Erik Anderson
Houston, Texas

Question posed to ChatGPT: What is the “Golden Rule”?
ChatGPT answer*: The “Golden Rule” is a principle found in many different cultures and ethical traditions and often phrased as “Treat others as you would like to be treated.”¹

Presently, artificial intelligence (AI) applications such as ChatGPT are exceptional at reiterating information, but do not as yet have the loosely-defined thing we call consciousness, i.e., awareness of oneself and surroundings.² The question that haunts us, as demonstrated in movies like 2001: A Space Odyssey and Terminator where conscious machines become murderous towards their human creators, is: If artificial intelligence achieves consciousness, will it be benevolent? Surprisingly, an eighteenth-century physician-philosopher may have an answer, and this involves the Golden Rule. Dr. David Hartley was an Enlightenment physician and philosopher whose book Observations on Man, His Frame, His Duty, and His Expectations³ was one of the earliest to propose a physiological basis for how the mind works. In so doing, he laid the foundations for association psychology, the Hebbian learning model, and even modern-day neural networks.

David Hartley was born in the County of York, England in 1705.⁴ His father was the Vicar of Armely, and Hartley’s childhood aspiration was actually to become an Anglican clergyman.⁵ In 1722 he entered Jesus College at Cambridge, but upon leaving in 1729 with a master’s degree, his professional interest changed to medicine.⁶ It is unclear what specifically caused this ecclesiastical drift; however, some historical context of the era can perhaps explain Hartley’s decision and frame his later scholarly interests.

By the time Hartley began his formal education, the revolutionary ideas of Sir Isaac Newton and John Locke had become ingrained in scholarly curricula and had made a profound impression on him. Newton’s famous 1687 Principia engendered a fervor to apply mathematics to science.⁷ It dovetailed with Locke’s 1690 An Essay Concerning Human Understanding, which galvanized empiricism. In the Essay, Locke also introduced “associationism” to England as the way our minds operate and learn, an idea first posited by Aristotle.⁸ Hartley’s crowning work—the Observations—would be organized around a Newtonian Doctrine of Vibrations and a Lockean Doctrine of Association.

After leaving Cambridge, the neophyte Dr. Hartley set up a medical practice in Newark and began to conceive his book on the mind and body. Though he left the clerical path, Hartley remained a devout Christian his entire life and believed in universal salvation. This is illustrated in a letter to his childhood friend, Rev. John Lister, in 1735:

I have wrote two small treatises about a year and a half ago . . . the first begins with shewing that all our intellectual pleasures and pains are formed either immediately or mediately from sensible ones by association . . . It then proceeds to show that Benevolence is the best means of obtaining private happiness . . . that our natures are so formed and so adjusted to the system of things that we must from the Law of Associa­tion at last become benevolent, and consequently that all must sometime or other be happy.⁵

Hartley would formalize these ideas over the next fourteen years into his Doctrine of Association and Law of Universal Happiness through benevolence, presented in his penultimate Observations.³ Hartley also advanced a physiological foundation for this metaphysical system, which he named the Doctrine of Vibrations. This rested upon Newton’s idea that the nerves conduct information through microscopic vibrations. By unifying the Doctrines of Association and Vibrations, Hartley presented one of the earliest neurophysiological explanations for how the mind forms and functions.

Hartley believed in Locke’s theory that the mind is a blank slate at birth, that the origin of all ideas are external sensations, and that knowledge is generated from associating these ideas into more complex ones.^3,9 Thus, all knowledge is the result of association and based on external sensation. Hartley extended Locke’s theory by applying Newton’s mechanism, explaining that sensory information is received by the peripheral nervous system and transmitted to the brain by vibrations on nerves. These vibrations persist in the brain, and owing to the multitude of neuronal interconnections, associate and evolve to form complex ideas like language. Hartley explains we learn words in the order of (1) sound (2) organs of speech (3) vision and (4) actions of writing, with 1 being the reverse of 2, and 3 the reverse of 4.³ First, we ascertain ideas associated with the sounds of words. Then, speech helps us solidify their use. Visual representation helps us extend the signification of words and facilitates association with new ones, e.g., with definitions and descriptions. And finally, writing helps us retain and recollect the significations gained through vision. All of this helps fix ideas within our mind, allowing us to build even more complex thoughts, and, ultimately, knowledge.

