Work in Process

There is a contemporary movement among a number of philosophers, inspired by biology, to pursue a process-oriented metaphysics. These theorists argue that standard approaches are inadequate because they do not capture the essentially dynamic nature of biological systems. Traditional ontologies begin with essentially stable building blocks (a “thing” such as a substance, object, or particle). But in biology, change is fundamental to existence. A biological system persists by virtue of its changes: what needs explanation is its achievement of relative stability amidst flux. So, a metaphysics of biology must use a different framework—one based on processes:

The world—at least insofar as living beings are concerned—is made up not of substantial particles or things, as philosophers have overwhelmingly supposed, but of processes. It is dynamic through and through […] More specifically, we propose that the living world is a hierarchy of processes, stabilized and actively maintained at different timescales (Dupré and Nicholson, 2018, p.3).

It turns out that, even among those who endorse this broad approach, there is controversy about what it means for a process to be the basic element of ontology. Also, while biology provides a motivation, many of these philosophers hope to furnish a metaphysical picture that applies across the board. After looking at some of the issues involved, touching on recent work by Anne Sophie Meincke and John Dupré, I will suggest what I think is a promising approach for developing a process ontology. This is one drawn from the account of causation I favor: a version of what is appropriately labeled “causal process theory”.

Putting Processes at the Foundation

While the debate is complex, the process critique of traditional ontologies usually focuses on how change over time is analyzed.¹ In a traditional object-based metaphysics, change involves an object having some different properties at different times. However, the object itself is said to persist by enduring through time: this is usually defined to mean it is not extended in time (as it is in space), but is wholly present at each moment of its existence. For the process theorist, the notion of change (in, e.g., some properties) in this view is inappropriately subordinated to the essentially static bedrock of the entity which bears the properties.²

A common alternative ontology is a four-dimensional (perdurance) model that views objects as entities extended in time: they are composites that have temporal as well as spatial parts (sometimes one is asked to picture a “spacetime worm,” only a slice of which exists at each moment). Of course, the object’s temporal parts will have varied properties, and this is how change is analyzed. A “process” might also seem to be amenable to the same analysis: a process is also something extended in time that changes through time. Of course, if we treat processes in this way, the distinction between object and processes becomes less sharp. But perhaps some might consider this a virtue.

However, some process theorists pointedly reject this alternative. For example, Meincke made a succinct cogent case in a recent twitter thread, arguing that in the 4D model, the temporal parts that make-up each entity are themselves unchanging, and if a “process” is composed of them, we again have landed upon an ontology that makes change derivative rather than fundamental.³

I find this critique to be persuasive. The challenge is to create a positive account of what it means to say that processes are fundamental—that is, not dependent on any static entities. One of Meincke’s criteria is that a process has “temporal extension that is not divisible (Meincke 2021, 444)”: any effort to divide a process into temporal parts (in analogy to spatial parts) is an idealization without basis in reality.⁴

Causal Process Theory has an Answer

This criterion immediately reminded me of another “process” proposal that did feature such persisting-but-temporally-indivisible basic entities. This was Wesley Salmon’s causal process theory, an approach to causation meant to underpin his account of scientific explanation. This feature of Salmon’s account garnered little support, but I believe a modified version I have advocated can make sense of the notion. And while it only provides a minimal ontology that leaves many questions unanswered, I think the approach can provide what process metaphysicians are looking for.

Before continuing, here I must take note of John Dupré’s recent work.  He has been a prominent voice advancing the case for a biology-inspired process metaphysics.⁵ However, his positive characterization has been criticized by Meincke as not sufficiently distinct from four-dimensionalism: in particular, Dupré thinks biological processes can be said to have temporal parts.⁶ But recently Dupré came at the topic from another direction. He has a 2021 article (“Causally Powerful Processes”) that invokes Salmon and causal process theory, and argues a similar framework can be apt for describing causation in a biological context (in this particular paper he eschews addressing foundational ontological issues). He also notes that causal process theories have previously been mainly discussed with physics rather than biology in mind, and says “I strongly suspect that this [a process account] is the right way also to think of the non-living world. (Dupré 2021, p. 10668).” I agree, and will discuss how to address both non-living and living phenomena while also meeting Meincke’s criterion.

