ChapterPDF Available

Causal Efficacy: A Comparison of Rival Views

Authors:

Abstract

The chapter critically assesses whether four rival accounts of efficient causation—transmission, mechanistic, powers-based, vs. powerful particulars accounts—are compatible with the scientific understanding of two often discussed types of causal phenomena: (i) collisions between billiard balls and (ii) how water dissolves salt. It is argued that only the powerful particulars view can be considered compatible with the scientific understanding, mainly because the other three characterize interactions—to varying degrees—in terms of a unidirectional exertion of influence of one thing on another, which is incompatible with the established scientific fact that all interactions are perfectly reciprocal.
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
1
Causal Efficacy: A Comparison of Rival Views
R. D. Ingthorsson
The paper critically assesses whether four rival accounts of efficient causation—
transmission, mechanistic, powers-based, vs. powerful particulars accounts—are
compatible with the scientific understanding of two often discussed types of causal
phenomena: (i) collisions between billiard balls, and (ii) how water dissolves salt. It is
argued that only the powerful particulars view can be considered compatible with the
scientific understanding, mainly because the other three characterize interactions—to
varying degrees—in terms of a unidirectional exertion of influence of one thing on
another, which is incompatible with the established scientific fact that all interactions
are perfectly reciprocal.
Key words: transmission accounts of causation, mechanistic accounts of causation,
powers-based accounts of causation, powerful particulars view of causation, efficient
causation, causal realism
1. Introduction
The aim of this paper is to critically discuss four rival views about how to best make sense of
causation as a real mind-independent feature of the world that involves the production of
changes through the exertion of influence of something on something else. Is causation a
question of (i) transmission of conserved quantities (transmission accounts), (ii) activities of
the parts of mechanisms (mechanistic accounts), (iii) mutual manifestation of powers (powers-
based accounts), or (iv) reciprocal action between powerful particulars (powerful particulars
account)? I argue in favour of (iv). Note that in this paper I will use the term ‘interaction’ to
denote any phenomenon where something exerts an influence on anything else, even though I
will argue that the scientific understanding of interactions, the one I promote, is of reciprocal
action between two entities, notably that whenever any entity A exerts any kind of influence
on any other entity B, B will at the same time exert the same kind of influence on A, and to the
same magnitude.
It bears to mention that transmission and mechanistic accounts are not always counted as
causal realist accounts—i.e., as treating causation as a mind-independently real phenomenon—
but only as attempts to clarify causal reasoning. This is partly because some proponents
explicitly take that stance, and partly because these accounts originate in the empiricist tradition
of accounting for everything in terms of what can be observed and are therefore unreflectively
grouped with neo-Humean accounts. Indeed, their proponents share an aversion to the
postulation of anything popularly regarded as ‘unobservable’ such as powers. However, I
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
2
believe the thinkers mentioned in this paper do treat causation as a mind-independent and real
phenomenon and therefore can count as causal realists.
I will first briefly sketch the core ideas of the different accounts before ventilating my
concerns about each view, first in general terms and then with respect to the way they explain
two kinds of causal phenomena: (i) collision between billiard balls, and (ii) how water dissolves
salt. It will be argued that transmission, mechanistic, and powers-based accounts are, to varying
degree, incompatible with certain scientifically established facts about the behaviour of
physical entities, in a way that the powerful particulars view is not.
2. Transmission accounts
I will initially use Wesley Salmon’s account of causation in terms of causal processes,
propagation, and production (Salmon 1980, 1984) to illustrate the core idea of transmission
accounts, since most contemporary accounts are influenced by it, such as Philip Kitcher (1989),
Phil Dowe (1992), and Max Kistler (1998). Later, I will refer to Dowe (2009) for a more recent
formal statement of the core tenets of transmission accounts.
It bears to mention that Salmon’s account is sometimes described as “neo-mechanistic”
because it was originally a revival of the idea that explanation is about identifying the causal
mechanisms that produce natural phenomena (Galavotti 2022). However, he ultimately moved
towards a transmission account of causal mechanisms.
According to Salmon, the persistent entities that form the basis of most scientific
ontologies—particles, molecules, cells, animals, planets, etc.—are to be understood as causal
processes; they are causal because able to exert influence on each other, and they are processes
because science reveals them to be continuously changing even though they may appear not
to. Ultimately, everything above the level of elementary particles, is either a continuously
changing atom or made of atoms of that kind.
Salmon denies that talk of ‘process’ is a commitment to event ontology, i.e., that entities
are structured series of events. However, he doesn’t offer an alternative analysis of process; he
just points to examples of things that remain the same through continuous intrinsic change
(1984: 139ff). His rejection of event ontology is not a commitment to substance ontology
either, but a rejection of any ontology that tries go beyond the observable. Of the transmission
accounts that developed in Salmon’s wake, some work with events as the relata of transmission
(Kistler 1998) while others stick to Salmon-style causal processes (Dowe 2009).
Propagation refers to the way causal processes conserve their structure over time when not
interacting with other processes and consequently conserve their ability to exert causal
influence in any future interaction. Production refers to the way interactions between causal
processes result in modifications in their structure and in their ability to exert causal influence.
Salmon resists commitment to causal powers for the same reason he resists commitment to
event ontology; they are unobservable and therefore mysterious. Instead, he first uses the term
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
3
‘marks’ to refer to the observable properties of causal processes recognised by the sciences,
and which play the same explanatory role that powers are meant to do (Salmon 1980). Later,
due to the criticism from Kitcher (1989) and Dowe (1992), Salmon adopted the view that
changes are produced in interacting causal processes because one transmits a conserved
quantity to the other.
Salmon’s view can be illustrated concretely by considering a batter hitting a baseball with
a bat in such a way that it leaves the pitch and breaks a window in a neighbouring house. The
bat and baseball are causal processes propagating the conserved quantity of momentum. On
contact, a conserved quantity of momentum is transmitted from bat to ball. The ball then
preserves its momentum as it travels across the pitch, propagating it from point A to point B.
When the ball interacts with a window it is in virtue of the preserved momentum that it breaks
the window.
Let us now look at a more formal statements of transmission accounts; one event-based and
another based on causal processes. Kistler presents the event-based account in what he calls a
‘reduction statement’:
(T) Two events a and b are causally related in the sense that one is a cause of the other if and
only if there exists a conserved quantity Q of which a particular amount P is transmitted between
a and b (Kistler 1998: 1)
Unfortunately, the statement is not specific enough about the relevant sense of ‘event’ to
distinguish causal and non-causal instances of transmission, and therefore doesn’t distinguish
between what Salmon calls propagation and production. If events are understood Kim-style,
i.e., as a particular a instantiating a property F at a time t, then conservation laws entail that
any quantity present in any Kim-style event e1 will be transmitted to the next event e2 whether
or not anything causal occurs between e1 and e2. For instance, a uniformly moving body will
through any given temporal interval constitute a succession of distinct Kim-style events
between which any conserved quantity possessed by any event ei will be transmitted to a
subsequent event ej (see definition of ‘world-line’ below). To be fair, Kistler obviously assumes
that everyone understands that (T) is only meant to apply to what happens in causal interactions
between distinct entities. However, this needs to be spelled out explicitly to avoid objections
by hair-splitting readers such as myself. More importantly, to make (T) more precise it seems
necessary to introduce some term for the persistent particulars that are the constituents of
events. There are other worries, but they apply just as much to the causal process version
discussed below.
The causal process version specifies that while a causal process is continuously transmitting
conserved quantities (really, propagating them), it is only when something is transmitted
between distinct processes to produce change that we can talk about causal interactions. This
position can be summed up in terms of three claims (for reference, see Dowe 2009: 219):
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
4
Causal process: an object that possesses a conserved quantity.
World‐line: a spacetime trajectory of a causal process in which it propagates a nonzero
amount of a conserved quantity between every spacetime point of that trajectory
Causal interaction/production: an intersection of world lines A and B (A B) that
involves exchange of a conserved quantity producing a modification in A and B.
Dowe’s causal process version of the transmission account, resolves the worries I raised about
the event-based version. However, there are other problems, most of which are also present in
mechanistic and powers-based accounts. It is important for what comes later, that while an
‘exchange’ implies a certain kind of reciprocity between A and B—both change in proportional
albeit different ways—then the reciprocity is always assumed to involve the loss of a quantity
by A, which B gains; exchange goes one way from A to B. It is this unidirectionality of the
exchange that is meant to constitute the direction of causation.
3. Mechanistic Accounts
The core idea of mechanistic accounts is that causation is the activities of compound parts of
organised wholes that produce changes in either whole and/or parts. It is in fact a requirement
that causes and effects must be connected by mechanisms (Glennan 2017: 145). There are some
disagreements to be found between accounts, say, between the ‘mechanism first’ approach
favoured by Stuart Glennan in earlier works (2009) and the ‘activities first’ approach advocated
by Machamer (2004) and Bogen (2008). However, I don’t think it is of any consequence for
this paper to confine the discussion to the most recent systematic presentation of the
mechanistic philosophy found in Glennan’s book The New Mechanistic Philosophy (2017).
Indeed, the influences from Machamer and Bogen, as well as from William Bechtel and Adele
Abrahamsen (2005), Carl Craver (2007), and Lindley Darden (2008), just to name a few are
clear to see. I will refer to this view as ‘NMP’.
NMP differs from the empiricist and reductive accounts of causation it is sometimes
associated with, in rejecting the idea that causes and effects can be understood as Kim-style
events. NMP stresses the processual nature of causation as a natural consequence of taking
causal production to consist in the exertion of influence between particulars, and that this
exertion must be understood as an activity that cannot fit in an instant (Glennan 2017: 177). To
be more precise, the causally relevant sense of ‘event’ is of particulars (plural) doing something
to each other since a cause is never the activity of just one particular but an interaction between
parts of a mechanism. This removes the need to further distinguish between causal and non-
causal events.
In more detail, Glennan suggests we should understand causation in terms of constitutive,
precipitating, and chained production. Constitutive production is what happens as a result of
interactions between parts of a mechanism, and which can only be understood in terms of the
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
5
activities of the whole mechanism. Glennan uses the example of searing a steak to illustrate.
When a steak is put on a hot skillet, energy is transferred to the steak (which is precipitating
production) until the temperature of the steak rises above a certain limit (+140°C). At that point
the constituents of the steak—various sugars and amino acids—react with each other in what
is known as the Maillard reaction to produce a variety of molecules responsible for a range of
pleasant flavours. So, the activities of the parts of the steak produce in those very parts a change
to turn raw meat to seared delicacy.
Precipitating production is “the way in which one or more events produce another event
[…] by creating start-up conditions for a different mechanism” (2017: 182). The skillet
transferring heat to the steak is an example of an event that creates the start-up conditions for
the Maillard reaction in the steak.
