Know Thyself: Augustine for the Present Age

Among the many problems associated with the rapid growth of artificial intelligence lies an existential one: the problem of how we conceive ourselves as human subjects, as agents of human intelligence in some kind of similarity to artificial intelligence, and yet dissimilarity to its artificial status. It is this existential dimension that Pope Leo has emphasized in his first encyclical, Magnifica Humanitas, where he identifies the “age of AI” as one where “human dignity is threatened by new forms of dehumanization” (§15). One of these forms comes from the trans-humanist and post-humanist rhetoric of many of the architects of AI, who see it as the next stage of human evolution and an advance beyond the natural limits of human intelligence (§§115-17). Will AI be “more human than human,” as Elon Musk claims? In order to address these concerns, we must raise the question of what human intelligence actually consists of. How do we, as we increasingly compare ourselves to AI, understand our own “intelligence,” which AI might mimic, advance, and/or usurp?

Perhaps these concerns are overblown; perhaps there is no “existential” issue, and perhaps no threat of “usurpation.” But evidence continues to mount that the architects of these programs have a vision of the future that is dangerous to human beings, and dangerous precisely on the existential level of how we understand ourselves and exist and act in the world as human beings. For example, earlier this year Sam Altman, the CEO of OpenAI, prophesied a world where “intelligence is a utility, like electricity or water, and people buy it from us on a meter.” The moral turpitude of such a statement should go without saying. But is it even a reasonable claim in the first place? Is intelligence the kind of thing that can be owned by a third party and externally distributed to human customers and consumers, like water or electricity?

No, it is not. In order to help us think about these issues, I suggest we turn to one who knew nothing of artificial intelligence: St. Augustine of Hippo, and in particular his massive work, De Trinitate (On the Trinity).[1] In an effort to uncover a “likeness” of the Trinity in the human mind (VIII, 8), Augustine performs a sustained investigation into his own subjectivity, consciousness, and intelligence.[2] He identifies the mind’s knowledge of itself as the innermost element of consciousness and intelligence, and thus as the site of the most profound likeness to the Trinity. It is precisely the kind of self-appropriation that Augustine performs in his development of this “trinity of the mind” which we, sixteen centuries after De Trinitate, urgently require. By following Augustine in this investigation, we will develop a standpoint from which we can understand our identity as human subjects invested with intelligence, and from which we can begin to navigate the possibilities of AI exceeding or replacing this intelligence.

Getting Mixed Up

I suggest we begin in medias res, in the middle of Book X of De Trinitate. Augustine is grappling with the ancient axiom of the Delphic oracle, gnothi seauton—know thyself. But where does the mind go to look for itself? (X, 11). Augustine draws our attention to the danger of the mind simply identifying itself with its objects, as the same in nature as the physical objects of which it forms images and concepts. In an especially vivid expression, Augustine speaks of the “force of love” with which the mind thinks about things for a long time and becomes “stuck to them with the glue of care”; by this force of love, it develops the tendency “to drag these along with itself even when it returns to thinking about itself” (X, 7). By making images of the things it knows, the mind gives something of itself to its objects, and when it joins itself with these images it can make the mistake of thinking of itself as just another thing of the same sort. It becomes lost in these images of things it has perceived with the senses and fails “to see itself alone” (X, 11). Because the mind spends so much time thinking about bodily objects, and thinking about them by virtue of images which it forms for itself, it makes the mistake of thinking of itself as just another bodily object; it “drags” them into itself when it returns to thinking about itself, and thinks about itself as just another sensible and imaginable thing. In other words, it “mixes itself up” with the objects of its knowledge (X, 11).

In general, Augustine contends that the more we focus our intelligence on other, external things, the more we actually become conformed to those things and begin to see ourselves in their image. AI simply intensifies this problem: the more we invest our intelligence into AI, outsourcing the patient labor of thought to a third party, the more we become “stuck” to it and imagine ourselves as images of it, rather than it as an image of us. As Jeffrey Bishop argues, this is the innate problem of all technology, and of AI in particular: we imagine ourselves as the masters of such technologies, but in reality they “recursively create us, and set out for us what it is possible to imagine.”

