Scientific Essentialism: Modernising The Aristotelian View 2015-2017
—Funded by a £186.000 project grant from The Bank of Sweden Tercentenary Foundation
Purpose and aims
The 21st Century has seen a dramatic revival of realist approaches in philosophy to the issue of properties and causation. An important part of this revival is the re-examination of the relevance of Aristotelian metaphysics for the construction of a scientifically informed metaphysics. A focal point in this re-examination is the notion of powers—the idea that natural properties are best understood as determinate ways an object can affect other objects and be affected by them—and how this notion might help to elucidate other notions such as causation, natural laws, natural kinds, and agency.
In this project I will critically examine recent discussions of powers and their role in causation, on the basis of my previous research (in particular 2002, but see also 2007, 2013, and forthcoming), which in turn rely on Bunge (1959). Particular focus is on the finding that a central feature of the Aristotelian view of causation is arguably falsified by modern physics. Roughly, the Aristotelian view depicts interactions between objects as involving a unidirectional exertion of influence of one object (the ‘agent’) upon another (the ‘patient’). The notion of power is accordingly understood to involve a distinction between active and passive powers, which account for the ability of an agent to influence, and the ability of a patient to be influenced, respectively. However, according to modern physics, unidirectional actions do not exist; all interactions are perfectly reciprocal (Resnick, Halliday & Krane 2001). If this is right, all notions deriving from or influenced by the idea of unidirectional actions, risk being false by the same measure. This flaw, although serious, is not fatal, and in my previous work I have sketched a way to modify the Aristotelian view to accommodate the reciprocity of interactions (2002). Indeed, I think this modification serves to strengthen the essentialist view of properties, causation and natural laws.
This project aims to critically examine current discussions of powers-based accounts in light of my earlier findings, and to elaborate on my positive account of causation, especially with respect to the following interrelated issues.
- Is there a distinction between active and passive powers, and a corresponding distinction between agents and patients?
- How should the notion of mutual manifestation of powers be understood?
- How should we understand and resolve the problem of ‘fit’?
- Are causal connections necessary or contingent, and in what way exactly?
- Can powers ground natural laws, perhaps even without appeal to ceteris paribus clauses?
The concept of causation is central to our understanding the world. For the same reason, it is a key to our understanding of the sciences, both with respect to the methods used to gain knowledge about the causal workings of the world, and the knowledge thus gained. Implicit, then, in our everyday thinking about the world, and in our understanding of science as the systematic exploration of the world, is the idea that objective reality is causal. However, mainstream philosophy has for some time predominantly treated causation as a construct of the mind; a feature of the way we make sense of experiences (or observational data) rather than an objective feature of an external reality. As a consequence, the philosophical study of causation in modern times has, until very recently, mainly focused on the analysis of how we think and speak, and not as much with the results and theorising of the sciences.
Today the tables have turned and realist approaches to properties and causation are again at the cutting edge of metaphysical debate (Mumford 1998; Ellis 2001; Ingthorsson 2002 & 2007; Molnar 2003; Heil 2003 & 2012; Lowe 2006; Bird 2007; Kistler & Gnassounou (eds.) 2007; Martin 2008; Marmodoro (ed.) 2010; Mumford and Anjum 2011; Gibb, Lowe & Ingthorsson (eds.) 2013; Greco & Groff (eds.) 2013; Cartwright & Pemberton 2013; Jacobs (ed.) forthcoming). This is also the case for a range of interrelated subjects, such as the laws of nature (Cartwright 1983; Ellis 2001; Mumford 2004; Bird 2007), and in the discussion about the relationship between various ontological levels, i.e. between the entities studied by physics, chemistry, and biology, respectively (Needham 2009; Hendry 2010, forthcoming), as well as the relationship between the physical and mental (Gibb, Lowe, & Ingthorsson (eds.) 2013). The notion of emergence, and how powers might shed light on emergence, is at the heart of this discussion about ontological levels (Humphreys 1997; O’Connor 2000; Kim 2006; Needham 2009). This project focuses on causation, but will collaborate with research groups whose focus is on laws, ontological levels, and emergence.
Several major European projects, recent and on-going, explore the relevance of powers for our understanding of properties, causation, laws, science, and agency, often explicitly with reference to the theories and findings of the natural sciences, e.g. The Metaphysics of Science (Bristol, Birmingham, Nottingham), The New Ontology of the Mental Causation Debate (Durham), Powers and the Identity of Agents (Innsbrück), Power Structuralism in Ancienct Ontologies (Oxford), Causation in Science (Aas), Causation, Laws, Dispositions, and Explanations (Cologne), Durham Emergence Project (Durham), A Process Ontology for Contemporary Biology (Exeter). This project aims to engage directly with recent and forthcoming outputs from these projects.
