Origins of a Subject-Agent

How did the modern equation of subjecthood and agency come to be? How did the notion of “I” or ego come to be substantialized? An extremely influential argument of Heidegger makes this an innovation of Descartes. Alain de Libera argues that this is too hasty, and that the groundwork for this identification was actually laid in the later middle ages. I’m continuing a high-level treatment of de Libera’s extremely important “archaeology of the subject“ (see also On a Philosophical Grammar).

Answering this question will involve an extended historical odyssey through complex interactions between Aristotelian and Augustinian views, and much more. De Libera sees Aquinas in his polemic against Averroes raising four interrelated questions of a more fundamental nature: Who thinks? What is the subject of thought? Who are we? What is man? The second of these seems to have been first asked by Averroes. The other three are largely attributable to Aquinas and his contemporaries, in their reactions to Averroes.

Several points of Aristotelian interpretation (What is substance? What is form? What is act? What is an efficient cause? What is the soul?) will be relevant to answering these, as will Augustine’s meditations on personhood and the nature of the Trinity. De Libera notes that John Locke — a major contributor to modern views on “the subject” — was deeply involved in debates on trinitarian theology. He also discusses Franz Brentano’s modern revival of the medieval notion of intentionality. The medieval version was closely bound up with a notion of “inexistence” or “existing in” of mental objects (forms separated from their matter) in the soul.

In the Categories, Aristotle gives substance the logical sense of something standing under something else. This influenced the Greek grammarians who formulated the notion of a grammatical subject. But in the Metaphysics, he treats this as only a starting point that is quickly superseded by an identification of substance with form or “what it was to have been” a thing, before moving into an account of substance as potentiality and actuality.

De Libera notes a historic division among readers of Aristotle’s treatise On the Soul between those who interpret the soul as an attribute of the body, and those who treat it as a substance in its own right. The latter position has different meanings, depending on whether substance is taken in the “standing under” sense or in the sense of form. De Libera will be particularly interested in the consequences of a further family of positions that make the non-obvious equation of human actions and passions with attributes of the soul.

He notes that “category” in Greek originally meant accusation, and relates this to Locke’s characterization of personhood as a “forensic” notion. We have here to do with subtle relations between attribution, inherence, and imputation with respect to actions and passions in relation to the soul. But what is an action? Must we explain an act in terms of a substantial subject’s power of efficient causation in a late scholastic sense that is far from Aristotle’s? (See also Expansive Agency; Brandomian Forgiveness.)

Mechanical Metaphors

Perhaps the greatest contribution of the Italian physicist, astronomer, and engineer Galileo Galilei (1564-1642) — regarded by many as the single most important originator of modern, mathematically oriented natural science — was a unified explanation of both astronomical and earthly phenomena by the same set of mathematical principles for analysis of the behavior of physical bodies and matter. This was a generalized mechanics of solid bodies.

The tremendous power of this new way of understanding the physical behavior of bodies is undeniable. At least until the computer age, it has been the main basis of modern engineering and technology.

A historical side effect of this immensely successful development has been the promotion of solid-body mechanics as a kind of privileged metaphor for causality in general. I’ve several times discussed the transformation of Aristotle’s notion of efficient cause (most fundamentally, the means to actualization of an end) into the very different notion of “driving” cause or “motor” by medieval and early modern authors (see Efficient Cause, Again; Suárez on Agents and Action; Effective vs “Driving”; Not Power and Action). In combination with a very un-Aristotelian tendency to reduce other causes to efficient causes, this created a ripe condition for the spread of a view of causality in general in terms of metaphors based on solid-body mechanics. We are now so used to this that it takes effort to imagine any other view.

But the solid-body interaction metaphor ultimately leads to an impoverished, overly narrow view of causality in general. (For an alternative, see Aristotelian Causes.) Even within mechanics proper, solid bodies are no longer the paradigmatic, privileged case. At scales that are too small or too large, analogies to the behavior of medium-sized solid bodies break down. In broader contexts, wave phenomena are as important as the analysis of solid bodies. The great Roman poet-physicist Lucretius already had the insight that in the general case, atoms in aggregate behave more like liquids than like solids.

Irreducible to any purely mechanical paradigm, disciplines like earth sciences, ecology, medicine, economics, and computer science provide many examples of more complex and subtle interactions and structures that suggest a new need for something more like an Aristotelian view of causality, as having more to do with forms of things than with force.