We can see this recursive creation in our language, where we use personal metaphors for AI—verbs like “think” and “know,” the category of “intelligence” itself, as well as gendered names and pronouns—while using mechanical metaphors for ourselves, as Warren Kinghorn observes. This results in an exchange of identities where persons grant personal qualities to machines that are in fact incapable of receiving them, while personhood itself is effaced in the offshoring of personal (intellectual and rational) activities. From an Augustinian perspective, this recursive creation is another way—indeed a uniquely powerful way—that the mind can get “mixed up” with something outside itself and begin to forget its own identity.

As soon as we start to think this way, we also quickly recognize our own inferiority: if human intelligence is simply a “natural” version of “artificial” intelligence, then it is in many ways a lesser image, a pale imitation of the extraordinary things that AI can do. If our minds are just large language models, but simply lower on the evolutionary scale, then they may be inferior models, and we should invest all our technological capital into developing superior programs that can perform the labor of intelligence in more effective and efficient ways than we can, as we have designed machines to do our laundry, dishwashing, etc. But—Augustine wants to tell us—intelligence consists in something other than these tasks that we can outsource to AI. It consists in something more fundamental than these tasks, and by cultivating our self-knowledge we will be able to recognize this.

Augustine wishes to draw our attention to the categorical difference between knowledge of other things and knowledge of oneself. When the mind is commanded to know itself, Augustine says later in Book X, it is not commanded to know an absent being, like the cherubim and seraphim, but to recognize its own presence to itself (X, 12). Augustine has a variety of ways of expressing this strict difference between self-knowledge and other-knowledge, present-knowledge and absent-knowledge. Augustine expresses this difference by noting the dissimilarities between the eye and the mind.

While admitting the unique fittingness of the metaphor of vision for knowledge—I “see” what you mean, as we often say—Augustine uses this fittingness as an occasion for recognizing the profound difference between the mind and the physical eye (XI, 1). For example, the eye sees external things without seeing itself, but the mind can only know other things because it knows itself first (IX, 3). If the eye wants to see itself, it needs an external object—a mirror—to help it do so; but if the mind wants to know itself, it needs no external medium, and needs precisely to abstract from all these external media (X, 5). To think of the mind as a kind of eye, and nothing more than an eye, is thus another way of “getting mixed up.”

Augustine draws a threefold distinction between things the mind knows through the senses, things it knows through testimony, and things it knows through itself (XV, 22). When we know things outside ourselves, our knowledge is “co-generated” by ourselves and by those external things; but when the mind is the object of its own knowledge, it is the sole parent of that knowledge, as both the subject and object of knowing, the knower as well as the known (IX, 18).

Augustine therefore draws another distinction between knowledge of the mind in terms of generic or specific knowledge (i.e. classifying the mind as a certain kind of substance, a certain kind of soul, à-la Aristotle’s De anima), and knowledge of the mind in oneself, which one can see for oneself and tell to others (IX, 9). Following Augustine, Bernard Lonergan makes a similar distinction between “soul” and “subject,” with the former pertaining to a metaphysical consideration of the soul’s objects, acts, and potencies, and the latter an introspection where one discovers not only the acts of consciousness but also their intending subject. Lonergan contends that Augustine’s “trinity of the mind” falls, at least in the first place, in the latter category, in a “phenomenology of the subject” before a “psychology of the soul.”[3]

But Augustine does not wish simply to tell his readers what is going on in his own mind, as if a kind of psychological autobiographer. Reflecting on the role of literature, Augustine believes that the reader should not simply take the author at his word, but should discover with the writer that these things are true, and discover this in himself or in truth guiding the mind (XIV, 9). Lonergan therefore remarks: “As Augustine’s discovery was part and parcel of his own mind’s knowledge of itself, so he begged his readers to look within themselves and there to discover the speech of spirit within spirit.”[4]