Quick Review of Salmon’s Original Idea

In Salmon’s account, the basic entity is called a causal process, and there are two dimensions of causation: propagation and production. Propagation refers to the progression of a process through time and perhaps space in the absence of interaction, while production refers to the change that causal processes undergo when an interaction occurs.

Salmon insisted that while interactions are events localized in time and space, propagating causal processes are extended as part of their basic nature: “A baseball colliding with a window would count as an event; the baseball, traveling from the bat to the window, would constitute a process (Salmon, 1984, p. 139).” It is a mistake to treat the extended process as itself divisible into events, however. Salmon called this stance the “at-at” theory of causal propagation (pp. 147-157; sometimes labeled “transmission”).  The causal influence possessed by the process (in this version of his theory characterized by an observable “mark”) is simply present at each point between interactions.  There is no further question of how a process “gets from” the interaction at point A to one at point B.⁷

This aspect of Salmon’s theory was criticized in a number of ways: some thought the idea was unclear, especially since many of his (macroscopic) examples involved entities undergoing incessant interactions (picture the baseball encountering air molecules). But even granting the absence of interactions, one might still want to ask about relations or connections between a causal process as it exists at one point in time compared to another. Also, Salmon often specifically talks about propagation between points in four-dimensional spacetime: in this case, a propagating process would trace a timelike interval in spacetime. But in relativity theory, each point in an interval can be identified and labeled as an event. Here again, a skeptic may ask how these events are connected, rather than accepting Salmon’s stipulation that a propagating process cannot be divided up and analyzed in this way.⁸

A Way Forward

Salmon ended up making substantial changes to his theory in response to a variety of objections. The details of this amount to a long story that I won’t get into here. Interestingly, however, he never abandoned this notion of propagation.  I do think the notion can be supported. There are two steps to doing so. First, one adopts a modification to Salmon’s account which enables it to more readily play a realist metaphysical role. Next, one considers how the theory handles quantum particles – the most fundamental “processes” a scientifically informed causal account needs to reckon with.  After discussing these steps, I will show how the analysis can connect back to the biological context that inspires many contemporary process theorists. The key to this is an account of how larger-scale composite causal processes are formed.

“Empowering” Causal Processes

The modification to Salmon I advocate has to do with the characterization of the causal influence propagated by a process. While Salmon struggled to find an empiricist-friendly way to do this, I straightforwardly define this influence in terms of a cluster of dispositions (or “powers”) toward possible interactions—aka a dispositional profile. Interactions produce a mutual change in profiles.⁹

Together, causal processes and their interactions form the world’s causal network or web. The entities and properties described by science correspond to (idealized) features of this web. For example, an electron corresponds to a causal process, and its properties (mass, charge, spin) represent regular features of its dispositional profile. Scientific explanations featuring these theoretical notions succeed in virtue of this correspondence.

The Causal Nature of Quantum Objects

Salmon was aware of a potential deficiency in his theory when it came to quantum phenomena (see Salmon 1984, pp. 242-259). The results of a two-slit experiment, where firing individual quantum particles somehow produces an interference pattern on a screen beyond the slits, cannot be explained if one assumes particles traverse a trajectory along a continuous spacetime path.¹⁰ Generally, the quantum phenomena of superposition and entanglement make assuming classical particle trajectories unworkable. And we can’t assume quantum objects are fields propagating in spacetime either. They do not possess physical quantities with definite values at spacetime points, but can only be ascribed (complex-valued) amplitudes that can be used to make probabilistic predictions. And even these amplitudes cannot be considered to have straightforward locations: once we move beyond the idealized case of a single, isolated particle, modeling a quantum system requires an abstract, high-dimensional configuration space.