Finally, chained production is a connection between events separated by a chain of
intermediary precipitating events which in turn involve chains of constitutive production. We
might say the connection between someone buying a steak and eating it, boils down to a series
of precipitating events from grabbing the steak at the store, putting it in the carrier bag, carrying
it home, putting it onto the skillet, searing it, placing it on the plate, and finally eating it.
There are certain points I puzzle over, regarding the distinctions between constitutive and
precipitating production and their relationship to each other. Chained production, on the other
hand, is straightforwardly understood as a connection between two non-adjacent events
consisting of a chain of precipitating production, but precipitating production is not obviously
constituted by constitutive production. Constitutive and precipitating phenomena, we are told,
marks the difference between “mechanisms that produce phenomena (non-constitutive) and
mechanisms that underlie phenomena (constitutive)” (Glennan 2017: 109). This could mean
that the heat transmitted from skillet to steak produces the phenomenon of Maillard reaction,
while the interactions between the constituents of the steak underlie the Maillard reaction. In
that case, I would have guessed precipitating production would correspond to what Salmon
calls ‘production’, i.e., an interaction between two distinct causal processes which produces a
modification in them (skillet cools down, steak heats up), while constitutive production would
correspond to the interactions that constitute the changes internally to each process when two
processes interact; something Salmon does not have a separate name for other than simply
‘modification’. However, Glennan says that constitutive production corresponds to Salmon’s
production and offers such examples as hammering a nail and a horse pulling a cart (Glennan
2017: 184). To my mind, hammering a nail is in the same category of phenomena as skillet
searing a steak. In both cases we arguably have two separate and well-defined wholes
interacting with each other—hammer and nail+board (skillet and steak)—which produces a
change in each other but neither whole appear to underlie the changes in the other, not in the
same way the constituents of the steak underlie the Maillard reaction. Similarly, the hammer is
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
6
not an underlying component of the phenomenon of nail being driven into the board, but it is
something that contributes to that change.
What I suspect we are seeing here, is the difficulty of applying a theoretical model to a
many-layered and thoroughly complex reality. There is a problem in demarcating clearly
between mechanisms on roughly the same level, and the many layers of mechanism within any
one of them. Is hammer and nail and board separate mechanisms or are they all united by the
person wielding the hammer and steadying the nail and board? And, underlying the amino-
acids and fats that underlie the Maillard reaction are the atoms of the amino acids and fats
whose interactions also must be constitutive of the wholes they are parts.
NMP and transmission accounts are similar in that NMP admits (or can admit) that
transmission of conserved quantities is one of the ways we can understand what goes on in
causal interactions, as for instance when energy flows from skillet to steak. NMP could even
admit that in the domain of particle interactions, transmission of conserved quantities may
possibly be all we need to understand what is going on. However, NMP claims to offer a more
general model that can be applied to other scientific disciplines that do not operate with
conserved quantities. Another difference between transmission and NMP accounts is that the
former treats a mechanism as a nexus between distinct entities, whereas the latter treats a
mechanism as a complex system. I take these differences to be large enough to treat
transmission and mechanistic accounts as distinct theories despite the similarities.
Finally, like transmission accounts, NMP tend to characterise the influence exerted
between distinct objects in precipitating production as unidirectional (from skillet to steak).
However, the assumption of unidirectionality is much less pronounced in NMP, especially
when it comes to constitutive production where it is not at all clear that interactions are assumed
to be unidirectional.
4. Powers-based Accounts
Powers-based accounts (for instance, Martin 1997; Ellis 2001; Molnar 2003; Mumford and
Anjum 2011; Heil 2012; Marmorodo 2017) are more heterogenous compared to transmission
and mechanistic accounts, mostly because of a disagreement about the nature of powers.
Dispositional essentialists accept the categorical/dispositional distinction—a distinction firmly
rooted in the empiricist tradition—and thus say there is a difference between the properties that
determine the qualitative state of the object at any given time, i.e., its qualities, and the
properties that determine what the object is liable to do if certain conditions arise, i.e., its
powers/dispositions (for instance, see Ellis 2001). Proponents of the identity theory reject the
distinction and insist that the very same properties that determine the objects qualitative state
also determine its ability to affect and be affected; they are powerful qualities (for instance, see
Ingthorsson 2013).
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
7
Dispositional essentialists (for instance, Ellis 2001; Molnar 2003) tend to think of causation
as a matter of pure potentialities (or unmanifested powers) being triggered by stimuli to
transition from potentiality to actuality, and thereby manifesting some qualitative state; the
manifestation. Identity theorists (Martin 1997; Heil 2012; Ingthorsson 2013) tend to think of
causation as a matter of two powerful qualities mutually modifying their bearers from one
powerful qualitative state to another powerful qualitative state; something called a mutual
manifestation. There is a greater heterogeneity within each view, but for the purposes of this
discussion I see no reason to delve into that. The two conceptions presented here, of how
something is manifested, are so close that dispositional essentialists have found it easy to
identify their manifestations with mutual manifestations, under the assumption that their
manifestations are also a joint product of power and stimuli.
Now, most powers-based accounts, regardless of the differences mentioned, accept
something equivalent to the distinction between active and passive powers that we find in
Aristotle, the Stoics, the Scholastics, and the natural philosophy of the Early Enlightenment,
i.e., between the ability to exert an influence on other objects/powers (active) vs. ability to
change in response to an external influence (passive). Indeed, they will represent the direction
of causation as the direction of the influence flowing from the object with the active power to
the object with the passive power. A dispositional essentialist will say that a ball at rest can
potentially move (it has the power to move while it is not actually moving), and that this
potentiality will become an actuality when influenced by a ball in motion colliding with the
ball at rest. The resulting motion will be a mutual manifestation of the active and passive
potencies. Most identity theorist will say that what we perceive as a ball at rest is really a ball
with an actual and determinate momentum, p, which is understood as a state of motion. If such
a ball, a, with momentum pa, collides with another ball, b, with momentum pb, the two balls
will mutually modify each other in accordance with the laws of motion, resulting in a transition
of a from pa to pa*, and of b from pb to pb*. However, despite differences in the manner which
dispositional essentialists and identity theorists characterise the collision between billiard balls,
they will agree that one of the balls exerts an influence on the other, while the latter receives
the influence. Like transmission accounts, most powers-based accounts characterise causal
influence as unidirectional.
It is now time to introduce the powerful particulars view, but it is best to first point out what
I think is problematic in the views already presented, because the powerful particulars view is
best understood as an attempt to overcome those problems.
5. Influence: Unidirectional or Reciprocal?
Transmission, mechanistic, and powers-based accounts agree that causation is the exertion of
influence of something on something else but disagree on the nature of the ‘something’ that
exerts influence and about the nature of the influence being exerted. Most importantly, they all
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
8
tend to treat influence as unidirectional. This is an overgeneralisation, as already hinted at
when presenting NMP. Indeed, it will emerge that I am an identity theorist rejecting the
unidirectional influence. But it is true enough for purposes of presentation. I’ll qualify it later
with respect to variations within each family of views.
In taking influence to be unidirectional the accounts under scrutiny are implicitly endorsing
an understanding of influence that has been a part of the causal realist tradition since Aristotle
wrote that “whenever the potential active and the potentially affected items are associated in
conditions propitious to the potentiality, the former must of necessity act and the latter must of
necessity be affected” (1998: 9, 5; 264). Indeed, we see a similar agreement about the
unidirectionality of influence in the Aristotelian, atomistic, Stoic, Scholastic, and natural
philosophy of the early modern period as the one we see today. They all share certain core
notions, that together make up what I have elsewhere called the standard view (for further
details, see Ingthorsson 2021: 45ff). The idea is that a new state is produced when an already
existing entity, or complex of entities, changes due to an influence external to that entity, one
without which the change would not have occurred, and the new state never exist. The external
influence comes from an entity possessing powers that allow it to influence other entities
(active), and that the entity upon which it acts possesses the power to receive the influence and
change in some particular way (passive).
It is true that if we limit the discussion to the kind of examples that philosophers typically
consider, the unidirectionality of influence and the distinction between active and passive
entities appear to make good sense. Philosophers talk about billiard balls in motion acting on
a ball at rest (Hume 1748: § 36); lead ball dropped onto a pillow produces a hollow (Kant 1787:
A203); a locomotive pulls a truck (Taylor 1973: 35); a baseball is hit by a bat to fly across the
pitch (Salmon 1980: 50). However, as Mario Bunge first pointed out (1959: ch. 6), it should be
recognised as a serious problem that modern science categorically rejects the reality of
unidirectional action even in the apparently asymmetric cases that philosophers typically
consider. It is instead insisted that all influence comes in the form of reciprocal action, or
‘interaction’ as the term is defined in classical physics.
To be sure, classical physics is in many ways an outmoded framework, but we need to be
careful in dismissing every component. While the first and second laws of motion are known
to fail in extreme situations, the core idea expressed in the third law is still believed to hold
good even in quantum and relativistic physics (albeit adapted to fit more generally to
conservation laws), and it is the third law that is the basis for the rejection of unidirectional
action. The core idea of the third law can be generalised in the following way:
Whenever any object whatsoever exerts any influence whatsoever on any other object
whatsoever, the latter exerts at the same time an influence of the same magnitude on
the former, but in opposite direction.
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
9
While classical mechanics offer a fairly accessible treatment of the notion of reciprocal action,
a full understanding of the concept still requires something of a shift of perspective. The notion
is most clearly expressed in the third law of motion which says that the force by which object
1 acts on object 2 is equal to the oppositely directed force by which object 2 acts on object 1
(F1on2 = –F2on1). However, because the prevailing understanding that influence is unidirectional
has been such an integral part of the popular and philosophical understanding of causation for
so long, the third law has always been widely misunderstood. Indeed, Newton’s unfortunate
decision to use the words action and reaction when explaining the law in plain English
contributed to that misunderstanding. Let me explain.
The correct understanding of reciprocal action is not that of a phenomenon composed of
two different kinds of actions, of which one (the action) gives rise to or provokes the other (the
reaction)—as two tennis-players returning each other’s strokes—but of a single phenomenon
of mutual influence occurring simultaneously between two objects such as when bat collides
with baseball. In Bunge’s words “physical action and reaction are, then, two aspects of a single
phenomenon of reciprocal action” (1959:153). Or, as physicists Resnick, Halliday, and Krane
put it: “Any single force is only one aspect of a mutual interaction between two bodies” (2002:
83). In reciprocal actions, neither side has priority, and the terms ‘action’ and ‘reaction’ can be
arbitrarily assigned to either.
However, researchers are not always interested in both sides of an interaction equally, and
therefore often focus on the effect that matters to them, neglecting other outcomes. The point
is stated beautifully by James Clerk Maxwell:
The mutual action between two portions of matter receives different names according to the
aspect under which it is studied, and this aspect depends on the extent of the material system
which forms the subject of our attention. If we take into account the whole phenomenon of the
action between the two portions of matter, we call it Stress […] But if […] we confine our
attention to one of the portions of matter, we see, as it were, only one side of the transaction
namely, that which affects the portion of matter under our considerationand we call this aspect
of the phenomenon, with reference to its effect, an External Force acting on that portion of matter.