To discover this self-knowledge requires a movement of the will. Because the mind becomes “mixed up” with its objects and forgets itself in them, it requires an act of will—a decision—to cultivate one’s self-knowledge and all it entails. This is implied by the fact that “know thyself” is a command, which one can follow or not. Augustine also states this explicitly when he states that the mind must “turn to itself the interest of its will” in order to recognize itself (X, 11). Augustine has a variety of reasons for recommending this kind of “non-bodily turning around” (IX, 3), and today we have urgent reasons of our own to add to Augustine’s. As significant technological, cultural, and economic forces seek to transfer intelligence from human minds into “artificial intelligence” (and the corporations which own these programs), it requires a decision to resist these forces and to cultivate the intelligence that they wish to usurp.

Fortunately, Augustine promises to guide his reader into recognizing the uniqueness and profundity of self-knowledge, this “subjective” knowledge which constitutes each human being as a unique subject, distinguishable from all other objects, whether natural objects or artificial intelligence.

Hidden Knowledge

Augustine begins Book X with the problem of the mind seeking to know itself. By seeking to know itself, it apparently does not know itself; but this is impossible, Augustine argues. Even if it seems not to know itself, the mind nevertheless knows some things, such as other minds; it therefore knows that it knows. Now, if it knows that it knows, then it necessarily knows itself knowing, and therefore knows itself (X, 5). Similarly, in the very act of seeking itself, it knows itself seeking and therefore “already knows itself” (X, 5). “It follows then that [the mind] simply cannot not know itself,” Augustine concludes, “since by the very fact of knowing itself not knowing, it knows itself . . . For it knows itself seeking and not knowing, while it seeks to know itself” (X, 5; cf. X, 16).

Augustine aims to prove that the mind knows itself, even and precisely when it seeks itself and seeks to know itself. In the previously cited passage, the most important word is “already”: the mind “already knows itself” when it seeks to know itself, and this act of seeking rests on this prior self-knowledge as its hidden ground. Augustine also uses language of “availability” for this truth about the mind: the whole of the mind is “available to itself” as it seeks for itself, and none of the mind is missing (X, 6). This also correlates with the language of “presence” indicated in the previous section: it is impossible for the mind to know other minds without knowing itself, because nothing is more “present” to the mind than itself (X, 5; cf. XIV, 14). With these key terms—already, available, present—Augustine seeks to show that all acts of the mind proceed out of a more fundamental availability and presence to oneself which constitutes a knowing of oneself; in other words, the apparent poverty of having to seek to know oneself paradoxically arises out of a profound wealth of self-knowledge, a wealth which is nonetheless hidden from our view.

Having recognized this wealth of self-knowledge, Augustine draws another key distinction between knowing oneself (se nosse) and thinking about oneself (se cogitare): even when one does not think about oneself, one still knows oneself (XIV, 7; X, 19). “Thinking about oneself” includes all deliberate acts of thought in which the mind directly turns its attention to itself, including Augustine’s composition of De Trinitate, and the thought processes that enabled it, as well as the reader’s participation in this thought. But all these specific acts presuppose the prior state of “knowing oneself” even when one does not think about oneself. Augustine refers to this primordial self-knowing as memory, as a kind of “confidential file of awareness” (XIV, 8; cf. XV, 40).

What Edmund Hill translates as “confidential file of awareness”—arcana notitia—could more directly be rendered as “hidden knowledge” or “secret awareness.” By directing his reader’s attention to this so-called “hidden knowledge,” Augustine does not mean to indulge in esotericism or obscurantism; on the contrary, he is attempting to show his reader that every mind necessarily enjoys this “hidden knowledge,” by which it knows itself even when it does not think about itself. Paradoxically, this seemingly opaque “hidden knowledge” can become most luminous to us, and Augustine aims to reveal precisely this hidden clarity.