Effective vs “Driving”

Modern thinking about causality remains haunted by the figure of “driving” causes on the model of billiard ball impacts, even though the use of sophisticated statistical methods in contemporary science already suggests the obsolescence of that model. I have characterized ordinary Aristotelian actuality as what is effectively operative in a process. This more general “pragmatist” notion of determination pays much more attention, e.g., to the role of the forms of things in shaping processes, but it also subsumes everything that was discussed on the narrower “driving cause” model.

Within the later tradition, the grounds for the early modern reduction of causality to an impact-modeled “driving” sense were laid by scholastic redefinition of Aristotelian efficient causality in terms of activity, and its simultaneous elevation above final and formal causes. I think that for Aristotle, the defining characteristic of an efficient cause is that it is a means by which an end is actualized.

The complete means by which an end is actualized undoubtedly involve activity, but in the example of building a house, for Aristotle it is the art of building that is the pre-eminent efficient cause of the actualized house — the pre-eminent means by which the building of the house is accomplished, even though there are many other contributing means and other causes. But the art of building is not itself an activity. This tells us that activity is not the defining characteristic of an efficient cause. It also brings the notion of efficient cause closer to those of final and formal cause. Taking into account the close relationship of Aristotelian form and matter, it turns out that all four of Aristotle’s causes are integrally related to one another.

I think that for Aristotle, activity is more a practical notion associated with broadly ethical doing, than a basis for theoretical explanation of events and properties. In Aristotle, activity is something inquired about. It does not play the role of an unexplained explainer. (See also Agency; Efficient Cause; Efficient Cause, Again.)

Aquinas and Scotus on Power

Gwenaëlle Aubry’s Genèse du dieu souverain (Genesis of the Sovereign God) concludes with chapters on Aquinas and Scotus. She finds that Aquinas systematically substitutes power and action for Aristotle’s less familiar and more subtle ends-oriented concepts of potentiality and act. Aquinas then distinguishes between active power and receptive or passive power, neither of which has much to do with Aristotelian potentiality.

For Aristotle, Aubry says, potentiality is an indwelling tendency of a being to be attracted toward an end. Pure act is the realization of an end (and, I would add, not itself a movement but an unmoved mover that is an attractor). For Aquinas, the receptive power of beings is the power to receive being from God. Pure act is equated with God’s creation from nothing. Aquinas strongly associates being with power; the power of God, pure Being, pure Existence, is for him an active and efficient cause, not an unmoved attractor. On my reading of Aristotle, it is only the less-than-pure acts of moved movers that are active and efficient causes; the “first” cause is an end that attracts beings.

Duns Scotus, according to Aubry, seems to have originated the modern notion of purely logical possibility. For Scotus, anything at all that is noncontradictory is possible, whereas Aristotle considered possibility more pragmatically, in relation to real-world conditions.

Scotus held that the order of the world is radically contingent, able to be reshaped by God’s will. According to Aubry, he explicitly speaks of God’s arbitrary choice, and attributes a power of arbitrary choice to the human will as well. For Aristotle, the source of contingency in the world is the potentialities of things. For Scotus, it is the absolute power of God.

Whereas Bonaventure, Aquinas, and the 14th century pope John XXII treated the “absolute” power of God as only logically distinct from the “ordained” power associated with the order of the world as we know it, and as not actually separately exercised, Scotus insisted that the absolute power of God is actually exercised. He identified the absolute power of God with a kind of pure fact, and insisted that God from eternity could choose to change the order of the world. (I’m inclined to think Abelard was right, and choice is incompatible with eternity.)

God’s choice for Scotus has no reason beyond itself. Scotus explicitly rejects the passage from Plato quoted by Abelard that everything that is has a cause or reason. Aubry says that for Scotus, the good is only good because God wills it so. This is the exact opposite of the argument of Plato, Abelard, and Leibniz that goodness comes first.

Scotus strongly emphasizes the infinity of God in contrast to the finitude of creatures; infinity for Scotus is God’s most important attribute. Moreover, God’s infinite power acts immediately in the world. This reminds me of the extreme positions on omnipotence articulated by Philo and al-Ghazali. According to Aubry, Scotus also says that a worldly prince enjoys a similar absolute power.

In passing, Aubry notes that Descartes — also a voluntarist — held that God creates eternal truths. This seems to be a somewhat Scotist position. (See also Aubry on Aristotle; Leibniz on Justice vs Power; Power of the One?; Disambiguating “Power”; Not Power and Action; Nature and Justice in Augustine; Peter Abelard; 1277; Being and Essence; Being and Representation.)