The other aspect of the stress is called the Reaction on the other portion of matter (Maxwell 1877:
267)
The lesson to be learnt is that even if it is recognised that interactions are perfectly reciprocal,
they are often treated, for the sake of convenience, as if they were instances of unidirectional
flow of influence from one portion of matter to the other.
However, even if it is a part of the amassed scientific knowledge that interactions are
reciprocal, misunderstandings are pretty common even among professional physicists. At least
if we are to believe physics educators Steinberg, Brown, and Clement (1990) as well as
Hellingman (1992). The main misconception they identify is the following:
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
10
The ‘inappropriate conception’ (as Steinberg et al call it) here at stake is the same as in Newton’s
time: action and reaction are conceived as separate agents instead of as two sides of one
interaction. Since they are conceived this way and since there are always two interacting bodies
in producing a force, the suggestion of an action belonging to one body and a reaction belonging
to another is virtually inescapable (Hellingman 1992: 112)
Anyone teaching classical mechanics faces the challenge of how to prevent students from
falling under the spell of the ‘virtually inescapable suggestion’.
But what are the philosophically interesting consequences that follow from the established
fact that interactions really are reciprocal and not unidirectional? Like Bunge, I think we must
accept that the polarization of interacting entities into an Agent that exerts an influence and a
Patient that suffers a change upon receiving an influence “is ontologically inadequate” (Bunge
1959: 170–1). The reciprocity of interactions shows that there are no strictly passive entities—
those who only receive influence but do not themselves influence other things—nor are there
entities who influence other things without being themselves equally affected. Furthermore,
since the mutual exertion of influence occurs simultaneously in equal magnitude in opposite
direction between two objects, there is no way to give priority to one or the other as chiefly
responsible for the outcome or even for initiating the interaction. It is because of this that the
terms action and reaction are considered arbitrary and why “we are free to consider either of
them as the force or the counterforce” (Hertz 1956: 185). If one comes first, we would not be
free to do this. Ultimately, the conclusion is that in so far as any causal account assumes
unidirectionality of influence, it is based on an empirical falsity in light of the current
understanding of physical interactions.
There are two worries I expect many readers will find difficult to let go of. One arises from
the fact that physical interactions are indeed so often described and understood in terms of
transmission of quantities even in physics. Is physics both right and wrong, but in different
ways, about how we should understand interactions? The other arises from the widespread use
of the terms ‘mutual’ and ‘reciprocal’ in the literature about powers-based accounts. How
exactly is reciprocal action as defined by physics different from mutual manifestations of
disposition-partners as defined by powers-based accounts?
To answer the first question, then yes, physicists often apply the strategy Maxwell
describes, notably to treat interactions as if only one side of the transaction mattered, such as
when describing interactions in terms of transmission of conserved quantities. The inadequacy
of this kind of characterization comes out clearest in the fact that while they often function fine
for the intents and purposes of a given occasion they fail to generalize: they don’t work for
symmetrical interactions. When we turn our attention to the whole phenomenon, the
transmission account only half-explains. As Maxwell points out, this is a choice of convenience
we can safely make when the other side of the ‘transaction’ doesn’t matter to us.
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
11
Consider a cue ball striking the black eight at a slight angle, pushing the black eight into a
corner pocket while the cue ball continues in a new direction to end in position for the next
shot. It seems easy to explain this by the transition of a quantity from the cue ball to the other,
but then we are neglecting the details about what makes the cue ball change its trajectory in the
way it did. Can we explain its change of trajectory merely by its loss of a quantity without
bringing in any notion of an influence being exerted by the black eight on the cue ball? I at
least find it difficult to entirely replace talk of forces with talk of transmitted quantities. You
would have to make it so that the loss of the momentum p by the cue ball not only takes the
role of the force usually meant to act on the black eight but also of the force presumed to act
on the cue ball to change its trajectory. This is difficult to do even in asymmetrical interactions,
but even harder in symmetrical interactions, say, when two cue balls move in opposite
directions with the same velocity and collide head on. The outcome cannot be explained only
in terms of something being lost by A which is gained by B. At least it must involve both A
and B losing and gaining something.
An interesting example to consider is the common practice of explaining particle
interactions by the transference of a virtual particle from one to the other, represented by
vertices in Feynman diagrams. This is a way of conceptualizing what happens as if something
carries a quantity from A to B. But when we consider that the term ‘virtual particle’ is not
generally believed to denote actual particles in motion (Jaeger 2021), then the imagery of a
transmission of something from one thing to the other appears to be misleading. Very
simplified, a virtual particle is a technical notion used to signify the presence of certain
theoretically calculated quantities considered to mediate interactions between real particles,
and which obtain within time limits that are too narrow for anything to be observed. The idea
is that the calculated quantities are of the kind that particles may have but it just isn’t clear
there are any, wherefore there is talk of a ‘virtual’ particle. The quantities kind of add up to
something that might belong to a particle but may better be said to belong to a quantum field.
However, if we were to assume there is a particle that actually moves between the two colliding
particles, then in light of the reciprocity of interactions stipulated by the third law, then we
would have to ask in what direction the particle is moving—from particle A to B, or vice
versa—and whether the assumption of it moving one way, or the other, would really explain
the whole interaction or only one side of the ‘transaction’. Again, it is easier to think of
asymmetrical interactions in terms of transmission of a virtual particle from A to B, than it is
to think of symmetrical interactions in that way.
Turning now to the question of whether powers-based accounts already operate with an
understanding of reciprocity when using terms like mutual manifestation of reciprocal
disposition-partners. I think clearly not. According to physics, interactions are reciprocal in the
sense that any two interacting things simultaneously influence each other and to the same
magnitude. Powers-based accounts describe powers as reciprocal even when it is assumed that
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
12
one thing influences and the other merely receives the influence. For them, to say that powers
are ‘reciprocal’ is only an admission of the fact that two powers are always involved, and that
both are considered equally important for the production of the outcome, but the outcome is
typically only the change produced in the passive recipient. An example would be that the
power of fire to heat and the ability of a hand to be heated both contribute to the end result of
a hot hand. This is the traditional way of distinguishing between an active and passive powers
as well as between Agents and Patients. The point is that if we accept what physics takes to be
an established fact—that all interacting objects influence each other—it is no longer possible
to argue that the distinctions between active vs. passive powers and Agent vs. Patient are based
on the unidirectionality of causal influence. This realisation is the starting point of the powerful
particulars view.
6. Powerful Particulars Account
I have now presented the core ideas of what must surely count as the main causal realist
accounts in recent times, the most serious objection to them all, and some worries specific to
each view. It is time to add my preferred account of causation in terms of reciprocal action
between powerful particulars into the mix. It first appeared in the paper ‘Causal Production as
Interaction’ (Ingthorsson 2002) but again in more developed form in a monograph (Ingthorsson
2021). The basic idea is that a single conception of causal interactions can account—in one and
the same way—for (a) changes produced by interactions between previously unconnected
entities, whether simple or compound, (b) composition of compound unities, (c) persistence of
such wholes over time, and (d) that changes internal to a whole are just as causal as the changes
resulting from interactions between wholes. It does so in a manner that easily relates to the
theories and findings of the empirical sciences, in two ways. First, it is transparent how the
philosophical account and the theories and data of the sciences offer complementary accounts
on different levels; philosophy providing a general model applicable to all the specific
phenomena, different scientific disciplines providing the details about each particular
phenomenon. Second, it is clear how future developments in the sciences could falsify the
general philosophical account. For instance, if physics ever finds truly asymmetrical
interactions that violate the third law, my view is falsified.
Accounts of causation in terms of interactions between powerful particulars is no novelty.
They are found in Aristotle, the Stoics, the atomists, the Scholastics, and the natural
philosophers of the early enlightenment (for further discussion, see Ingthorsson 2021: ch. 3.7).
In the 20th Century we find Dorothy Emmet (1985) and Ingvar Johansson (1989) defending
this view. However, with the exception of Johansson’s ‘action by mixture’ view, these earlier
versions either assumed the unidirectionality of influence or construed the reciprocity of
interactions more in the form of mutual manifestations and therefore are subject to the same
kind of criticism levelled at powers-based accounts. More recently—indeed, while making
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
13
final revisions of this paper—Andrew Newman has independently offered an account of the
constitution of objects in terms of interactions between their constituent parts, one that appeals
to our scientific understanding of how particles form atoms (Newman 2022). Newman doesn’t
address water dissolving salt, but his understanding of the collision of billiard balls (2022: 307–
8) agrees exactly with the analysis I offer below, as well as with my earlier analysis of a brick
hitting a window (Ingthorsson 2002: §3).
My powerful particulars view aims to offer a unified explanation of what Glennan calls
precipitative and constitutive production (2017: ch. 7), what Craver and Darden think is needed
to ‘maintain’ certain structures (2013: 65–6), what explains the production and destruction of
the kind of entities that Salmon calls ‘causal processes’ as well as the unified wholes that NMP
would denote as ‘mechanisms’. Bear in mind that I only describe my view in enough detail to
make comparisons between views meaningful (for full detail, see Ingthorsson 2021).
The powerful particulars view bears very little resemblance to transmission accounts,
although it is meant to fit well to the scientific conception of the world as ultimately constituted
by particles carrying properties of the kind described as conserved quantities. It resembles
powers-based accounts, some more than others, to the degree that I think the best way to
explain the behaviour of the entities that are parts of organized wholes (both why they
constitute wholes and how they and the wholes behave in interactions) is to attribute powers to
them. However, I prefer the conception of powerful qualities and I like powerful qualities to
correspond to the natural properties that the sciences think we have good reason to think are
real; I don’t think powers come in addition to the natural properties, as something an object has
‘in virtue of’ its other qualities (Ingthorsson 2013). This still leaves place for plenty of emergent
properties, i.e., properties of wholes that cannot be reduced to the sum of the properties of the
parts.
I also prefer to think that it is the particulars that bear the properties that exert influence on
each other, and which change, because I cannot make much sense of the idea that powers act
on other powers to change the powers. Do we say a force makes a quality of velocity speed up
from 10 to 16 km/hr? No, we say a force exerted between bodies makes a body accelerate from
10 to 16 km/hr. My reasons for preferring a ‘particulars first’ ontology are closely connected
to my views on time, change, and identity, but this is not the place to elaborate (for details, see
Ingthorsson 2002, 2016).
I am not convinced about the legitimacy of the distinction between active and passive
powers or between agents and patients, at least not as they are drawn today. That is, as power
to affect (active) vs. power to be affected (passive), whose possession determines whether an
object is active (an agent) or passive (a patient). I might be persuaded to think there is a real
distinction between active and passive powers of a single object, but not that they determine
that some objects are active while others are passive. The only distinction between agents and
patients that I am prepared to admit at present is the one between intentional agents and
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
14
inanimate objects. If there is anything that can truly initiate an action, it would be an intentional
agent.
