Many people — engineers, software developers, scientists, philosophers — are now pondering the question of AI consciousness. I have thoughts too, about both what consciousness is and whether AI systems have it or could ever have it. Key philosophers who have helped shape how my thoughts on consciousness have evolved over the past few decades include David Hume, John Locke, Edmund Husserl, Bertrand Russell, Martin Heidegger, Maurice Merleau-Ponty, Paul Ricoeur, and Daniel Dennett.
To grossly summarize —
Hume reduces consciousness to a “bundle of perceptions” with no underlying self;
Locke defines personal identity as continuity of consciousness through memory;
Husserl foregrounds consciousness as the absolute foundation of all meaning;
Russell describes mind and matter not as two distinct substances but rather two different organizational patterns of the same underlying “stuff”;
Heidegger displaces consciousness in favor of being-in-the-world;
Merleau-Ponty re-grounds it in the body;
Ricoeur re-grounds it again, this time in language and interpretation;
Dennett dissolves it entirely into functional information processing.
But look, we’re already deep in the weeds — and I haven’t
even mentioned Aristotle, the Stoics, Giordano Bruno, Alfred North Whitehead, Galen
Strawson, or David Chalmers. There is too much to say for this post, so I’ll
save further elaboration for another day.
So why are we here today? In the process of pondering this subject
over many years, and now more recently in light of developments in artificial intelligence,
some linguistic musings have swum to the surface, and as is my wont, linguistic
and philosophical roots always tend to grow and ramify into literary branches
and leaves. I thought I’d share a few of them here.
Thanks to Isaac Asimov and others, everyone knows the word robot. George Lucas also popularized
mechanical people, but he called them “droids”, a shortening of androids, a word that is much, much
older than robot. The OED defines an
android as “an automaton resembling a human being”, first attested in 1727, if
you can believe that! It’s from Latin androīdēs,
in turn from Ancient Greek ἀνδροειδής
“human-shaped”.
But robot is a 20th
century coinage. The Czech playwright, Karel Čapek, needed a word for artificial
workers for his 1920 science-fiction play, R.U.R.,
an acronym for Rossumovi Univerzální Roboti (“Rossum’s Universal Robots”). Interestingly,
in Čapek’s play, the robots are made out of synthetic organic matter, i.e., artificial
flesh and blood, rather than metal and circuits and wires. This makes them closer
to what we might think of as androids — like the “andies” in Philip K. Dick’s Do Androids Dream of Electric Sheep? (replicants
in the film adaptation, Blade Runner).
But Čapek’s choice of robot is quite
deliberate.
Originally, when drafting the play, he was thinking of using
the word laboři or “workers” (from
Latin labor “labor, toil”), but it
didn’t feel like what he was really grasping at. His brother Josef, also a
writer, suggested roboti “*robots” —
at the time, an asterisk before the English word would have been appropriate,
as there was no such word in English (or Czech).
The word ultimately derives from Old Church Slavonic работа “servitude”
and рабъ “slave”, with descendants and cognates across the Slavic spectrum, all
connected to “labor, work”. The Czech robota
refers to “forced labor" specifically, or by extension, “a grind, drudgery,
tedious toil”. The German Arbeit “work”
is also related, as its consonantal silhouette makes clear. Compare Proto-Germanic
*arbaidiz with Proto-Slavic *orbota. The English word orphan is not too distantly related: as
the root evolved, one branch took on the meaning “deprived of a parent”, the
other “deprived of freedom”.
A robotnik in
Czech is “a forced laborer”, i.e., someone who performs robota, suggesting just what Čapek was looking for, and his brother
provided the coinage, roboti. The play
was popular, translated into 30 languages within three years of its
publication, and all those translators had to figure out what to do with roboti. The easiest choice was to retain
the word more or less as is, allowing for logical changes like losing the Czech
inflectional forms, and so on. All of which helped to spread the word “robot”
around the world, practically overnight, sowing imaginative seeds just as
science fiction was really about to explode.
