Thoth’s invention
What information machines might leave us free to become
In Plato’s Phaedrus, Socrates tells the story of the Egyptian god Theuth, or Thoth, who presents writing to King Thamus as a remedy for memory and wisdom. The king is unconvinced. Those who rely on written marks will neglect the exercise of memory, and those who encounter many accounts without instruction will acquire the appearance of wisdom rather than wisdom itself.[1]
The objection concerns more than forgetting. A person may possess the products of an intellectual activity without having developed the capacities through which those products acquire meaning. Writing preserves what someone has said, but cannot answer when questioned. It appears to speak intelligently, yet offers the same words again.
Language models seem to overcome precisely this limitation. They answer questions, respond to objections, elaborate arguments, and work through unfamiliar problems. The written word has become something we can converse with.
Yet the ability to answer does not settle the question of wisdom. An explanation can be produced without anyone acquiring the capacity to judge it. A problem can be solved without its solution changing the person who posed it. These machines may make thought available without making thinkers.
There is another way to understand what such machines might accomplish. In Cybernetics and the Origin of Information, published in 1954, Raymond Ruyer welcomes the possibility of liberating human beings from the mechanical demands of their own civilisation. Information machines would relieve the brain just as industrial machines had relieved the muscles. The accumulated knowledge of libraries had already become too great for an unaided mind to employ. The era of the intellectual foot soldier, he suggests, was coming to an end.[2]
Ruyer is not simply an enthusiast. His critique concerns the postulates of cybernetics, not its immense practical promise. Drawing upon Bergson, he distinguishes the enlargement of human technical powers from the cultivation of the soul. A supplement of brain is not a supplement of soul. Self-regulation is not wisdom, although it may provide the conditions in which wisdom becomes possible.
This distinction is easily lost. Machines that free us from intellectual drudgery may permit forms of inquiry previously beyond our reach. But they may also leave us producing more of what we already produced, only faster. Where the value of intellectual work is measured principally by its outputs, greater productive capacity need not produce any corresponding development of the person.
Much of our institutional life already turns upon the visible products of competence: reports, plans, publications, demonstrations, decisions. Language models make these forms extraordinarily cheap to produce. We may become increasingly successful at obtaining what our institutions reward while becoming less capable of the activities those rewards were intended to cultivate.
The same difficulty attends the larger promises of automation. The production of a plan is not the achievement of agreement. An accurate prediction does not determine what ought to be done. Political, moral, and personal transformation involves forms of work that cannot simply be completed by generating their apparent results. Extraordinary technical progress may therefore coexist with much slower social development.
Martin Heidegger approaches something of this problem from another direction. In The End of Philosophy and the Task of Thinking, he identifies cybernetics with the increasingly technological organisation of human activity. The sciences become capable of ordering and regulating what was once the concern of philosophy. Yet the essay begins with an unusual qualification: an answer to its question, if one could be given, would consist in a transformation of thinking, not in another proposition.[3]
The distinction is not between true and false statements. It concerns what it would mean to undergo a change in one’s relation to the matter of thought. No collection of correct propositions can itself establish that such a transformation has occurred.
In an earlier essay, Heidegger considers Hölderlin’s line that man dwells poetically upon the earth. The poetic is not an ornament added to an otherwise complete existence. It concerns the taking of a measure through which dwelling becomes possible at all. To dwell is not merely to occupy a place. Nor is building exhausted by the construction of something that can be occupied.[4]
There is an echo here of another passage in the Phaedrus. Socrates contrasts the gardens of Adonis, where seeds produce an attractive display in eight days, with the patient cultivation through which a farmer brings a crop to maturity. Both involve growth, but the purposes and the processes differ. The appearance of the result does not tell us what has been cultivated.[1:1]
Perhaps the same is true of intellectual life. We can now obtain in moments something resembling the outcome of months of work. The resemblance may be useful, even remarkable. But what has grown, and where?
This is not an argument for abandoning information machines. Ruyer was right to recognise the extraordinary possibilities of relieving the brain of servile labour. Nor must the use of an external instrument diminish the person employing it. Writing itself made possible forms of sustained thought that could scarcely have existed without it.
The question is what we do with the freedom these instruments make available.
Warren is being developed from this concern. We want a place for thought to dwell: somewhere an observation can be retained without becoming a conclusion, an objection can survive the summary of an argument, and an unfinished undertaking can be resumed through different instruments and interlocutors.
Such a place cannot itself provide the transformation of thinking. It can preserve records, relationships, and histories; it can make information available while allowing its use to remain under deliberate control. It cannot determine what that information ought to mean, or what someone should become through their engagement with it.
But perhaps it can help make room for that engagement.
The invention of writing did not end thought. Neither will the invention of machines that answer. Their significance lies not only in what they allow us to produce, but in what they may leave us free to cultivate.
Sources
Plato. Phaedrus, 274c–276d. Translated by Harold N. Fowler (1925). See the myth of Theuth, the discussion of living speech, and the gardens of Adonis. ↩︎ ↩︎
Raymond Ruyer (1954). La Cybernétique et l’origine de l’information. English translation: Cybernetics and the Origin of Information, translated by Amélie Berger-Soraruff, Andrew Iliadis, Daniel W. Smith, and Ashley Woodward (2023), Introduction, pp. 8–9. ↩︎
Martin Heidegger (1964). The End of Philosophy and the Task of Thinking. In On Time and Being, translated by Joan Stambaugh (1972). ↩︎
Martin Heidegger (1951). “…Poetically Man Dwells…”. In Poetry, Language, Thought, translated by Albert Hofstadter (1971), pp. 209–227. ↩︎