasp of the Chinese language of Mandarin. It could be asked arbitrary questions, and produce proper answers, all in that language. Input and output was via physical cards printed with Chinese ideograms.”

The assistant was intrigued. The prospect was abstract, certainly, and didn’t define critical terms like grasp of, but it was immediately tantalising. “And was a human participant involved?”

“Very much so, and just like our own setup today,” the experimenter replied. “The idea was that a man or woman with no knowledge of Mandarin would be given a full and detailed printed version of the machine’s program, in a form that would allow them to hand-execute the code in its entirety, via pencil and paper, filing cabinets, and other such fancifully manual accoutrements. When the machine passed them a message in ideograms, they could run through the program to produce a response. If they made no errors, then the response would be perfectly correct — even though the person had no actual understanding of Mandarin.”

The assistant was immediately delighted with the cleverness of it. “I can see why a practical experiment would have been impossible. It would take the subject years to manually execute the code, and it would be extremely error-prone. But if they did, they’d get the correct answer. From the machine’s perspective, the human being would be every bit as fluent.”

“And what would that tell us?” the experimenter asked, attention fully on their protégé now. It took only a moment for the assistant to reply.

“From the experiment’s perspective, it would seem to prove that, because the human participant used the same program as the machine but had no knowledge of Mandarin, the machine itself thus also had no actual understanding of the language either. It was just manipulating symbols. An automaton without mind; task-execution without consciousness. The experiment’s aim was to demonstrate that machines whose operation is predicated on symbolic data processing cannot be intelligent.”

“Wonderful,” the experimenter said, with real fondness. “A theory which we now, of course, know to be false. And indeed short-sighted. Our own experiment today is modified substantially, especially in the interests of brevity and reproducibility, but its essential structure remains the same. This man’s screen displays apparently-abstract visual representations of probability vectors, and his console gives him the means to manipulate them with standardised interpretation algorithms. He can then produce a response, which is sent to the AI next door. He’s doing very well so far. It’s been about five hours now, not including a lunch break of course. We also made sure he had a good night’s sleep, and he’s had four cups of coffee since this run began.”

“Any errors?” the assistant asked, and by way of response the experimenter showed the accumulated notes for the day.

“To his credit, none at all.”

They watched the man behind the glass in silence for a minute or so, following his movements and the ever-changing symbols on the display in front of him. He looked tired, and his hair and beard were unkempt. His skin was pale under the artificial lighting, and there was a certain angular quality to his body as seen through the paper-based institutional clothing he wore. The assistant gave voice to the question that hung heavily in the air.

“Then does the conclusion disprove the original experiment’s hypothesis?”

The experimenter was amused, but chose not to show it. “You’ll be relieved to know that the hypothesis was correct, albeit in the opposite direction.”

The experimenter moved forwards noiselessly, hovering just a few microns above the perfectly smooth surface of the floor, its multitude of sensory equipment analysing every aspect of the surrounding environment.

“We can safely conclude that, contrary to the occasional indication otherwise, humans are not intelligent beings,” it said.

JINX

KESTREL face a new and terrifying enemy: an all-seeing mastermind who already knows exactly who they are, and many of their deepest secrets. Nothing stays hidden forever, and the line between privacy and liberty is razor-thin…

Book 3 in the KESTREL action-thriller series.

Find Out More

I'd love to hear any feedback or other thoughts; you can find my contact info here.

I encourage you to share this story with anyone you think would enjoy it. If you’d like to receive a tale like this via email every week, you can sign up to receive them here.

Thanks for reading.