I think my view resembles NMP the most, and if its proponents could be persuaded to think
that the concept of powers need not be of unobservable mysterious potencies but simply of
powerful qualities such as spin, charge, momentum, valency, etcetera—the kind of properties
that are ubiquitous in the ontology of all the natural sciences—then I think they could warm up
to a powerful particulars view. In the end, the resemblance is mostly that we seem to identify
the same kind of phenomena as belonging to the class of relevant causal phenomena of which
we need to develop a unified view, and this I believe is a consequence of the shared idea that
the theories and findings of the empirical sciences is an important input, and that an agreement
between the philosophical and scientific images of the world is desirable.
It is relevant to note that the powerful particulars arose from the realisation that for more
than 333 years the philosophical implications of one of the most significant results of classical
physics has either not registered as relevant in the minds of philosophers of causation or been
misunderstood and dismissed. Mario Bunge (1959) is a notable exception. I am talking about
the result that there are no unidirectional actions, only reciprocal interactions. To me it
appeared foolish to challenge the validity of this result, because it would require me to show
that our current physics is fundamentally wrong. It seemed more fruitful to explore the
consequences of accepting the result as true (as already noted, I treat it as a provisional and
falsifiable truth). Bunge had already argued convincingly that the reciprocity of interactions
demonstrates that the prevailing idea must be wrong that causes are active objects or events
that exert influence on some passive recipient, and that effects are the changes suffered by the
recipient. This idea can only be considered an approximation (Bunge (1959: 151ff) calls it the
‘causal approximation’) which in many or most cases might be good enough given our
particular explanatory interests, pretty much in the way Maxwell describes. However, Bunge
comes to the conclusion that an account of causation based on reciprocal action is equal to a
reduction of the asymmetric and productive relation of causation, to a non-productive and
symmetric functional relationship of the kind Russell favours (1912). I disagree with this
conclusion. Yes, interactions are symmetrical in the sense that two interacting objects influence
each other simultaneously and to the same magnitude, but that influence produces a succession
of states between which there is an asymmetric relation of one-sided existential dependence:
one of producer to product.
Let me repeat that the kind of reciprocity we can see traces of in NMP and powers-based
accounts is not drawn from the realisation that all interactions are reciprocal. It rather boils
down to a realisation that even the scientific explanations that only pay attention to one side of
an interaction—in a so-called ‘causal approximation’—cannot entirely ignore the role of the
perceived ‘passive’ recipient. To fully embrace the reciprocity of interactions as expressed by
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
15
the third law, one need also embrace the idea that interactions are not composed of two actions
belonging to two entities but is a single phenomenon of reciprocal action.
In the end, despite undoubtedly being a conceptual revolutionary, I think Bunge was not
fully able to embrace the conclusion of his own argument (for details see Ingthorsson 2021:
ch. 4.5). He insists that interactions are a single phenomenon of mutual action, but then rejects
interaction as a basic principle for causation because it would involve the error of singling out
the ‘action’ as cause and ‘reaction’ as effect and therefore characterise the relationship between
cause and effect as symmetric, functional, and non-productive. I am still puzzled why he did
not realise that his own argument suggested instead that we must accept that neither action nor
reaction can count as cause, but that the interaction as a whole—being a single unified
phenomenon—must count as the cause to whatever changes it brings about. On this view, it is
not the baseball hitting a window that causes a breaking; that is only half the story. The
interaction causes not just a breaking but a change in the state of the ball such that we end up
with a ball at rest in a pile of broken glass. Similarly, the mutual action between billiard balls,
regardless of their initial state of motion, causes quantitatively proportional changes in the state
of both balls. It takes an equal amount of work to stop a ball as to move a ball.
It is often objected that the assumed reciprocity of interactions fails to account for the
asymmetry exhibited by many interactions. The window breaks, but the ball does not. The
standard explanation to asymmetries of this kind is simply that since the initial states of the
two things are different, then one and the same influence exerted on them will lead to different
changes. If you hit a window with a hammer it will break. If you hit a rubber ball in exactly
the same way, it will not break. Instead, the hammer may bounce back to hit you in the face.
The observed asymmetry is perfectly compatible with the reciprocity of interactions.
I will admit though that if a case is to be made for a distinction between active and passive
powers and agents and patients—one that still respects the reciprocity of interactions—is to
argue that in many cases the consequences on either side of the interaction must be considered
in some sense graver for one than the other, even though they are otherwise proportional. Davis
Kuykendall (forthcoming) argues just that, developing an earlier suggestion from Anna
Marmodoro (2017). For instance, he suggests that enzymes, as biological catalysts, speed up
chemical reactions without themselves being destroyed by the reaction (Kuykendall
forthcoming: § 4.1). He argues in a similar vein, in §4.2 of the paper, that the interaction
between H2O and NaCl results in the destruction of NaCl but not H2O. My answer is, first, that
asymmetries of that kind do not violate the reciprocity of interactions; they can be explained
by appeal to the differences in the initial states of the two interacting things. Second, that to
find a handful of asymmetrical examples does not justify the conclusion that causal interactions
are generally asymmetrical. It can at best show that sometimes interactions bear signs of some
kind of asymmetry, but not of the kind that contradicts the third law of motion.
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
16
According to the powerful particulars view, causation is best described in terms of
reciprocal action, and that the particular ways that entities can affect each other is determined
by their intrinsic and powerful qualities; qualities whose determinate nature is best discovered
by science. Indeed, science has found out that there are four different kinds of fundamental
interactions, some of which are attractive while others are repulsive, some stronger than others,
and they operate at different distances. We are talking about the four fundamental forces of
nature. However, for all of them it is true—despite differences—that they exert their force
reciprocally between interacting entities.
If causation is reciprocal action which sometimes comes in the form of attraction and
sometimes repulsion, the lesson I draw is that philosophy has only been looking at a subclass
of causal phenomena, notably where previously unconnected entities suddenly interact, often
with very disruptive consequences; brick breaking windows, balls in motion disturbing balls at
rest, etcetera. Philosophy has focused on this subclass because of the assumption that causation
involves a unidirectional influence of something on something else and therefore essentially
involves an external compulsion. Also, of course, because the interactions taking place
between the component parts of objects could not be observed until quite recently. From the
assumption that causation involves an external compulsion, it follows that anything happening
inside an object, especially if it does not produce visible changes on the surface, is judged to
be non-causal. Consider that the definition of spontaneous change is of a change that happens
in the absence of an external compulsion. Even with the knowledge that only compound entities
decay and knowing that compounds are held together by interactions between the component
parts (which, on the hypothesis being considered, are causal), the idea that causation involves
external compulsion has made it difficult to think of decay as a causal process. Hence it is
called spontaneous decay. However, if causation is the phenomenon of reciprocal action, and
such interactions occur within compound entities, then interactions between the parts of
compound objects also are causal. Even if such interactions do not always produce visible
disruption and change but instead appear only to maintain the inner structure of the entity, this
is no reason to conclude that constitution and persistence are non-causal. Indeed, considering
our current knowledge of the physical constitution of compounds it is misleading to say that
they maintain an inner structure, at least if you interpret ‘maintain’ as simply remaining the
same. As far as I know there are no interactions that do not involve some form of continuous
change, even though many of them also preserve certain structures. All known compounds,
beginning with protons and neutrons as compounds of quarks, are thoroughly dynamic entities
in the sense that while they may stay the same structurally, they are still continuously changing
on a more fine-grained level. Indeed, that change is what gives them their stability, which is
why it makes sense to say that continuous reciprocal actions between the parts of a whole can
produce stability, since that stability is dynamic. Just consider any atom of your choice. They
are composed of a nucleus of protons and neutrons ‘encircled’ by a cloud of electrons. Every
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
17
component part of that atom, as well as the atom as a whole, is continually changing. It follows
that everything made up of such atoms is continually changing too. Consequently, such wholes
are processes, if by process we mean any entity for which change is essential.
I hope I have now said enough to make it possible to compare the four different accounts
with respect to how well they handle two different concrete cases: (i) collision between billiard
balls, and (ii) water dissolving salt.
7. Case Study I: Collisions Between Billiard Balls
7.1 Transmission Accounts
The inadequacies of the transmission account when it comes to explain collisions between
billiard balls has already been drafted in some detail above, so I’ll be brief but still add a couple
of details. I said that transmission accounts appear to make intuitive sense in cases like ‘ball in
motion acts on ball at rest’, but not when two identical billiard balls moving at the same speed
in opposite directions collide head on; the account doesn’t generalise to fit all the cases.
Furthermore, it is relevant to point out that it is only under the assumption that the balls are
perfectly rigid and friction against the table is ignored, that transmission is only between the
balls. In the real world of billiard—considering only symmetrical collisions—the balls take
away from the interaction an equal share of conserved quantities, but some energy is dissipated
as heat and as sound. We now have at least three directions in which conserved quantities flow.
Which of these directions represents the direction of causation? To say it is the direction that
matters most to the player is to decide to only look at one side of the story for anthropocentric
reasons. My suggestion is that a deeper account of how interacting particulars exert an
influence on each other is needed, one which explains why the conserved quantities are
distributed/changed the way they are when particulars interact. A mere description of the actual
exchange of quantities doesn’t answer that question. If you look at the full range of cases, there
is no single direction in which quantities flow between entities.
Traditionally, the concept of ‘force’ has served as the explanation of how interacting
particulars influence each other. However, in the friction between reductionistic empiricism
and anti-reductionist rationalism, the concept of force has been just as controversial as the
notion of power. Empiricists like Hertz (1956) and Mach (1919) wanted to get rid of the
concept. To them the notion of force was an unnecessary postulate based on a redundant
inference from observed changes to some imagined invisible cause to those changes. They
instead wanted to describe interactions between material entities merely in terms of the changes
in the state of motion that they can be observed to suffer. Accordingly, the second law of motion
should not really be understood as saying that in any interaction there is this special thing, a
force, that is proportional in magnitude to the object’s mass times the acceleration it suffers,
but as stating an identity of force and change in state of motion. Really, the third law of motion
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
18
(F1on2 = –F2on1) can then just as well be expressed by a reconstructed two-way second law; m1
x a1 = m2 x a2. In plain English, when two material systems interact, the observed change in
the state of motion of the first (m1 x a1) is always equal in magnitude to the observed change
in the state of motion of the second (m2 x a2). But, having effectively removed the notion of
force from the equation, we no longer have an explanation as to what it is that causes the
changes in the two material systems; we just have a description. Why should we accept a
reduction of this kind? The empiricist answer is that we otherwise must appeal to the
mysterious notion of force.