Augustine’s derivation of this “hidden knowledge” represents one of his most salient insights for the age of AI. Much of the discourse concerning AI centers (and quite reasonably so) on its extraordinary products, which seem to outstrip the capabilities of human intelligence, at least in speed and efficiency (cf. Magnifica Humanitas §99). Augustine is contending, though, that the essence of intelligence lies in something else, neither in the products of intelligence nor even in the processes of thought which give rise to these products (and which AI mimics). Rather, the innermost nature of intelligence lies precisely in this “hidden knowledge,” this wealth of self-knowledge “out of which the gaze of thought is formed” (XIV, 8). The key word in this formulation is “out of” (unde). Augustine is drawing our attention away from both the product of knowledge as well as the thought-processes pertaining to knowledge, and toward the immanent source of knowledge, that “out of which the gaze of thought is formed.”

Rowan Williams, one of the most eminent contemporary interpreters of Augustine, highlights the fact that self-knowledge for Augustine is not simply another piece of information, and that this provides a crucial point of philosophical interest today in navigating the “mental” status of artificial intelligence.[5] For Augustine, whatever information the mind possesses, this is not the most fundamental element of the mind. In uncovering this most fundamental element as “memoria” and “arcana notitia,” Augustine is developing a thoroughly non-informational understanding of the mind which is crucial for comparing “artificial intelligence” to authentic human intelligence. The source and “nature” of the latter resides in this surplus of knowledge, intelligibility, and meaning out of which thinking arises, and which is intensely personal and “subjective” (in Lonergan’s sense) in nature. When we see the nature of intelligence as such, we can see the inescapably improper and metaphorical sense of “intelligence” in the moniker “artificial intelligence,” because it is precisely this abiding personal subject, constituted in self-knowledge, which AI lacks.

Augustine provides us with further reflections that can help clarify and solidify this point. At a few points in De Trinitate, Augustine recommends prescinding from everything outside the mind in order for the mind to discern itself from what is other (e.g. X, 12). Augustine does not mean thereby to embrace any kind of idealism or solipsism for which the mind alone is true; when he distinguishes between what the mind knows through the senses and what the mind knows through itself, he explicitly upholds the former as a valid site of knowledge (XV, 21). Nevertheless, he remains confident that the mind can discern itself as distinct from everything outside itself, even while maintaining the reality of things outside itself, and he insists that we conduct this exercise in order to discover the necessary role that the mind plays in every act of knowledge, a role which, despite its fundamental status, frequently becomes obscure to us.

By “drawing off what it has added to itself,” the mind will be able to discern itself from what is other by recognizing itself as “more inward” than sensible things (X, 11); by putting aside all outward things, we will be able to discern the “nature” of the mind (X, 14). Though the mind gives something of itself by forming images and concepts of its objects, Augustine contends that that is “more truly mind” and more truly “rational intelligence” that the mind withholds from its objects, i.e., that which it keeps apart by making judgments on its objects (X, 7). In order to further distinguish the “inner” of the mind from the “outer,” Augustine recommends we pay attention to the various acts of the mind (such as judgment) which originate with the mind and are not simply reducible to the objects and information conveyed in these acts. Augustine therefore contends, “If [the mind] refrains from affixing to itself any of these image-bound objects of its thoughts in such a way as to think it is that sort of thing, then whatever is left to it of itself, that alone is what it is” (X, 16).

Augustine observes the many arguments about the nature of the mind (arguments which continue today), as some contend that the mind is nothing but a body or an organization of body, such as air, fire, or (as we are inclined to think today) brain (X, 14). But all of this is uncertain and subject to debate. Augustine therefore recommends that we disregard (at least for the moment) what is uncertain and focus on what is certain:

Nobody surely doubts, however, that he lives and remembers and understands and wills and thinks and knows and judges. At least, even if he doubts, he lives; if he doubts, he remembers why he is doubting; if he doubts, he understands he is doubting; if he doubts, he has a will to be certain; if he doubts, he thinks; if he doubts, he knows he does not know; if he doubts, he judges he ought not to give a hasty assent. You may have your doubts about anything else, but you should have no doubts about these; if they were not certain, you would not be able to doubt anything (X, 14).

Even if one doubts, Augustine is saying, one cannot doubt the very act of doubting, or the very act of thinking, judging, etc. Those who question whether the mind is air, fire, or brain fail to notice what we discovered above: that the mind knows itself even when it is looking for itself (X, 16). If we want to know what the mind most truly is, we need to attend to this self-knowledge and to the acts which proceed out of it.