Experimental measurements of quantum systems, of course, do take place at specific spacetime locations.¹¹  But between preparation of the system and measurement, it doesn’t make sense to talk about a trajectory in spacetime, or ask what connects events along such a path. So, as in the theory of causal processes, quantum systems appear to be extended (because they bridge two or more locations where measurements are made), but not divisible (because they can’t sensibly be assigned locations between measurements). In keeping with this correspondence, the properties of an electron, for instance, represent its dispositions toward possible interactions with various other systems at different locations. But this cluster of dispositions is not itself located in spacetime and cannot be straightforwardly divided into parts in spacetime. (In an appendix to this post I will consider an objection asserting that the evolving quantum system, even if necessarily modeled in higher-dimensional space, can still thought of as divisible into temporal parts.)

This causal theory then, in a way that comports with physics, delivers on the idea of an ontology based on extended-but-not-divisible processes.  But, on the other hand, it might seem to give us a strange view of reality. What is the world like outside the laboratory where quantum particles are measured?  Here, the answer is to treat all physical interactions in the same way as we do measurements, as in the interpretation called relational quantum mechanics (RQM). In this view, there are no privileged observers. The properties of all physical systems have definite values relative to other systems only when they mutually interact. A system is in superposition, in contrast, only when considered from the perspective of another system in the absence of direct interaction.  This interpretation dovetails with our causal ontology. Quantum systems (when not interacting) are extended processes that bear dispositional profiles. These processes give rise to a punctuated network of events in space-time by virtue of their change-producing mutual interactions. (For more on RQM and how it fits with the causal theory, see this earlier post).

From Physics to Biology?

But this focus on physics may seem to miss the point of recent work by process metaphysicians. They have been inspired by the idea that biological systems in particular must be viewed as essentially processual.  Can our causal process theory satisfy this notion?

One needs to have a theory of how basic causal processes (those that elementary particles map onto), compose larger ones (such as atoms, molecules, and organisms). I have already discussed the details in other posts,¹² but the basic idea is that a persisting composite causal process is formed when two or more sub-processes repeatedly interact, sustaining a pattern. These interacting sub-systems, along with their (changing) dispositions, constitute a dispositional profile for the composite, accounting for how it will itself interact with other systems of like scale. A hierarchy of composite processes is formed in this way, giving rise to productive causal interactions at a variety of scales. While the details are always messier than any quick summary will suggest, a biological system in this framework is a composite process: one whose persistence is due to both the internal causal interaction pattern of its sub-processes and the character of the persisting dispositions toward external interactions that this internal pattern helps to sustain.

In the natural world, interactions and the changes they impart are happening relentlessly, and we inevitably use idealized static models in our efforts to explain atomic, chemical, and biological phenomena. From a metaphysical standpoint, however, there is no ultimate static foundation. There is a hierarchy of processes that take us from one interaction to the next without ever coming to rest.

Conclusion

This is an overview of a fairly minimal ontology originally developed to support an account of scientific explanation. It leaves many of the questions that concern metaphysicians unanswered. Some of these I will try to address in a follow-up.  One possible issue with what I have described above is that that the most elementary processes (associated with quantum systems) still differ somewhat from the biology-inspired conception. For instance, in additional to their distinctive manner of persistence, Meincke says “movement and change are at the heart of what it is to be a process (Meincke forthcoming, 444).” But it is not clear that elementary processes undergo “change” outside of the bare fact that they provide a bridge between interactions at distinct locations (for more, see the discussion in the appendix below). Still, I think that the ability of these powerful processes to build these bridges connotes a kind of underlying vitality that is in the spirit of a process metaphysics.

Appendix: The Nature of Persisting Quantum Systems

While the discussion above, following Salmon, sometimes discussed propagation between spacetime points, here I will look more closely at the question of persistence mainly in the context of non-relativistic quantum mechanics.  A brief comment about quantum field theory follows at the end.