Contemporary proponents of transmission should not have any qualms about the concept
of force, in so far as they seek to ground their view in the notions already in use in the natural
sciences. Appeals to forces is ubiquitous in physics. However, to accept it, transmission
theorists have to accept that there is a more fundamental feature to physical interactions than
the transmission of conserved quantities, notably something (force/influence) that makes the
quantities be transmitted in a particular way. Transmission accounts avoid commitment to the
reality of forces and therefore end up being merely descriptive and in fact empirically
inadequate when we consider the full range of interactions.
7.2 Mechanistic Accounts
The worries I have about mechanical explanations of billiard balls colliding, are, first, that it
isn’t clear to me whether they would be treated as reciprocal (and therefore more like
constitutive production) or unidirectional (and therefore more like precipitative production).
The difficulty is partly to decide whether the change suffered by a ball in motion would be
considered of ‘lesser’ importance than the change suffered by the ball at rest, say, because it
matters more to the player that a ball at rest goes into a pocket. However, for an experienced
player it is equally important to down a ball as it is to place the cue ball in position for the next
shot. They must consider the effects on both balls equally. Is it the novice or experienced player
that is best equipped to decide which side of the transaction matters? To my mind, the criteria
for judging the direction of interactions in terms of ‘importance’ seem clearly anthropocentric
and will be a case of deciding on the basis of the interest of individual players which side of
the interaction they favour. At the very least, assuming the third law is valid, whether any
change is ‘lesser’ in importance must be wholly unrelated to the purely physical magnitude of
any change. We must then still accept the reciprocity of interactions but, like Kuykendall
(forthcoming) and Marmodoro (2017), may perhaps look elsewhere for objective criteria for
treating interactions as asymmetric. Kuykendall and Marmodo´s suggestion is that if A interacts
with B with the result that B breaks but not A, then we have a kind of ‘directedness’. Perhaps
not one that could be accounted for in terms of quantities, but still be objectively real. However,
the handful of examples that might plausibly be considered asymmetric in this sense, would
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
19
only show that some interactions are directed; we are short of a generalisable account of
causation. More about that when discussing water dissolving salt.
Second, I am uncertain of how to apply the notion of mechanism to collisions between
billiard balls, especially if we are to follow the idea that every interaction is an activity between
parts of a mechanism. How exactly are the billiard balls part of a mechanism? I am worried
that the senses in which they are part of a mechanism is either anthropocentric or turns out to
include the entire universe. To be sure, the balls and table are designed to allow the balls to roll
in certain restricted ways, for the purposes of the game. We also have the cue and the players,
but where do we draw the lines for this particular mechanism? Is it only the table, cue, balls,
and players, or is it also the room, the building, the planet, and solar system? There doesn’t
seem to be any particular physical bond between balls, table, cue, and players that organise
them in any particular way in which they are not equally connected to the floor on which the
table is standing, or the planet on which the building stands. The molecules of each individual
ball are however clearly connected to each other in a way the ball is not connected to the table.
But such bonds do not obtain between the balls, or between them and table. To be sure, the
balls, table, cue, and players form a unity in our minds, but that would again introduce an
unwanted anthropocentric feature. If proponents of NMP want the theory to be a contender
among mind-independent theories of productive causation they must provide an account of
how billiard balls are parts of a mechanism that doesn’t rely on human cognition.
The question perhaps ultimately is whether it is unnecessarily confining to require that
every causal interaction be between parts of an already existing organised whole. Can we not
talk about interactions between previously unconnected entities? If the case of billiard balls
colliding is not persuasive enough, consider the random collision between two pieces of rock
drifting aimlessly in space. How can we understand their interaction as an activity of the parts
of an organised whole? And if we can’t, is their collision not causal?
An alternative is to ask whether previously unconnected entities may achieve some kind of
unity when the interaction starts, such that they become parts of an organised whole during the
interaction even if they were not so connected before. I have suggested that this is in fact what
happens in interactions (Ingthorsson 2021: ch. 4). An interaction, on the realist stance taken
here, is a substantial connection between two or more entities. This is clearest in cases when
two entities attract each other to form a compound, but the same is the case in repulsion. Indeed,
once interactions are accepted as the basic mechanism of causation, we have already accepted
that interacting entities are parts of an organised whole. In any interaction—even between
previously unrelated entities—the entities achieve a substantial connection in virtue of the
forces they exert on each other (attractive or repulsive) and then form a unity of parts acting on
each other, albeit the unity sometimes is very short lived. Indeed, the general understanding of
interactions as a unified phenomenon rather than a composite of two separate actions implicitly
support this understanding. Accordingly, two colliding billiard balls become a mechanism on
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
20
contact and continue to be one as long as they are in contact. Indeed, on this view there really
is no difference between precipitating and constitutive production, other than this: precipitating
production is an interaction between previously unconnected entities while constitutive
production is an interaction between already connected parts of a whole. Indeed, I have
suggested that reciprocal action allows us to understand not just the changes that happen inside
a steak when it is being seared as being causal, but also the material constitution of the steak
during periods when the interactions between its constituent parts only serve to maintain the
steak, for instance, while it is being carried home from the butchers (Ingthorsson 2021: Ch. 6).
In sum, my worries about NMP revolves around an uncertainty in the application of the
model to particular cases, an uncertainty that arises partly because it hasn’t been designed to
take an explicit stance on the problems I raise here and elsewhere concerning the ago old view
that exertion of influence is unidirectional (Ingthorsson 2002; 2021), and partly because NMP
does not have an account of what it is for entities to constitute a mechanism that seems to cover
all putative cases of causal interactions. As far as I can see, if NMP were to embrace my account
of causation as reciprocal interaction as the fundamental feature of causation, both these
problems would be solved.
7.3 Powers-based accounts
Powers-based accounts are more heterogenous than transmission and mechanistic accounts and
therefore more difficult to present and criticise all in one go. One general worry already
mentioned is that they tend to explain collisions, like transmission accounts, in terms of an
exertion of influence of one ball on another which receives the influence, which is in conflict
with the result that there are no unidirectional actions. Then there are worries connected to the
specific conceptions of powers.
Dispositional essentialists characterise the powers of the balls in terms of potencies, i.e.,
properties that are not instantiated by the object until the change is being manifested in a
collision. Accordingly, the ball at rest can potentially move, and the moving ball can potentially
make another ball move. My worry is how to reconcile it with the scientific picture. According
to science, the billiard balls are solid and have a shape that together allow them to roll. More
importantly, they have at any given time a quantifiable and directed momentum, which is
represented as a property that the balls have before the collision, and which is what supposedly
makes the balls able to influence each other. None of these properties that figure in the scientific
explanation are pure potencies but actual and occurrent properties of the balls. If there are any
pure potencies in that picture, it would have to be something in addition to these determinate
and fully realised properties. The dispositional essentialist must postulate that in addition to,
and independent from a given balls actual momentum, it has a potency to change its momentum
p1 to another momentum p2 by going continuously through the intermediaries. But on this view
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
21
momentum is itself an inert property, but perhaps one from which emerges a power to change
momentum.
My objection to dispositional essentialism is not that it is incoherent but needlessly
complicated and does not easily link up with the scientific image, which is an ambition for
many dispositional essentialists (for instance, Ellis 2001). Why not just accept, since you have
accepted the reality of properties like momentum, that it is momentum that is responsible for
both resistance to change (inertia) and ability to exert forces on other balls? Why insist that in
addition to having momentum an object has a particular power for every kind of visible
consequence of interactions between objects with momentum (power to make other balls move,
power to change one’s own state of motion, power to make a hollow in a pillow, power to pull
trains, power to break windows), especially when you can’t avoid appealing to momentum
when explaining all these various consequences? Do we want to say that the ball broke the
window because it had the power to break windows? These conceptual quirks are a worry in
addition to the main problem, that powers-based accounts do not take into account the
established fact that whenever any object whatsoever acts on any other object whatsoever, the
latter always acts on the first in the same way to the same magnitude and at the same time.
The identity theory of powers identifies an objects power to change its own state of motion,
as well as of other objects, with momentum. Hence it does not add to the set of natural
properties defined by the sciences an infinite set of powers or dispositions corresponding to
every distinguishable kind of behaviour. However, with the notable exception of myself
(Ingthorsson 2002 and 2021) and John Heil (2012), proponents of the identity theory have not
generally taken to describing interactions in any other way than in terms of mutual
manifestations of reciprocal disposition partners where passive and active powers jointly
contribute to a change in the object with the passive powers. I outlined what I think is wrong
with that view in §5.
It bears to mention that both C. B. Martin (1993) and Mumford and Anjum (2018) are
sceptical to the distinctions between active/passive and agent/patient and so do not characterise
mutual manifestations in those terms. But they do not base this scepticism on the fact that
whenever any object whatsoever acts on any other object whatsoever, the latter always acts on
the first in the same way to the same magnitude and at the same time. In conclusion, powers-
based accounts are by and large incompatible with the reciprocity of interactions, although this
is not true of one or two alternatives. However, those who are compatible with the reciprocity
of interactions are so for completely different reasons than the one’s I outline in §5.
7.4 Reciprocal action between powerful particulars
According to the powerful particulars view, colliding balls each have their own directed
momentum p1 and p2. Momentum is at the same time the power to resist changes in the state
of motion and to change the state of motion of other balls. On contact the balls exert an equal
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
22
and oppositely directed force on each other, that mutual exertion of influence being the cause
to a change in their respective momenta from p1 to p1* and p2 to p2*. To repeat: the cause is the
interaction as a whole and the effect is the sum total of changes suffered by the interacting
entities. This formula applies to any collision regardless of initial state of any ball.
Note that there is an unresolved issue about the exact understanding of the nature of force.
I lean to the understanding that forces are not properties of the balls either prior to collision or
during the collision. To be exact, I do not consider them properties in virtue of which the balls
exert influence; forces do not really push or pull, objects do. I understand the terms ‘force’, in
this particular case, to denote the magnitude of the influence that balls with momentum exercise
on contact. Similar understanding can apply to the force objects exercise in virtue of various
other powers, e.g., those related to charge, spin, etc. In this sense the concept of force is an
abstraction, not the least in light of the fact that forces only arise in interactions, and the fact
that interactions are not considered to be composed of two separate entities, the action and
reaction. However, by saying the concept is an abstraction I am not saying it does not relate to
a real phenomenon. The real phenomenon is reciprocal action, the mutual exertion of influence
between two entities, and the abstraction is the conceptualisation of reciprocal action as having
two sides because it affects two (or more) entities. My understanding of forces comes close to
that of Johansson (1989: 167–8) and Massin (2009), in that they treat them not as properties of
an object but as something real that essentially holds symmetrically between objects, but I am
uneasy about understanding reciprocal action as a relation. If it is a relation it is a very special
relation since it is an efficient relation and so more like an activity or process. Note that Massin,
like Bunge, takes mutual forces to be non-causal relations because he thinks of them as
symmetrical and therefore cannot be productive, and yet he thinks they are relata of production.