Augustine thus makes the following recommendation: if you want to know what intelligence is, do not—at least in the first place—ask Google or an AI chatbot (perish the thought) about what neuroscientists, psychologists, or philosophers say about intelligence. Rather, attend to your own acts of intelligence by which you pose these questions, by which you seek their answers, by which you judge their truth. It is only in these acts that the meaning of intelligence can be determined, and can serve as the foundation for a comparison with AI: “When a human being is able to think about the nature of his consciousness and find out what is true about it, he will not find it anywhere else but inside himself” (XIV, 8).[6]

For Augustine, the key characteristic of all the aforementioned acts is that the mind knows them “in itself” (X, 16). Unlike their extrinsic objects, these acts cannot be doubted because they arise out of that ineradicable wealth of self-knowledge that we discovered above. As Augustine puts it, it is “not from the outside” that the mind receives what it remembers and loves within this self-knowledge (XIV, 11). Though many things are adventitious to the mind, “the mind, after all, is not adventitious to itself” (XIV, 13). By momentarily prescinding from everything that comes to the mind from without, Augustine has discovered the nature of the mind in that which does not come from without, in that which constitutes the most inward source of its life and its acts—its wealth of self-knowledge out of which all acts of thought are formed, and which exists distinctly from everything that comes to it from without.

Herein also lies Augustine’s most important idea for understanding the status of “artificial intelligence,” which, unlike the mind, is precisely adventitious to itself. Whatever remains after removing everything adventitious to the mind—that is what the mind is, according to Augustine (X, 16). What remains of AI when you remove everything that comes to it from the outside? Nothing. Everything comes to it from the outside: the programming of a human engineer, the “training” of human agents, the data it receives and processes, the prompting of a human agent who needs their homework done or their vacation planned. If AI-generated literature, art, or music seems to lack depth and soul—it is because it literally does. If it is improper to describe artificial intelligence as artificial intelligence, it is exactly right to describe it as artificial intelligence: unlike the mind, which has a nature discerned in its abiding presence to itself and knowledge of itself, AI has no nature, no immanent ground of life and being which persists independent of its extrinsic influences.

The Drama of the Image of God

If Augustine encourages us to consider the source of intelligence over its products and objects, he simultaneously insists that we move beyond this immanent source toward its transcendent source. Augustine considers all of the preceding necessary for determining what constitutes the “image of God” in the human being, which will be the most intense likeness to God in creation and thus the most adequate site for the “likeness” to the Trinity for which Augustine is searching (XI, 8). In other words, by laboring to uncover the immanent nature of the mind in its “hidden knowledge” of itself, Augustine is simultaneously uncovering the mind’s transcendence of itself and its orientation toward the God who is “more inward than my inmost and higher than my highest” (Confessions III, 11).

Augustine identifies the mind as the image of God to the extent that it has the “capacity to use reason and understanding in order to understand and gaze upon God” (XIV, 6). This “capacity to reason and understand” is provided by the mind’s immanent constitution as the kind of thing that always already knows and loves itself; by being this kind of thing it is capable of knowing and loving God. The mind’s natural love of itself constitutes its potency to cling to God (XIV, 20), and the mind is always “in God” to the extent that it is able by nature to remember him (XIV, 21). Augustine therefore identifies the mind’s trinity of remembering, understanding, and loving itself as “already the image of God,” which is innate to the mind and cannot be lost (XIV, 11).

The immanence of the mind thus breaks forth into transcendence: “This trinity of the mind,” Augustine famously writes, “is the image of God not because the mind remembers and understands and loves itself, but because it is also able to remember and understand and love him by whom it was made” (XIV, 15). This passage might seem to contradict Augustine’s earlier claim that the mind is “already the image of God” when it remembers, understands, and loves itself (XIV, 11). But it actually confirms and deepens this claim: the mind is the image of God because it “is able to remember and understand and love him by whom it was made,” and it receives this ability in its natural knowledge and love of itself.