Disambiguating “Power”

As Aristotle might remind us, “power” is said in many ways. Each of these is different.

There is the power that Plato suggests as a distinguishing mark of being in the Sophist. There is the greater power he attributes to the Good more ancient than being. There is Aristotelian potentiality, which I normally prefer to distinguish from “power” altogether, but is referred to by the same Greek word. There is the related notion of power as capacity, of the sort developed by Paul Ricoeur. There is efficient causality, itself said in many ways. There is physical force. There is legal or political authority. There are repressive apparatuses. There is the positive, distributed social power involved in the formation of selves, discussed by Michel Foucault. There is the artistic and inventive power with which Nietzsche was especially concerned. There are claims of supernatural power beyond possible human understanding.

I haven’t yet found where in her French text Gwenaëlle Aubry clarifies how her identification of Aristotle’s god with pure act — involving neither Aristotelian potentiality nor Platonic power — goes together with her identification of the efficacy of the pure act with a final causality realized through “potentiality as tendency toward the end”. I think this has to do with the pure act’s role as an end or attractor, so that the potentiality in question belongs to the things it attracts, rather than to Aristotle’s god. Aristotle’s god for Aubry is what might be called an “inspiring” or attracting cause rather than a ruler and a driving cause.

It seems to me that in order to even be intelligible, a power of any kind must be understood as having definite characteristics related to its efficacy. I therefore think “infinite power” is devoid of sense. Even the “omnipotent” God of Leibniz who selects the best of all possible worlds at the moment of creation only selects an inherent, coherently realizable possibility that is also in accordance with non-arbitrary criteria of goodness. He does not create arbitrarily.

Suarez on Agents and Action

Among the greatest of the Latin scholastics, Francisco Suárez (1548-1617) was a profoundly original and highly sophisticated theologian-philosopher who significantly influenced early modern thought, and also produced monumental summaries of several centuries of Latin scholastic argument. A full third of his gigantic Metaphysical Disputations was devoted to an extremely detailed and systematic discussion of causality. A large volume entirely dedicated to efficient causes has been translated to English, and a web search popped up several secondary discussions. My comments here will be very high-level, mostly based on those.

In this scholastic context, traditional Aristotelian terms like cause, being, and substance are all given very different explanations from the nonstandard but hopefully both more historical and more useful ones I have been giving them. Latin scholastics tended to have a somewhat neoplatonizing, substantialized notion of Aristotelian causes. A common view was that any cause must be a substantial entity of some sort, whereas causes in the common modern sense are events, and I read Aristotle himself as identifying causes with “reasons why”.

Suárez held to the view of causes as substantial entities, and apparently went on to argue that all causes give Thomistic being (esse) either to a substance or to an accident in a substance. This influx or “influence” is described as a kind of immaterial flowing of being that makes or produces, without diminishing the agent. In the case of an efficient cause, this influence occurs through action, and the substantial efficient cause is called an agent. (By contrast, in the above-linked article, which has brief additional remarks on Suárez, I quoted Aristotle saying in effect that an agent’s action is more properly an efficient cause than the agent, and that something like a technique used in an action is more properly an efficient cause than the action.)

Suárez’s metaphysical emphasis on actions producing being in things has been characterized as transitional to a modern, event-based view of causality. While Suárez himself held to the idea that causes were substantial agents, early modern mechanism indeed seems to have kept his emphasis on action but moved to an event-based view.

It seems to me to have been a historical accident that mathematical natural science arose on the basis of an event-based view. While mathematics certainly can be used to develop precise descriptions of events, any mathematical analysis relevant to this can also be construed as a “reason why” rather than a mere description. On the frontier of analytic philosophy, Brandom is again suggesting that a consideration of reasons actually circumscribes — and is necessary to underwrite — consideration of events and descriptions. This suggests a new motivation for recovering Aristotle’s original reason-based view.

Efficient Cause, Again

Yesterday, I changed my thinking about Aristotle’s “efficient cause”, making a somewhat surprising connection to the modern notion of “structural causality”. Then I had to update my account of generalized unmoved movers to add a case for an unmoved efficient cause.

Aristotle’s whole framework of “causes” (answers to “why” questions) is often misunderstood, and it is especially bad with the so-called efficient cause. A quick web search on the latter turns up mostly accounts that are just wrong. (A wonderful exception is the outstanding article in the Stanford Encyclopedia of Philosophy).