My view also has affinity with Jessica Wilson’s view that forces are aspects of the objects that
exert them and therefore not something in addition to the object and its properties, although I
am uneasy with her description of the objects as non-causal entities (Wilson 2007). As I
mentioned earlier, Newman has recently offered an analysis of the collision of billiard balls
that coincides with mine (Newman 2022: 307–8).
8. Case Study II: Water Dissolving Salt
The scientific explanation of water dissolving salt appeals, first, to the properties of H2O and
NaCl molecules. Molecules of H2O are covalent dipole compounds (have negatively and
positively charged poles). NaCl is a nonpolar ionic compound. When molecules of H2O and
NaCl come into contact there will be attraction between H2O and either the Na or Cl in NaCl,
depending on the spatial orientation of H2O (negative pole attracted to positive ion, positive
pole to negative ion). The covalent bond between O and H in water is stronger than the ionic
bond between Na and Cl, which is why the tug of war between the various compounds (Na and
Cl also attracting each other) ends in the breaking of the ionic bond of NaCl but not the covalent
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
23
bonds in H2O. The dissolution continues as long as individual water molecules can interact
with individual NaCl molecules. Important to note that the mutual attractions between the parts
of each compound are all considered to be reciprocal.
8.1 Transmission accounts
Transmission accounts are bound to explain the dissolution of salt in water in terms of
transmitted quantities, but I have been unable to find in the philosophical literature any attempt
to do this. Transmission accounts typically chose examples from the domain of
thermodynamics such as the heating or cooling of water (for instance, see Fair 1979; Kistler
1998). I myself find it difficult to see an explanation of water dissolving salt solely in terms of
transmission of conserved quantities.
The standard explanation of why and how water dissolves salt—the one given above—
appeals to electrical charges and electrostatic interaction. One can however come across
thermodynamic accounts of solubility, some of which give the impression of explaining
solubility rather than just describing the thermodynamic aspects of dissolution. They claim to
explain what happens in terms of systems striving for equilibrium. However, in so far as they
only appeal to least-energy principles or say that NaCl breaks because that requires less energy
than H2O breaking, they turn out to be half-explanations. The reason that the breaking of NaCl
is the most energy-efficient outcome is because in water covalent bonds are stronger than ionic
bonds, wherefore it takes more energy to break the covalent bonds. But the relative strength of
those bonds is not decided by transmission of conserved quantities.
In the end it seems difficult to explain attractions of any kind solely in terms of transmission
of conserved quantities or a strive towards equilibrium. Indeed, as Tracy Lupher argues, it is
even difficult explain any kind of static interactions in terms of transmitted quantities (Lupher
2009). I think his result is really the same as the more general conclusion I have reached that
transmission accounts fail for symmetric interactions, whether dynamic or static.
The criticism here is that even in the domains of physics and chemistry that operate with
conserved quantities, transmission accounts can at best only be applied with some plausibility
to asymmetric interactions, but then only as approximations.
8.2 Mechanistic accounts
Mechanistic accounts can easily be applied to explain the dissolution of salt by water. All the
interactions taking place are either between the parts of organised wholes, or between organised
wholes that consequently morph into other kinds of organised wholes where interactions
continuously preserve the whole. My only complaint is that NMP does not really address the
question of unidirectionality vs. reciprocality of the interactions, wherefore it appears NMP
can treat interactions sometimes as reciprocal when a given scientific explanation clearly tells
us so, and sometimes as unidirectional when that is suggested by the scientific account. That
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
24
raises the question whether the deciding factor of whether any given interaction is treated as
unidirectional or reciprocal hinges on whether the scientific explanation for that particular
interaction is considering the whole phenomenon or is only concerned to explain one side of
the transaction. This ambiguity derives, arguably, from the fact that its proponents have not
been aware of the problems I have raised here and elsewhere, and so have not taken any
measures to respond to it.
8.3 Powers-based accounts
Powers-based accounts explain water dissolving salt in two different ways, depending on
whether they relate only to the manifest or also to the scientific image. According to the former
approach, the power to dissolve salt is attributed to water as a body of matter, and to salt the
power to be dissolved. Accordingly, when salt is put in water the two powers mutually manifest
the dissolution of the salt. This kind of explanation is open to two objections: (i) that it presents
causation as a phenomenon involving unidirectional influence of the kind the natural sciences
say does not exist, and (ii) that it simply does not even address the fact that the scientific
explanation of the phenomenon ties the ability to dissolve not to water as a body of matter, but
to the properties of individual H2O and NaCl molecules. That raises the worry that these
accounts at best relate to the way we ordinarily think about water dissolving salt, but not to
what really happens.
The second approach takes the scientific explanation as its starting point and says that
covalent dipole H2O molecules have the power to break the ionic bond in NaCl molecules
(Marmodoro 2017; Kuykendall forthcoming). When such particles interact, the power to break
and the power to be broken manifests the breaking of NaCl. This explanation is also open to
the objection that it assumes interactions are unidirectional and is therefore in conflict with the
third law. However, it does ground the powers of water and salt on the physical properties of
the molecules, and therefore cannot be said to ignore the scientific explanation. On the other
hand, the account does not identify the powers with the physical properties recognised by the
sciences. The assumption is that in addition to H2O having one slightly positively charged end,
and one slightly negatively charged end, which gives the molecule the ability to interact
electrostatically with other charged molecules (mutually attract or repel), it also has the more
specific power to ‘break NaCl’. As far as I know, chemistry does not postulate any such specific
property in addition to the already mentioned physical properties, but it acknowledges that for
NaCl to break is a known consequence of electrostatic interaction between H2O and NaCl.
Importantly, it is clear that the properties of H2O alone are not enough to ground its power to
break NaCl. It can only be said to have that power with respect to the specific properties of the
NaCl, notably that it is an ionic compound. Indeed, power-based accounts rarely explicitly
address the forces of push and pull operating between the molecules, but only on the
consequences of the forces exerted—NaCl breaks—and then say that the strength of the
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
25
covalent bond, and/or the strength of the ion-polar electrostatic interaction between water and
salt, gives H2O the power to break.
I should address explicitly that Anna Marmodoro (2017) and Davis Kuykendall
(forthcoming) complain that my account of causation in terms of reciprocal action “fails to
capture the directionality of the causal process, which is underpinned by the different ‘actions’
of salt on water and water on salt respectively” (Marmodoro 2017: §7). On the one hand, it is
true that I say that molecules exert an equal and oppositely directed force on each other, and
that the interaction is in that sense symmetrical. On the other hand, I say that this symmetry
allows for very different outcomes for each interacting entity in so far as they are different from
the outset. One and the same kind of influence can result in different types of changes in each
interacting entity. In other words, we can talk about reciprocal actions as always being
symmetric in terms of magnitude of influence and yet distinguish between symmetric and
asymmetric interactions when talking about (i) interactions between two similar entities that
produce similar changes in both, and (ii) interactions between two dissimilar entities that
produce different changes in each. My account therefore perfectly well explains why NaCl but
not H2O is destroyed in the interaction, and apparently without having to appeal to the kind of
directionality that Marmodoro and Kuykendall are talking about. Indeed, I argue we should not
appeal to it if we are to arrive at an account that applies to the full range of interactions and in
a manner consistent with the third law. My argument doesn’t show that there is no way to
consider one side of the transaction as ‘graver’, only that previous ways of motivating that
conclusion conflicts with established scientific facts.
I can in turn complain that Marmodoro and Kuykendall’s accounts of water dissolving salt
bear some signs of the kind of half-explanations I have mentioned before. To be fair,
Kuykendall recognises that the breaking of NaCl is not the only outcome. Another outcome is
the production of two different kinds of formations, which significantly alter the properties of
the resulting liquid. Several molecules of H2O will surround each Na+ and Cl- ion because of
strong ion-dipole interactions between them to form what is called a ‘hydration shell’. We now
really have a liquid that no longer is made up only of H2O molecules connected by hydrogen
bonds, but a mixture of molecules of which some are connected by hydrogen bonds but others
by ion-dipole bonds. All of this, I argue, can be explained by reciprocal actions between the
parts in the liquid. The point is that Kuykendall’s explanation, detailed as it is, is not complete
in relation to the whole phenomenon since it focuses almost entirely on the ‘water dissolves
salt’ aspect, rather than ‘water and salt merge to form saline’.
Mumford and Anjum indeed take the position that we shouldn’t think in terms of water
dissolving salt but instead that salt and water together produce saline, and similarly that sugar
and water produce a sweet solution (2011: 123). However, while much more fully recognizing
the reciprocity of interactions it is not because they accept the kind of reciprocity I have been
arguing for here, or even are aware of the problem stemming from the rejection of
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
26
unidirectional action, and their account still retains a notion of ‘agency’, i.e., of something
being the pivotal trigger to a change that upsets a state in equilibrium, say, a match being struck
in a place with oxygen and flammable material. Indeed, they don’t think so called
‘countervailing’ powers should be included in the notion of cause (2011: 33–4). So, while we
often come to similar conclusions, our accounts are decidedly different.
8.4 Powerful particular account
According to the powerful particulars view, the dissolution of salt by water can be explained
fully in terms of reciprocal interactions between the component parts of water and salt. The
mutual electrostatic attraction between H2O molecules on the one hand, and the components
of NaCl on the other, results in separation of NaCl into Na+ and Cl- ions and the formation of
hydration shells around each ion. The result is saline. Not only does the powerful particulars
view agree in this way to the scientific understanding of ‘water dissolves salt’ but also of the
formation of H2O and NaCl molecules respectively, as well as of the constellations of Na+ and
Cl- ions surrounded by a hydration shell. It does so without postulating a distinct second order
class of dispositions/powers. It is all down to a question of who is the winner in various tugs
of war between different entities, tugs of war whose strength is determined ultimately by the
different strengths of the fundamental interactions on which they are based. Indeed, reciprocal
interaction is a viable candidate for being what can give rise to the kind of organised wholes
that mechanistic philosophers call ‘mechanisms’.
9. Conclusion
In sum, I have argued that transmission, mechanistic and powers-based accounts do not offer
a generalizable explanation of physical phenomena, mainly because they either assume that all
influence is unidirectional or open to it being both unidirectional and/or reciprocal.
Transmission accounts appear unable to account for mutual attractive influence and so are
unable to account for the creation of the kind of organised wholes that would fit the description
of mechanisms. Indeed, mechanistic accounts also struggle to explain how previously
unconnected entities could become parts of a mechanism, and so struggle to explain how
interactions between such entities would count as causal. I have suggested that my account of
causation in terms of reciprocal action would resolve that problem.