Let’s start with the familiar wave function representation of a simple quantum system.  If we want information about the position of a single particle of a given mass, we solve the for the (position space) wave function using the (time-dependent) Schrödinger equation (assuming for simplicity one dimension of space along the x-axis):

Here, we are solving for ψ, which is a function of position and time. Then we can go on to calculate the probability of finding the particle in a particular location at some time (between points a and b for example) in accordance with the Born rule:

Given that the results depend on the time parameter, it is typically said that a quantum system “evolves” according to the Schrödinger equation. The question under consideration is this: why can’t we say that the quantum system has temporal parts or stages in keeping with this notion of “evolution”?

Before moving on, I note that the most general mathematical representation of the state of a quantum system is as a (normalized) vector in a (complex valued) Hilbert space. The wave function presentation corresponds to a vector expressed in a particular basis chosen based on the observable of interest (position in the example above). Observables are mathematically represented by operators, which perform transformations on the state vectors. Eigenvectors of an observable are those vectors left unchanged by the operator, except for multiplicative numbers called eigenvalues. Eigenvalues of the operator are the only possible results of a measurement of the associated observable.

With this more general setting in place, the question is as follows. According to the Schrödinger picture, the system persists across various moments in time, but its state changes continuously. Why not take the state at each moment in time to represent the stages or parts of the system?¹³ A response might start by pointing to the fact that the dynamics of a quantum system can be represented in another way: the Heisenberg picture. Here, the quantum state vector is not indexed to time; instead, the operators vary with time.  The calculation of the probable position of a particle at a particular time will give the same result as in the Schrödinger picture. To the extent that the state represents the quantum system, then it seems that in this picture the system is extended in time but it is not evolving in time (its state is not changing). Therefore, interpreting the system’s state at each moment in time as temporal parts is, at best, unmotivated.

It may be possible to make a stronger claim, however. Thomas Pashby has an interesting 2013 paper that explores this issue (“Do Quantum Objects Have Temporal Parts?”).  Temporal parts are posited in analogy with spatial parts. So, to answer the question about temporal parts, we might first ask, does it make sense to ascribe spatial parts to a quantum system?  In the paper Pashby describes a principled way to ascribe spatial parts to a quantum object (such as a spinless particle) by associating regions of 3D space with a partitioning of the Hilbert space associated with the position observable (for details, please see the paper). He then shows that when one attempts to find an analogous partition corresponding to the time parameter, the effort fails. The root of the problem is that time is not strictly an observable (representable by a self-adjoint operator), and this in turn is related to the fact that its conjugate, the energy of the system (represented by the Hamiltonian), must have a minimum value to be physically realistic. Pashby concludes quantum systems, while extended in time, do not have temporal parts.

Pashby’s label for this model of persistence (an object is extended in time but lacks temporal parts) is “temporal holism.” Another paper exploring this concept (purely from a philosophical rather than quantum physical perspective) is by Paul R. Daniels, who names it “transdurantism.” (Daniels 2019). Whichever label one chooses, this idea seems to meet the demand of a process ontology discussed in the main body of this post.

Lastly, I’ll try to briefly address whether the situation improves if we consider quantum systems in the context of quantum field theory (QFT). Here the question of persistence gets very murky because of the general difficulties one has interpreting what QFT implies for ontology (see the discussion in the relevant SEP article). On the one hand, how the QFT picture connects to our usual (classical) idea of a field is surprisingly unclear, since there are no physical values associated with spacetime locations. There are instead operators associated with each point. But note these are not like position operators in non-relativistic QM: when a particular quantum state is assigned, there is still no notion of a link to localizable particles or even definite localized field values. Much more could be said I’m sure (and I need to read more of what has been written on QFT ontology), but it certainly doesn’t seem to be any clearer how one might divide a system into appropriate “parts” in QFT than it is in non-relativistic QM.

---

Notes

¹ See for example, Meincke (2019) “The Disappearance of Change: Towards a Process Account of Persistence.”

² For an argument that this sort of criticism of substance-based ontology is sometimes too quick and may be misguided, see this recent paper by Morgan (2021) “Are Organisms Substances or Processes?”