All of this provides important ontological considerations for Augustine, i.e., truths about the being of the mind that cannot be lost, even in sin (XIV, 6, 11, 19). But the mind as the image of God is not merely a stable ontological reality; it is also situated within a drama of creation, sin, and grace where it can either forget and reject its status as the image of God or become more true to its innate dignity. If the mind does not remember God, then it is foolish, even as it always remembers itself (XIV, 15). Similarly, the mind that always knows itself can forget itself and begin to bury and obscure the identity which it always possesses (X, 7). Even more starkly, Augustine contends that the same mind that is constituted in its love of itself can grow to hate itself: “The man who does not love God, even though love of himself is innate in his nature, can be said to hate himself” (XIV, 18). For Augustine, the condition of the fallen human being is precisely this: to exist within this contradiction of self-love and self-hate, of remembering and forgetting. To be the image of God has incredibly high stakes: the mind constituted in its love of itself is the same mind that learns to love itself unduly, and in that twisted love grows to hate itself and hate others, exchanging love of God for love of self, justice for power, service for domination (De Trin. XIII, 18; City of God XIV, 28, passim).

As Augustine interprets it, the ancient axiom “know thyself” therefore has a fundamentally moral valence: by knowing itself the mind should live according to its nature, placed beneath God and above other created things (X, 7). But because of its “twisted desires,” it cannot do this on its own and must receive grace from God in order to preserve the honor of its status as image of God (XII, 16). When it finally turns to love God, it can then heed God’s commandment to love its neighbor (XIV, 18). This grace also enables it to achieve its highest dignity as the image of God, since it is only in seeing God face to face that it becomes most like him and achieves the deepest possible likeness to God (XIV, 23, 24; XV, 21, 26).

At risk of stating the obvious, AI cannot do any of this: it cannot know itself and love itself, it cannot know and love God or know and love its neighbor, it cannot sin, repent and receive grace, and it cannot see God. For Augustine, “intelligence” is that which allows us to do all these things. Anything that does not exist for this purpose and cannot achieve these ends is not intelligent.

Conclusion

Along with Augustine, we have seen how the mind has a tendency to obscure itself by attaching itself to external things and confusing itself with them; how the mind is immanently constituted in an abiding act of self-knowledge that serves as the ground for all intellectual activity; and how the immanence of the mind is ordered beyond itself to God, in whom alone it can find fulfillment. By following Augustine in his odyssey through the mind, we have also found crucial principles for navigating the question of “artificial intelligence,” focusing less on the products of intelligence than on its sources, both immanent and transcendent: to the extent that artificial intelligence lacks that abiding self-knowledge from which flows all intelligence, it is not truly intelligent; and because it lacks that abiding self-knowledge, it lacks the kind of rational nature which is capable of knowing God, in whom alone the mind, as an intelligent reality, can find its end.

It would seem, then, that Augustine’s message to the age of AI is one of comfort: however dazzling its products are, AI is not “actually” intelligent, because the essence of intelligence lies in something which AI, precisely as artificial, lacks. But Augustine’s message is not solely one of comfort, since in identifying the knowledge of oneself and love of oneself which constitute the mind, Augustine also discovers the dangerous tendency of the mind to forget its identity, to forget itself and to hate itself.

In an age when certain technological, cultural, and economic forces aim to strip people of their intelligence and then to commodify that intelligence, Augustine has two, seemingly contradictory, things to say: first, that intelligence cannot be stripped from people or externally distributed to them, because intelligence consists most fundamentally in that knowing and loving of oneself which exist as soon as one’s individual human nature exists; but also, second, that this same mind always bears the potential of tragedy, that it can forget itself and hate itself and so bury its identity as the image of God in any combination of moral, cultural, and technological accretion, that it becomes alienated and estranged from what lies most deeply within itself.

Intelligence may not be able to be lost or stolen, but it can be buried. In order to unearth itself from this existential entombment, the mind must heed its own acts and its own nature, discovering within itself the capacity for true knowledge of God and neighbor.