Wednesday, February 25, 2026

The filioque

Suppose we model the Trinity as three vertices with arrows between them representing relations of procession, such that:

  1. No two vertices have two arrows between them and arrows only go between distinct vertices.

  2. The arrows define the vertices in the sense that there is no way to relabel the vertices while preserving which vertex has an arrow to which vertext (i.e., there is no non-trivial directed graph automorphism).

There are seven different structures possible subject to (a).

Condition (b) is a way of capturing part of the traditional Western idea that the Trinity is relational, so that the distinctions between the persons arise from the processional structure. Four of the graphs violate this condition: in (1) we can rearrange the labels in any way without changing the relational structure; in (4) and (5) we can swap A and C; in (7), we can rotate the whole graph by 120 degrees.

Aquinas captures the relationality idea by the principle that person x is distinguished from person y if and only if there is a (real) processional relation between x and y. I think this principle is too strong in one way and too weak in another way. It is too strong, because (2) would allow the persons to be distinguished, even though there is no processional relation between A and C (or B and C): A would be the unique person from whom someone proceeds; B would be the unique person that proceeds; C would be the unique person unrelated by procession. It is too weak, because in (7) we have a processional relation between each person, but it fails to distinguish the three persons. Of course, (2) and (7) have other flaws.

The following options satisfy condition (b): (2), (3) and (6).

Graph (2) is theologically unacceptable, as it has a person not related to any other. And so while the relationality in (2) would be sufficient to define the persons, because C would be unrelated to A and B, the graph does not represent a fully relational Trinity. Besides which, (2) has two persons, A and C, who do not proceed from another, and the East and West agree that only the Father does not proceed.

Graph (3) is interesting. If we adopted this model of the Trinity, we would have to take A to be the Father, since everyone agrees that the Father does not proceed for anyone. Moreover, since everyone agrees that the Son proceeds from the Father, we would hae to

In graph (3), we would have to take A to be the Father, since the Father does not proceed from any other person, as the East and West agree. Likewise, the East and West agree that the Son comes from the Father. So B would have to be the Son. That would make the Holy Spirit be C, and proceed only from the Son. Neither the East nor the West will like this. And it would violate the principle that the Son eternally has nothing beyond what the Father has, excepting what follows from the Son’s being generated by the Father, since now the Son has the Spirit proceeding from him, but the Father does not. I suppose one might try to reconcile (3) with the Western view by saying that the Spirit proceeds mediately from the Father and immediately from the Son. I don’t like this.

That leaves graph (6), the Western view of the Trinity, with A being the Father, C being the Son and B being the Holy Spirit.

This isn’t really a great argument against the Eastern view of the Trinity, because the East will deny (b), on the grounds that one can distinguish the persons not merely by the directions of the procession, but by the types of processions, with generation and spiration being different types. One can think of this Eastern move as allowing the arrows to have different colors, and then instead of (b) having the condition that there is no way to relabel the vertices while preserving which vertex has an arrow of what color to what vertex.

The point here is rather to allow us to state with some precision what principles force the western view. Namely, we have:

Theorem: Suppose:

  1. There are exactly three persons in the Trinity

  2. The (binary) procession relation is asymmetric: if y proceeds from x, x does not proceed from y

  3. At most most one person does not proceed from another

  4. There is one person from whom both of the other persons proceed

  5. There is no way to relabel the persons while preserving the procession relations.

Then there is a way of respectively naming the persons “Father”, “Son” and “Holy Spirit” such that it is correct to say: “The Father does not proceed, the Son proceeds from the Father, and the Holy Spirit proceeds from the Father and the Son.”

4 comments:

  1. What is the difference (assuming there is one) between a procession relation and a grounding relation?

    ReplyDelete
  2. I never thought about this. My quick thought is that when entity A is grounded in entity B, then entity A does not "really exist"--or does not exist as really as B does. For instance, plausibly, a rock's existence is grounded in the arrangement of its molecules--so rocks aren't as real as molecules. This is unacceptable in the Trinitarian context. Or at least we can say this: grounded entities are not fundamental, and so we would have a kind of subordination with grounding in place of procession.

    Besides which, I think we want to say that the Trinitarian relations are sui generis.

    ReplyDelete
  3. I agree we should want things to do with God to be unique. But it seems like there is at least an analogy between "proceeds from" and "grounded in,
    as they're both irreflexive and asymmetric (not clear if they're transitive, maybe it would depend on where you come down on "from the Father, through the Son"-type descriptions of the Holy Spirit)?

    ReplyDelete
  4. At least you'd need to say that procession is a *special case* of grounding. For not all grounding is procession.

    ReplyDelete