³ The priority of temporal slices is made explicit in a distinct four-dimensionalist view called stage theory, which, in contrast to perdurantism, denies that a composite entity made up by summing temporal parts is part of the world’s ontological furniture and instead treats the instantaneous slices as the only fundamental entities.

⁴ Meincke’s 2021 article discusses Henri Bergson’s emphasis on this notion and related ideas.

⁵ See for instance 2018’s “A Manifesto for a Processual Philosophy of Biology” co-authored with Daniel J. Nicholson. I note that while the authors acknowledge the long history of process thought in philosophy (featuring figures such as Heraclitus and Whitehead), they feel a fresh approach is needed.

⁶ He put it this way in a recent article: “My own view is that all persisting entities have temporal parts, but that these are all in the past and present, not the future (Dupré, 2021, p.10674, footnote 7).”

⁷ In this section Salmon notably quotes Bergson.  He also draws a connection to a proposal for solving Zeno’s flying-arrow paradox of motion. He does not mention Whitehead, despite having once studied him.

⁸ The most common answer, that the events are mereological parts of the larger process, is what the process theorist is motivated to avoid (in line with the discussion of perdurance above). Another answer, that the events are causally connected, is not available: in this theory the extended process is itself playing the theoretical role of a causal connection (between interaction events).

⁹ In my approach, this change is the mutual manifestation of the respective dispositions of the interacting processes. The idea of mutual manifestation (among “reciprocal disposition partners”) is due to philosopher C.B. Martin (see his 2008, Ch. 5). In addition to the terminology of dispositions and powers (which I treat as synonymous), I sometimes use the word propensity when the manifestation is probabilistic. Finally, one may define a “capacity” as a regular feature or aspect of a dispositional profile.

¹⁰ Early on, Nancy Cartwright noted this was a problem for Salmon’s theory; further, she suggested that giving up on the “mark” notion in favor of a transmission of a real causal influence or capacity (akin to what I propose here) might be a way to address the issue. See Cartwright, 1989, Ch. 6.

¹¹ In causal terms, this is a mutual manifestation of dispositions of the system and measuring apparatus.

¹² A post giving an overview of the theory of causal composition is here; and a deeper dive into molecular composition is here. Importantly, this approach to composition avoids problems with the concept of emergence that plague non-causal models.

¹³ Here I am ignoring the option of an endurance interpretation, since that is rejected by process theorists. In this interpretation, the system would be posited to have some unchanging nature that transcends the changes represented by the evolution in the Schrödinger picture.

References

Cartwright, N. (1989). Nature’s Capacities and Their Measurement. Oxford: Oxford University Press.

Daniels, P.R. (2019). Persistence, Temporal Extension, and Transdurantism. Metaphysica, 20(1), 83-102.

Dupré, J. & J. Nicholson, D. J. (2018). A Manifesto for a Processual Philosophy of Biology, in: Nicholson, D. J. & Dupré, J. (eds.), Everything Flows: Towards a Processual Philosophy of Biology, Oxford: Oxford University Press, 3-45.

Dupré, J. (2021). Causally Powerful Processes. Synthese, 199, 10667-10683.

Martin, C. (2008). The Mind in Nature. Oxford: Oxford University Press.

Meincke, A.S. (2019). The Disappearance of Change: Towards a Process Account of Persistence, International Journal of Philosophical Studies, 27(1), 12-30.

Meincke, A.S. (2021). Bergson and Process Philosophy of Biology, in Sinclair, M. & Wolf, Y. (eds), The Bergsonian Mind, London: Routledge, 432-445.

Morgan, W. (2021). Are Organisms Substances or Processes? Australasian Journal of Philosophy, DOI: 10.1080/00048402.2021.1931378.

Pashby, T. (2013). Do Quantum Objects Have Temporal Parts? Philosophy of Science, 80(5), 1137-1147.

Salmon, W. C. (1984). Scientific Explanation and the Causal Structure of the World. Princeton: Princeton University Press.

---