At Physics 195b22-25, referring back to an earlier example of a man building a house that was initially said (metonymically, it turns out) to be an efficient cause, Aristotle wrote (Complete Works, Barnes edition) “In investigating the cause of each thing it is always necessary to seek what is most precise (as also in other things): thus a man builds because he is a builder, and a builder builds in virtue of his art of building. This last cause then is prior; and so generally.” So it is the art of building — not the carpenter or the hammer or the hammer’s blow — that primarily “builds” the house (i.e., governs the details of the process its construction), and is properly (not just metonymically) called its efficient cause. Once again, for Aristotle it is something at the level of adverbial detail — not the coarse level of agents or action — that is most important.

Having said the other day that Foucaultian discursive regularities are a kind of efficient cause like the art of building, it occurred to me they are also a good example of structural causality, and then that one might say the same about, e.g., the art of building.

Previously, I had been thinking about the efficient cause as functioning like a sort of catalyst. This relatively modest role had led me to privately think of the efficient cause as an “accidental” cause (i.e., one not really contributing to the essence of the thing).

This was at the opposite extreme from the tendency of late scholastics like the great Jesuit Francisco Suárez (1548-1617) to make efficient causes paradigmatic for causes in general, conceived in a proto-modern sense of being responsible for the fact of a thing’s being or existence rather than for the manner of its being, and as involving an “influence” from an agent rather than reasons. (In this context, ends or “final causes” were also reduced to mental intentions of a natural or supernatural agent, as al-Ghazali (d. 1111 CE) had done earlier, quite contrary to Aristotle’s own non-mental account. It was this mentalist reinterpretation that was the real target of Spinoza’s eloquent polemic against against teleology.) There is a nice article by Stephan Schmid on these issues in Suárez.

Anyway, an efficient cause as a point of application of structural causality clearly has a much bigger role to play in determining the detailed nature of a thing than the purely external one of a mere catalyst. (I am using the word “nature” here in a sense broader than Aristotle’s, similar to essence but particular to things that come to be, whereas Aristotle further limited it to nonartificial things, which he thought all contained at least a rudimentary internal principle of motion not shared by artificial things.) On my new account, the efficient cause also exemplifies the interweaving of Aristotelian essence with accident or contingency, due to the role of the semantic materiality as well as form of the means of realization of a nature that is its efficient cause.

This is also more conformant to the idea that all Aristotelian causes are supposed to contribute to explaining the natures of things. Ironically, my previous “catalyst” view made the efficient cause a kind of exception that looked more like the sort of cause of existence I have generally been arguing is un-Aristotelian. As a point of application of structural causality, an efficient cause now fits the general pattern of explaining natures, rather than the mere factual existence of things with natures more or less taken for granted.

Efficient Cause

Each of Aristotle’s four “causes” or kinds of reasons why a thing is the way it is picks out a distinct kind of conceptual content. Actually, none of them — including the efficient cause — should be thought of in terms of anything like a mechanical impulse or force or the exertion of a force. An efficient cause is also not primarily a thing that exerts a force. Rather, an Aristotelian “efficient” cause exercises what in modern terms most closely resembles a sort of structural causality, associated with the form and materiality of the means by which a thing is realized as the sort of thing it is. It acts in an instrumental way that is more “logical” than physical.

In an example of the production of a statue, the efficient cause is not the sculptor, or the sculptor’s will, or the blows of the sculptor’s hammer and chisel. It is the art (objectively characterizable technique) by which the statue is produced. Many people have certainly made contrary assertions about this, but there is, e.g., a good discussion in the Stanford Encyclopedia of Philosophy that supports the above interpretation. (In this simple example, the end is the finished work of the statue. The form and matter are the form and matter of the finished work.)

While the efficient cause is perhaps a little closer to a “cause” in the usual modern sense, it is still far from the same, even though it has a much closer connection to the less common notion of structural causality. Aristotle himself put either form or end first, but influential late scholastics such as Suarez elevated the efficient cause above the other three, perhaps on the ground that God was considered to be pre-eminently an efficient cause (whereas for Aristotle, the “First” cause is primarily an end). I seem to recall some reference to late scholastics treating creation ex nihilo as an example of efficient causality. In any event, Suarez is regarded as treating all four Aristotelian causes on the model of the efficient cause. This helped pave the way for early modern mechanism’s reduction of all causality to a single, univocal form.

Aristotle’s semantically oriented science aims not so much at prediction of what we would call physical events as at a retrospective understanding of why things have turned out the way they have, in a humanly relevant, pragmatic way. Aristotelian “causes” are pluralistic and nonunivocal. They are just reasons why something came out the way it did.