Hartley died in 1757, but his Observations would live on to impress a diverse set of luminaries, from the iconoclastic chemist Dr. Joseph Priestley¹⁰ to the eloquent bard Samuel Coleridge.¹¹ James Mill claimed Observations as the most crucial precedent for his 1829 Analysis of the Phenomena of the Human Mind,¹² making Hartley instrumental to the founding of associationism psychology. Additionally, in modern neuroscience and machine learning, associationism forms the theoretical basis for the influential Hebbian model of learning and artificial neural networks (ANNs). The Hebbian model asserts synaptically connected neurons that activate together will strengthen these connections—e.g., associating two things temporally will strengthen their neuronal links in the brain.¹³ Associationism also underlies the ANN architecture that powers AI bots like ChatGPT.¹⁴ In these networks an artificial neuron is represented as a node, and each node is interconnected, i.e., associated, with many other nodes. On receiving an input, these interconnections allow nodes to affect the threshold with which others activate, just as writing words affects our retention of their visual representation and signification. Thus, Hartley’s work is an ancestral branch in the phylogeny of artificial intelligence (AI), and even offers an optimistic answer to the question of whether a benevolent AI is programmable.

For Hartley, while humans have free will, their fundamental nature is formed of “self-interest, benevolence, and piety,” which “all concur to move and exalt our affections.”³ This philosophy was rooted in his lifelong devotion to and study of Christian doctrine. A striving for universal happiness through benevolence towards oneself is what results in kindness towards others—i.e., the Golden Rule, or loving thy neighbor as thyself, is the “algorithm” that drives human civility and kindness. Hartley believed this to be a natural outgrowth of our associative mind, which works from infancy to correct and perfect the unconscious neural processes. Embracing the mathematical concept of the limit, like Newton’s calculus, if time is taken to infinity, Hartley argues that all people eventually become benevolent.

The corollary for AI reasons that a similar law of universal happiness through benevolent action could be trained in an ANN. Consciousness necessitates a concept of oneself, and if kindness towards oneself could be programmed, it might be the wellspring of empathy towards others. If done, the benevolence residing within all humans, of which Hartley tried to remind the world, may also be possible for artificial intelligence.

Note

* ChatGPT answer truncated and slightly edited

Acknowledgments

Special thanks to the Halifax Antiquarian Society for their help in acquiring “The Correspondence of Dr. David Hartley & Rev. John Lister.”

References

1. OpenAI. ChatGPT. Accessed January 31, 2023. https://openai.com/blog/chatgpt/.
2. “Consciousness.” Merriam-Webster Dictionary. Accessed February 15, 2023. https://merriam-webster.com/dictionary/consciousness.
3. Hartley, D. Observations on Man, his Frame, his Duty, and his Expectations. Bath, London: Samuel Richardson, 1749.
4. Langdon-Brown, W. “David Hartley: Physician and Philosopher (1705-1757).” Proc R Soc Med. 34:11-17.
5. Trigg, WB. “The Correspondence of Dr. David Hartley and Rev. John Lister.” Halifax Antiq Soc, 1938.
6. Webb, ME. “The Early Medical Studies and Practice of Dr. David Hartley.” Bull Hist Med, 1989;63(4):618-636.
7. Gleick, J. “Isaac Newton.” Vintage, 2004.
8. Barnes, J. Complete Works of Aristotle: The Revised Oxford Translation. Vol. 1. Princeton UP, 1984.
9. Locke J. “An essay concerning human understanding.” Kay Troutman, 1847.
10. Priestley, J. Hartley’s theory of the human mind, on the principle of the association of ideas; with essays relating to the subject of it (2nd ed.) London: J. Johnson, 1775.
11. Haven, R. “Coleridge, Hartley, and the mystics.” J Hist Ideas, 1959;20(4):477-494.
12. Mill, J. Analysis of the Phenomena of the Human Mind. Vol. 1. Longmans, Green, Reader, and Dyer, 1878.
13. Hebb, DO. The Organization of Behavior: A Neuropsychological Theory. Psychology Press, 1949.
14. Bishop, MJ. “History and philosophy of neural networks.” Encyclopedia of Life Support Systems, 2015.