It bears to mention that there is one salient feature of the world that I haven’t worked out
how, or whether, the powerful particulars account could explain, and which may perhaps be
the main reason Bunge, Marmodoro and Kuykendall are persuaded that all interactions cannot
really be reciprocal and that my account fails to account for some kind of causal directionality
present in the world. I am talking of the kind of feedback loops that we find in biological
systems. Or, really, self-organising structures whose parts act on each other in a manner that
suggest linear progression in a certain direction. The Krebs cycle is a good example. Every
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
27
particular interaction that takes place in the cycle appear to be reciprocal and yet the process
as a whole is directed in such a way that it repeats the same pattern again and again. Indeed,
this would apply to the explanation of the role of enzymes as catalysts, as Kuykendall mentions
(forthcoming), to maintain a directed process. An explanation of that kind of direction is
needed, but I doubt it will come in the form of overthrowing the third law of motion and
vindicate the active/passive or Agent/Patient distinctions. One possibility is that this kind of
directionality can be grounded partly on the directionality of the relation between the
successive states of any organised whole of reciprocally acting parts, A—the one-sided
existential dependence between producer and product—but which will have to be constrained
somehow by the structure of a larger organised whole B of which A is a part. To work out the
viability of that idea is work for the future.
Acknowledgements
I am very grateful to Yafeng Shan, Ingvar Johansson, and an anonymous referee, whose
comments and suggestions decidedly helped to improve on the initial draft of this paper.
References
Aristotle (1998), Metaphysics, H. Lawson-Tancred (trans.), London: Penguin.
Bechtel, W. and Adele Abrahamsen (2005). ’Explanation: A Mechanistic
Alternative’, Studies in History and Philosophy of the Biological and Biomedical
Sciences 36: 421–41.
Bogen, J. (2008). ‘Causally Productive Activities’, Studies in History and Philosophy of
Science, 39: 112–23.
Bunge, M. (1959). Causality and Modern Science, New York: Dover. Rpt. 1979.
Craver, C. F. (2007). Explaining the Brain: Mechanisms and the Mosaic Unity of
Neuroscience, Oxford: Clarendon Press.
Craver, C. F. and Lindley Darden (2013). In Search of Mechanisms: Discoveries across the
Life Sciences, Chicago: University of Chicago Press.
Darden, Lindley (2008). ‘Thinking Again About Mechanisms’, Philosophy of Science 75 (5):
958–69.
Dowe, P. (1992). ‘Wesley Salmon’s Process Theory of Causality and the Conserved Quantity
Theory’, Philosophy of Science 59: 195–216.
Dowe, P. (2009). ‘Causal Process Theories’, in Helen Beebee, C. Hitchcock & P. Menzies
(eds.), The Oxford Handbook of Causation, Oxford: Oxford University Press.
Ellis, B. (2001). Scientific Essentialism, Cambridge: Cambridge University Press.
Emmet, Dorothy (1985). The Effectiveness of Causes, Albany: SUNY Press.
Fair, D. (1979). ‘Causation and the Flow of Energy’, Erkenntnis 14: 219–50.
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
28
Galavotti, Maria Carla (2022). ‘Wesley Salmon’, in E. N. Zalta (ed.), The Stanford
Encyclopedia of Philosophy, available online:
https://plato.stanford.edu/archives/fall2022/entries/wesley-salmon/.
Glennan, S.S. (2009). ‘Productivity, Relevance and Natural Selection’, Biology and
Philosophy 24: 325–39.
Glennan, S.S. (2017). The New Mechanical Philosophy, Oxford: Oxford University Press.
Heil, J. (2012). The Universe as We Find It, Oxford: Oxford University Press.
Hellingman, C. (1992). ‘Newton’s Third Law Revisited, Physics Education 27(2): 112–5.
Hertz, H. (1956). The Principles of Mechanics, New York: Dover.
Hume, D. (1748). Enquiry Concerning Human Understanding, Harvard Classics, Vol. 37,
Boston: MA: P.F. Collier & Son, 1910.
Ingthorsson, R.D. (2002). ‘Causal Production as Interaction’, Metaphysica 3: 87–119.
Ingthorsson, R.D. (2013). ‘Properties: Qualities, Powers, or Both?’, Dialectica 67 (1): 55–80.
Ingthorsson, R.D. (2016). McTaggart’s Paradox, New York: Routledge.
Ingthorsson, R. D. (2021). A Powerful Particulars View of Causation, New York: Routledge.
Open Access: https://doi.org/10.4324/9781003094241.
Jaeger, G. (2021). ‘Exchange Forces in Particle Physics’, Foundations of Physics 51:13
https://doi.org/10.1007/s10701-021-00425-0
Johansson, I. (1989). Ontological Investigations: an Inquiry into the Categories of Nature,
Man and Society, New York: Routledge. Rept. 2004 by Ontos Verlag. Open Access:
https://www.degruyter.com/downloadpdf/title/303559.
Kant I. (1787). Critique of Pure Reason, N. Kemp Smith (trans.), New York: St. Martin’s Press,
1965.
Kistler, M. (1998). ‘Reducing Causality to Transmission’, Erkenntnis 48: 1–24.
Kitcher, P. (1989). ‘Explanatory Unification and the Causal Structure of the World’, in P.
Kitcher & W.C. Salmon (eds.), Scientific Explanation, Minnesota Studies in the Philosophy
of Science, vol. 13. Minneapolis: University of Minnesota Press: 410–506.
Kuykendall, D. (forthcoming). ‘In Defense of the Agent and Patient Distinction: The Case from
Molecular Biology and Chemistry’, British Journal for the Philosophy of Science:
https://www.journals.uchicago.edu/doi/abs/10.1086/715470.
Lupher, T. (2009). ‘A physical critique of physical causation’, Synthese 167: 67–80.
Mach, E. (1919). The Science of Mechanics: A Critical and Historical Account of its
Development, T. McCormack (transl.), Chicago: The Open Court Publishing Company.
Machamer, P. (2004). ‘Activities and Causation: The Metaphysics and Epistemology of
Mechanisms’, International Studies in the Philosophy of Science 18(1): 27–39.
Marmodoro, Anna (2017). ‘Aristotelian Powers at Work: Reciprocity Without Symmetry in
Causation’, in J. Jacobs (ed.), Causal Powers, Oxford: Oxford University Press, 57–76.
This is a draft of a chapter/article that has been accepted for publication by Oxford University Press in the
forthcoming book Alternative Approaches to Causation: Beyond Difference-Making and Mechanism edited by
Yafeng Shan due for publication in 2023.
29
Martin, C. B. (1997). ‘On the Need for Properties: the Road to Pythagoreanism and Back’,
Synthese 112: 193–231.
Martin, C. B. (1993). ‘Power for Realists’, in K. Cambell, J. Bacon and L. Reinhardt (eds),
Ontology, Causality, and Mind: Essays on the Philosophy of D. M. Armstrong, Cambridge:
Cambridge University Press: 175–83.
Massin, O. (2009). ‘The Metaphysics of Forces’, Dialectica, 63(4): 555–89.
Maxwell, J.C. (1877). Matter and Motion, Dover, New York.
Molnar, G. (2003). Powers: A Study in Metaphysics, Oxford: Oxford University Press.
Mumford, S. & Anjum, Rani (2011). Getting Causes from Powers. Oxford: Oxford University
Press.
Mumford, S. & Anjum, Rani (2018). ‘Powers and Potentiality’, in K. Engelhard and
M. Quante (eds), Handbook of Potentiality, Dordrecht: Springer: 261–78.
Newman, A. (2022). ‘A causal ontology of objects, causal relations, and various kinds of
action’, Synthese 200(4): article 308. https://doi.org/10.1007/s11229-022-03752-5
Resnick, R., Halliday, D., & Krane, K. (2002). Physics (5th ed.), New York: John Wiley &
Sons.
Russell, B. (1912). ‘On the Notion of Cause’, Proceedings of the Aristotelian Society, New
Series 13: 1–26.
Salmon W. (1980). ‘Causality: Production and Propagation’, in Sosa, E. and Tooley, M.
(eds.), Causation, Oxford University Press, Oxford, 154–71, 1993.
Salmon, W. (1984). Scientific Explanation and the Causal Structure of the World, New Jersey:
Princeton University Press.
Steinberg M. S., Brown D. E., and Clement J. (1990). ‘Genius is not immune to persistent
misconceptions’ Int. J. Sci. Educ. 12: 265–73.
Taylor, R. (1973). Action and Purpose, New York: Humanities Press.
Wilson, Jessica (2007). ‘Newtonian Forces’, Brit. J. Phil. Sci. 58: 173–205.
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
The basic kinds of physical causality that are foundational for other kinds of causality involve objects and the causal relations between them. These interactions do not involve events. If events were ontologically significant entities for causality in general, then they would play a role in simple mechanical interactions. But arguments about simple collisions looked at from different frames of reference show that events cannot play a role in simple mechanical interactions, and neither can the entirely hypothetical causal relations between events. These arguments show that physics, which should be authoritative when it comes to the metaphysics of causality, gives no reasons to believe that events are causal agents. Force relations and some cases of energy-momentum transfer are examples of causal relations, with forces being paradigmatic in the macroscopic world, though it is conceivable that there are other kinds of causal relation. A relation between two objects is a causal relation if and only if when it is instantiated by the two objects there is a possibility that the objects that are the terms of the relation could change. The basic metaphysics of causality is about objects, causal relations, changes in objects, and a causal primitive. The paper also includes a discussion of the metaphysics of forces and a discussion of the metaphysics of energy and momentum exchanges.
Article
Full-text available
The operation of fundamental forces in quantum field theory is explicated here as the exchange of particles, consistently with the standard methodology of particle physics. The particles involved are seen to bear little relation to any classical particle but, rather, comprise unified collections of compresent, conserved quantities indicated by propagators. The exchange particles, which supervene upon quantum fields, are neither more fundamental than fields nor replace them as has often previously been assumed in models of exchange forces. It is argued that this explication of quantum field-theoretic exchange forces definitively improves upon its predecessors.
