Quantity Versus Quality in (Mere) Generation

Arthur C. Clarke’s short story The Nine Billion Names of God memorably captures the remarkable yet typically unreflective capacity of computers to generate candidate answers. In the fictional setting of the story, a lamasery of monks in the Himalayas have spent centuries pursuing their goal of listing all the possible names of God, in the belief that since the world began with the name of God, it must also end when the last name of God has been uttered. The monks have devised a holy alphabet, and for reasons known only to themselves, have determined that any name of God can contain no more than nine of these letters. They impose a variety of other constraints too, such that no name of God can contain a sub-sequence of three repeated letters.

This monastic Bletchley park has always enumerated these names by hand, painstakingly adding each to a giant book of names. But the advent of modern technology offers an obvious means of automating this process, and so the monks engage a computer company to send them both a computer and a pair of programmers for the task. The computer is easily programmed with the alphabet of holy symbols and the rules for name generation, and it is quickly put to work enumerating and printing all nine billion or so names.

However, as the task nears completion, the programmers worry that the monks will become violent when their religious beliefs are falsified by the world’s refusal to end on schedule. They make their excuses and leave early, so as to be far away when their program terminates. As they trek down the mountain on ponies, laughing at the monks’ strange mix of superstition and technological savvy – after all, the monks had shown more faith in the value of computer-generated answers than they themselves had – they notice an alarming sight: “Overhead, without any fuss, the stars were going out.”

That which is easily and cheaply acquired in bulk is valued lightly and rarely respected. The computer in Clarke’s story is oblivious to the value of the many letter strings it generates – they may as well be social-security numbers, or the tracking numbers on engine parts – and it certainly does not understand the import of the task as a whole. It falls then to an external agency to recognize their value, such as the monks of the story, or God, or (at the bitter end) the computer’s programmers. Value and meaning often arise from curbing our capacity to generate.

For instance, at any given moment there is a vast number of possible sentences we could utter, given our knowledge of the grammar of our language and a sizeable vocabulary of words. But the truly meaningful sentences that can be uttered in a context form just a tiny subset of this overall capacity. The linguist Noam Chomsky has argued that this capacity to generate is a necessary part of human creativity, and founded the highly influential field of generative linguistics to model the different ways in which humans can generate surface utterances from underlying (or deep) structures. Each grammatically-valid surface utterance is an answer in the sense that Picasso frowned upon (when he said "Computers are useless; they can only give you answers"), for there are so many possible answers, like there are so many names of God, that few have any real value in any given context.

The more we generate just because we can, then the more our true meanings get lost in the noise of mere possibility. In a very real sense, the measure of creativity is not how much we generate and keep, but how much we generate and discard. An ability is generate with automated means is called Mere Generation in Computational Creativity Research. The engineers in Clarke's story are skeptical of the monks and their beliefs not just on grounds of religion/superstition, but because they recognize mere generation when they see it. But in CC mere generation is not always so easy to spot, because the phenomenon is more of a gradable notion than a clear-cut and absolute one.


Scholarly Lite is a free theme, contributed to the Drupal Community by More than Themes.