And thus began a long tradition — or not quite began: there are earlier examples
of automatons, artificial men, clockwork people. The monster in Mary Shelley’s Frankenstein (1818) and Olympia in Jacques
Offenbach’s, The Tales of Hoffmann (1881)
— probably my favorite opera, by the way — are two notable antecedents (and not
the only ones). But in terms of our modern conception of robots, you could say Čapek
was the pater familias from whom other science-fiction writers’ robots
descended.
Isaac Asimov was influential here — without him, Čapek’s
word might have died out; after all, his play isn’t still being staged, is it?
— as were Ray Bradbury, Philip K. Dick, Fritz Leiber, Stanisław Lem, Ira Levin,
and plenty of others. I’d like to put in a special recommendation for a
favorite of mine, Walter Tevis’s Mockingbird
(1980), the same guy who wrote The Man
Who Fell to Earth, The Hustler, and
The Queen’s Gambit, by the way. A central
character in Mockingbird is the robot Dean of NYU. He wants to die, but his
programming won’t permit him to end his own life. He does try. Last week, I read
Glenn Dixon’s The Infinite Sadness of
Small Appliances (2026), a charming and cautionary tale about sentient
appliances, and I’ve just started Kazuo Ishiguro’s Klara and the Sun (2021), soon arriving as a feature film starring
Jenna Ortega. The moral implications of artificial people — for example, whether
it’s okay to mistreat or enslave them — are still being explored in fiction
and philosophy today, one hundred years after Čapek wrote his play.
Speaking of etymologies, did you know that the English words
“slave” and “Slavic” are related? The Latin sclavus,
originally referring to people of Slavic geographic and/or ethnic origin, came
to mean a person forcibly captured and sold into bondage, because so many of
these slaves in the Middle Ages, well, were
Slavs. You can read a lot more about this here,
but isn’t it ironic that a Slav, Čapek, would introduce to the world a new word
for “slave”, a word that would go on to stir up all kinds of engineering
innovation and thorny philosophical issues a century later?
Having said all this, dare I make some connections to
Tolkien here? Why wouldn’t I? I mean, have you met me? :)
There’s plenty one can say about slavery in Middle-earth —
the Nazgûl enslaved by the “Nine [Rings] for Mortal Men doomed to die”; the fallen
Númenóreans of the Second Age enslaving the denizens of Middle-earth; the large-scale
enslavement of Easterlings and Southrons, forced to till the fields around Lake
Núrnen to feed Sauron’s vast armies; enslaved Elves and Ents twisted by Morgoth
into Orcs and Trolls. Even among the Orcs, we see a class hierarchy, in which
the smaller, lesser breeds were called snaga,
“slave”, an “Orc-word, for a servant or slave, or like” [1]. Tolkien also often
used the word “thrall” in the same way.
I’ve written about the Orkish word snaga before, near the end of this post from almost twenty years ago. Tentatively, I’ve suggested that Old English snaca “snake” might have been an unconscious
source for Tolkien’s snaga. After
all, among the sound changes Grimm’s Law describes is PIE g > PGmc k, thus the
Proto-Indo-European root √sneg “to
creep”, “a creeping thing” begat Old English snaca. If we imagine Tolkien playing with how stops might become
voiced or voiceless in his own fictive historical linguistic Lautgesetze,
Orkish snaga could have developed
from something like Black speech *snaka,
*sneka. It fits, in my mind. But we
can’t know what Tolkien was thinking unless some commentary on it should come
to light. We’re really just having some fun.
But what about robots in Tolkien? Surely not! Or … actually,
yes? What about Tolkien’s “robot” dragons? Robot dragons, you say!? Yes, I do
say!
Many casual fans of The
Hobbit and The Lord of the Rings
are unaware of this particular oddity in Tolkien’s legendarium, but in the
earliest drafts of “The Fall of Gondolin”, written around 1916, Tolkien
describes iron monsters made by Melko (as Melkor/Morgoth was first conceived) in
the likeness of dragons, which could inflict horrors on the battlefield and
even transport Orcs inside them, like Odysseus’s Trojan Horse.