Book
Full-text available
Short abstract: (The DOI link brings you to a downloadable copy of the book) This book critically examines the recent discussions of powers and powers-based accounts of causation. The author then develops an original view of powers-based causation that aims to be compatible with the theories and findings of natural science. Recently, there has been a dramatic revival of realist approaches to properties and causation, which focus on the relevance of Aristotelian metaphysics and the notion of powers for a scientifically informed view of causation. In this book, R.D. Ingthorsson argues that one central feature of powers-based accounts of causation is arguably incompatible with what is today recognized as fact in the sciences, notably that all interactions are thoroughly reciprocal. Ingthorsson’s powerful particulars view of powers-based causation accommodates for the reciprocity of interactions. It also draws out the consequences of that view for issue of causal necessity and offers a way to understand the constitution and persistence of compound objects as causal phenomena. Furthermore, Ingthorsson argues that compound entities, so understood, are just as much processes as they are substances. A Powerful Particulars View of Causation will be of great interest to scholars and advanced students working in metaphysics, philosophy of science, and neo-Aristotelian philosophy, while also being accessible for a general audience. Precis of the book: 1. Introduction In the 20th Century, philosophy has favoured the neo-Humean view that there are no substantial connections in nature. The 21st Century has seen a revival of causal realism and opposition to neo-Humeanism. This book is a constructively critical discussion of the realist revival. It is critical because it aims to show (i) that contemporary realists often inadvertently incorporate into their accounts of causation certain ideas that really belong to the neo-Humean framework they object to, (ii) that the incorporation of these neo-Humean ideas are an obstacle to the further development of causal realism, and (iii) that some of the Aristotelian components that causal realists are seeking to revive are arguably incompatible with the theories and findings of modern science; in other words, they are falsified by science. The book is constructive in that it suggests a way to avoid the fallacies involved in (i)–(iii) to arrive at a modified powerful particulars view that is compatible with modern science. In this chapter the author briefly sets the stage for the discussion, introducing the relevant competing views and central concepts. 2. Powers-Based vs neo-Humean Approaches This chapter introduces a distinction between philosophical attempts to characterize vs. explain a phenomenon and argues that powers-based accounts of causation are best understood as attempts to explain, while neo-Humean accounts can only hope to characterize causation. As a consequence, it is argued, powers-based accounts need not be understood as rivals to neo-Humean approaches. The explanation that powers-based approaches offer, entail that causation appears to us in the way that neo-Humean approaches characterize causation. Neo-Humean accounts are only contrary to powers-based accounts when seen as committing to the basic tenet of neo-Humean metaphysics, notably that there are no substantial connections in nature. However, it is not obvious that everyone that endorses neo-Humean accounts of causation also commits to the ontology of neo-Humean metaphysics. 3. Causal Realism Causal realism, the author argues, is not merely a commitment to the reality of causal connections but also an attempt to explain the nature of such connections. All realists do not appeal to powers in trying to explain causation, but instead to notions such as causal processes, transmission of conserved quantities, and mechanisms. These accounts, it is argued, fail to explain causation because they implicitly endorse neo-Humean strictures about sticking to observable correlates. Some powers-based approaches also fail to explain, because they too implicitly endorse some neo-Humean ideas, such as the dispositional analysis of powers, and the two-place relation model of causal connections. Accounts that reject the dispositional analysis and two-place relation model represent a return to what used to be the standard view of causation prior to the rise of empiricism, notably that causation is the production of changes brought about by the interaction of powerful particulars. There used to be a general agreement about this general idea across the range of different philosophical schools of thought from Aristotle and well into the early modern period. It is argued that this is the most promising form of causal realism. 4. Causal Production Most powers-based accounts assume, in accordance to the standard view, that interactions are unidirectional; one object acts while another passively receives the influence. This assumption arguably contradicts the theories and findings of natural science, which takes it to be established that unidirectional actions do not exist. They insist that every interaction is perfectly reciprocal; all interacting entities simultaneously act on each other, to the same magnitude, and in the same way. The author shows how the standard view can be modified to accommodate the reciprocity of interactions. The modification requires that we neither treat the action of one, or any other object, as a cause, but only the interaction as a whole. Also, that we treat the sum total of changes produced in all the things involved as the effect. This allows a perfectly intelligible analysis of the idea that cause and effect are simultaneous, and a novel analysis of the generic connection between cause and effect: the effect is made out of the same entities that were involved in its production. That is, it is argued that the most fundamental feature of causation is reciprocal action, or interaction as this is defined by the natural sciences. 5. Causal Necessity This chapter argues that the bulk of the arguments against causal necessity are either arguments against realism in general or they only address a caricature of causal realism—here called relational realism—a view that was constructed by Russell (1912), but which has never been put forward in defense of causal necessity. Hume’s argument against necessary connections is discussed in Ch. 2. Here the focus is on Anscombe’s criticism of a logically/conceptually necessary connection, Russell’s problem of action at a temporal distance, and the problem of interference and prevention. It is shown that the standard view is immune to these arguments, even to Mumford & Anjum’s modified version of the problem of interference and prevention that is meant to apply also to powers-based accounts. 6. Constitution and Persistence Constitution and persistence are typically construed as non-causal phenomena. However, if the account of causation outlined in Chapter 4 is accepted, then it follows that the relationship between the constituents of compound objects, as they are described by the sciences, are causal in exactly the same way as the relationship between previously unconnected particulars are causal when they suddenly interact. The difference is that the interaction between the component parts of a unity does not only produce a change, but also the preservation of a certain structure; they produce persistence. It is argued that this causal account of constitution and persistence provides support for Karen Bennett’s claim that some building relations are causal and has a definite explanatory advantage over Kit Fine’s account of constitution in terms of the sui generis relation of ‘embodiment’. 7. Substance and Process Substance and process ontology are typically construed as complete opposites. In this chapter it is argued that contemporary process ontologists are primarily in opposition to what is identified as the modern analytic view of substance, which is the brain-child of the neo-Humean framework, but not obviously in opposition to the Aristotelian view of substance. It is argued that a process ontology that frames itself in perfect opposition to the Aristotelian view, in particular by denying any common constituent that underlies the distinct stages of a process, is one that can only make sense of change either as variation between temporal parts, the view taken by neo-Humean substance ontologists, or in terms of absolute becoming; which is a becoming that violates the idea that everything has a natural origin. Finally, it is argued that the author’s account of compound entities as unities of interacting parts represents such compound objects as entities for which change is essential for their continued existence, and that they therefore are processes while still being substances in the way dictated by the Aristotelian tradition. 8. Powers Proponents of powers disagree on whether to accept or reject the categorical/dispositional distinction, i.e. the idea that our conception of powers is conceptually incompatible with our conception of quality. Here it is argued that the categorical/dispositional distinction has its roots in the modern analytic view of properties which belongs to the neo-Humean framework. Its allure for contemporary realists is based on (i) misguided epistemological concerns, (i) on the mistaken view that it originates in Locke, and (iii) that Hume’s argument against necessary connections is somehow relevant for realist conceptions of properties. It is shown that the distinction is not present in Locke—he in fact saw properties as both qualities and powers—and that Hume’s argument about necessary connections is just as irrelevant for the contemporary realist debate about properties as for the debate about causal necessity. Finally, building on Locke’s view, a way of thinking of fundamental properties is presented, as primitive natures that we can only understand in virtue of what they do but which we should not think of as being ontologically constituted by these doings. According to this view, properties are both qualities and powers. 9. A Critique of Counterfactual Accounts of Causation In this chapter, it is argued that the intuitive and ontological virtues of counterfactual theories of causation (CTC) are exaggerated. First, contrary to what proponents of CTC assume, then tradition actually dictates that causation explains counterfactual dependence rather than the other way around. It is part and parcel of the standard view that effects are counterfactually dependent on their causes. This fact is rarely, if ever, addressed by proponents of CTC. Second, ontologies according to which the world develops in determinate ways because of substantial connections of some kind—for instance, Platonic laws, natural laws, or powers—have no use for counterfactual theories (even though they may possibly find talk of possible worlds a useful heuristic). Finally, and more surprisingly, Humean metaphysics is ill suited to ground the truth of causal claims on the truth of counterfactual claims. If reality really is contingent, there is nothing about any world that prohibits two near duplicate worlds to differ only in the respect that makes a given counterfactual false. 10. The Contrast to Alternative Views This concluding chapter summarises the contrast drawn up in previous chapters between various approaches to elucidate the nature of causation. That is, between powers-based and neo-Humean views, between powers-based and causal objectivist approaches, as well as between those powers-based approaches that still incorporate neo-Humean ideas and those that explicitly reject them.
Book
Full-text available
This volume is devoted to problems within analytic metaphysics. It defends an ontology and theory of categories inspired by Aristotle, but revised in such a way as to be compatible with modern science. The ontology of both natural and social reality is addressed, starting out from the view that universals exist but only in the spatiotemporal world (immanent realism). In attempting to bring Aristotle's ontology up-to-date, the author relies very much on the thinking of Edmund Husserl, conceiving the cement of the universe as Husserlian relations of existential dependence and regarding intentionality as a non-reducible category in the ontology of mind. The work is thoroughly realistic in spirit, but large parts of it should nonetheless be of interest to conceptualists and nominalists, too.
Chapter
Analytic philosophers have in recent decades rediscovered powers as the basis for an all-encompassing metaphysics and philosophy of nature. What recommends the powers view is its explanatory utility, including a putative explanation of potentiality. Powers can be understood as the elements in the world that provide the grounding for potentiality in actuality. They can be productive of their manifestations but typically do so only in certain circumstances. This will explain why there can be some potentialities that are not actualised and also why there are some constraints on what can be. The powers account also makes potentialities a possible subject of scientific investigation as powers are to a degree empirically accessible. It is important, however, that we provide a plausible account of how powers are able to bring things about: how they able to make some of the potentialities actualities. A mutual manifestation model is preferred to the stimulus-response model of production. It was C. B. Martin who introduced the mutual manifestation model but it is argued that his account needs to be amended so that it resembles less mereological composition and more causation.
Article
A theory of causality based upon physical processes is developed. Causal processes are distinguished from pseudo-processes by means of a criterion of mark transmission. Causal interactions are characterized as those intersections of processes in which the intersecting processes are mutually modified in ways which persist beyond the point of intersection. Causal forks of three kinds (conjunctive, interactive, and perfect) are introduced to explicate the principle of the common cause. Causal forks account for the production of order and modifications of order; causal processes account for the propagation of causal influence.
Article
If the core idea of process theories of causation is that causation can be understood in terms of causal processes and interactions, then the approach should be attributed primarily to Wesley Salmon (1925-2001). Salmon takes causal processes and interactions as more fundamental than causal relations between events. To express this Salmon liked to quote John Venn: 'Substitute for the time honoured "chain of causation", so often introduced into discussions upon this subject, the phrase a "rope of causation", and see what a very different aspect the question will wear'. According to the process theory, any facts about causation as a relation between events obtain only on account of more basic facts about causal processes and interactions. Causal processes are the world-lines of objects, exhibiting some characteristic essential for causation.
Book
This book develops the theory of causal dispositionalism. Others have already suggested that a theory of causation would follow from an ontology of real dispositions or powers. This book attempts to show how. The book argues that powers come together in complex partnerships producing something together that they could not have produced alone. They will do so in a distinctly dispositional way that is not reducible to necessity. The mode of composition of causes can vary and is sometimes non-linear. The book calls this view compositional pluralism. The book argues for the simultaneity of cause and effect as causation is the process that occurs when partnered powers produce their effect. It begins once those partners are together and ends either when the process is complete or is interrupted. The theory explains how causal claims are distinguished from others and why there is no causation by absence. The book distinguishes the distinct, sui generis dispositional modality of causation and show how it can be known directly through experience. The book applies the theory to the science of biology, where it is corroborated. © Stephen Mumford and Rani Lill Anjum 2011. All rights reserved.