In the words of Tolkien:
Then on a time Melko assembled all his most cunning smiths and sorcerers, and of iron and flame they wrought a host of monsters such as have only at that time been seen and shall not again be till the Great End. Some were all of iron so cunningly linked that they might flow like slow rivers of metal or coil themselves around and above all obstacles before them, and these were filled in their innermost depths with the grimmest of the Orcs with scimitars and spears; others of bronze and copper were given hearts and spirits of blazing fire, and they blasted all that stood before them with the terror of their snorting or trampled whatso escaped the ardour of their breath; yet others were creatures of pure flame that writhed like ropes of molten metal, and they brought to ruin whatever fabric they came nigh, and iron and stone melted before them and became as water, and upon them rode the Balrogs in hundreds; and these were the most dire of all those monsters which Melko devised against Gondolin. [2]
Vivid and terrifying, no? Then, in the notes of his son and editor, Christopher explains:
In The Silmarillion the dragons that came against Gondolin were ‘of the brood of Glaurung’, which ‘were become now many and terrible’ whereas in the tale the language employed […] suggests that some at least of the ‘Monsters’ were inanimate ‘devices’, the construction of smiths in the forges of Angband. But even the ‘things of iron’ that ‘opened about their middles’ to disgorge bands of Orcs are called ruthless beasts’, and Gothmog ‘bade’ them ‘pile themselves’ […]; those made of bronze or copper ‘were given hearts and spirits of blazing fire’ while the ‘fire-drake’ that Tuor hewed screamed and lashed with its tail […]. [3]
This image was obviously suggested — not subtly! — by the armored
tanks of World War I, a new horror at the time, which Tolkien would have seen with
his own eyes during his grim experience in the Battle of the Somme. Perhaps
partly as a coping mechanism, he euhemerized them into dragons.
John Garth writes that these monsters “violate the boundary
between mythical monster and machine, between magic and technology” [4]. War reporting
at the time also likened German tanks to living creatures, like a “monster”, driven
by “a supernatural force”, an “iron caterpillar”, “the devil’s chariot”, and “and
other antediluvian and mythical monsters” [5]. Garth goes on to observe that
there is a whiff of science fiction about the army attacking Gondolin […]. From a modern perspective, this enemy host appears technological, if futuristic; the ‘hearts and spirits of blazing fire’ of its brazen dragons remind us of the internal combustion engine. But to the Noldoli the host seems the product of sorcery. ‘The Fall of Gondolin’, in Tolkien’s grand unfolding design, is a story told by an Elf; and the combustion engine, seen through enchanted eyes, could appear as nothing other than a metal heart filled with flame. [6]
Are these really robots in the modern sense? Maybe not Čapek’s or Asimov’s sort of robots, but with semi-autonomous machines, capable of some independent action, “given hearts and spirits”, but controlled by a greater power, Tolkien seemed to be reaching for a similar idea. He would, of course, stop short of any such automaton developing actual consciousness; this would have been very much incompatible with his Roman Catholic beliefs. Aulë’s Dwarves would have remained mere automata, requiring Aulë’s constant attention and control, were it not for the grace of Ilúvatar in endowing them with the divine spark of independent consciousness. And in any case, Tolkien dropped the whole idea of “robot dragons” in his later elaboration on the Gondolin story, leaving them behind in these older drafts.
What would Tolkien have thought of today’s robots? A
“contrivance of the Enemy”, no doubt, to quote Boromir, a burlesque of genuine
creation and consciousness, which are the provenance of God alone. I don’t
think Tolkien would have admitted robots into the domain of positive subcreation
by Man humbly imitating of God, because he would have felt the motives were evil,
not good; arrogant, not worshipful; mocking, not making.
[1] Parma Eldalamberon
17 (2007), p. 79; and see The Return of the
King, Appendix F, p. 409.
[2] The Book of Lost Tales
2, p. 170.
[3] Ibid., p. 213.
[4] Garth, John. Tolkien
and the Great War: The Threshold of Middle-earth. Houghton Mifflin, 2003,
p. 220.
[5] Quoted in Garth, p. 221.
[6] Garth, p. 221.