Current conceptions of powers and associated powers-based accounts of causation, assume—explicitly or implicitly—the Aristotelian view of the nature and causal role of powers. The few who do not explicitly assume it, still use examples to illustrate their view that comply with the Aristotelian view. Accordingly, these conceptions characterise causation as the unidirectional action of an object possessing active powers (the ‘agent’), on an object possessing passive powers (the ‘patient’). As a consequence, they also depict the outcome of causal interactions as the ‘mutual manifestation’ of active and passive powers (Heil 2003; Molnar 2003; Martin 2008; Mumford & Anjum 2011; Marmodoro forthcoming), and for them the ‘problem of fit’, i.e. why each given power only work with a select few other powers to mutually manifest something, is framed on the assumption that the fit concerns the fit between active and passive powers (see, for instance, Williams 2010). Similarly, arguments concerning the necessity or contingency of causal connections revolve around the connection between the exertion of an active power, and a change determined by a passive power (Schrenk 2009; Mumford & Anjum 2011). Finally, the distinction between the cause and the conditions in which, ceteris paribus, the cause invariantly produces a change, is standardly drawn on the same model. A cause is the exertion of influence by an object possessing active powers, and the conditions are specified in terms of an object or state of affairs whose passive powers, all things being equal, always changes in exactly the same way in response to the action of the agent. Consequently, in philosophy, lawful connections are typically modelled on the connection between active and passive powers, for instance, the way water dissolves salt, or rocks break windows.
As Mario Bunge points out, the idea that causation is the unidirectional action of objects with active powers on objects with passive powers “disregards the fact that all known actions are accompanied or followed by reactions” wherefore Bunge concludes that the “polarization of interacting objects into agents and patients, is ontologically inadequate” (Bunge 1959: 170–1). When Bunge says it is a ‘fact’ that all known actions accompanied by reactions, he is saying that physics, and natural science generally, do not recognise the occurrence of any form of unidirectional action; all the kinds of influences that science recognises are taken to be reciprocal. For instance, all the fundamental forces of nature work reciprocally between the entities subject to them. This represents a very strong argument against unidirectional action. In fact, it indicates that the Aristotelian view may be a potential source to serious misconceptions, e.g. in our understanding of the notion of power itself, of the notion of ‘mutual manifestation’ of powers, of the so-called ‘problem of fit’, of the role of ceteris paribus clauses in our understanding of natural laws, and of the necessity of causal connections.
However, it is important to note that if science is right to think of all influence as reciprocal rather than unidirectional, then this does not contradict the idea that properties bestow upon their bearers distinctive powers to influence and/or be influenced. It only contradicts the idea that the bearers divide into agents that unidirectionally exert their powers upon patients. In ‘Causal Production as Interaction’ (2002), I sketch an account of causation that accepts the reciprocity of actions, but which still retains all other components of the Aristotelian view.
It is also important to note, that even if science is right to represent as reciprocal all the fundamental forms of influence—represented by the fundamental forces and basic chemical reactions—then it may still be possible to argue that macroscopic objects possess emergent properties, not addressed by natural science, that do exert unidirectional influence. Putative candidates of properties of this sort abound in common sense, such as the power of water to dissolve salt. The connection between the fundamental properties that figure in our best scientific theories, and those of common sense is an issue pursued in the Durham Emergence Project, and the Power-structuralism in Ancient Ontologies project, with which I aim to collaborate.
In the exploration of the role of powers in the elucidation of other issues, one has to keep in mind that there is no consensus about the best way to characterise powers, even among their proponents. On the one hand we have the view that powers only represent one of two mutually exclusive types of properties, sometimes called pure powers and pure qualities; one type whose essence is exhausted by its power-bestowing nature, and another type that has no such power-bestowing essence but well enough a causally inert quality (Ellis 2001; Molnar 2003; Bird 2007). On the other hand we have the mixed view that all properties are both powers and quality all at once (Heil 2003; Martin 2008; Jacobs 2011; Ingthorsson 2013).
The main difference between the two views, from the perspective of this project, is that pure powers are typically represented as essentially tied to some particular manifestation, while a mixed view is not. On the pure view, each power is for something very particular, say, power to break is for breaking. On the mixed view, a breaking need not be a manifestation of a power whose only essence is to break, even though the power(s) responsible have a determinate nature. The mixed view takes the manifestation to be an effect of the exertion of the power in particular situations, but does not insist it must always manifest the same thing whenever it is exerted.
The disagreement between the pure and mixed views, bring to light that the term ‘manifestation’ is used ambiguously in the literature. It is sometimes used to denote what a power is potentially capable of, and sometimes it is used to denote what a power actually produces. In the former sense, ‘manifestation’ is taken to be a determinant of the nature of the power whether or not it is exerted, and in the latter to the effect produced when the power is exerted. This ambiguity is closely related to the dispute between a pure and mixed conception of powers, in the sense that proponents of the pure conception tend to characterise the nature of a power in terms of its potential manifestation (Molnar 2003; Bird 2007), while proponents of the mixed view tend to reject a characterisation of powers in those terms (Heil 2003; Martin 2008; Jacobs 2011; Ingthorsson 2013). Indeed, the proponents of the mixed view think the characterisation of a power in terms of its manifestation is to confuse a cause with its effect.
The difference in the way the pure and mixed view represent manifestations of course has repercussions for their This is because a mixed view can accept that a power has a nature independently of whatever it actually manifests when exerted, while the pure view ties its characterisation of the nature of the power to whatever the power actually manifests. This allows the mixed view to be more flexible as to what a power can manifest, which gives it a definite advantage today when it is increasingly admitted that the effect produced by the exertion of powers is never a manifestation of a single power, but always the result of the mutual exertion of two or more powers in an interaction between powerful particulars. This is what is now referred to as mutual manifestation. In the current literature, mutual manifestation is typically illustrated as the combination of active and passive powers, e.g. the power to dissolve and the power to be dissolved, the power to heat and the power to be heated. I take this to indicate that proponents of powers, whether they endorse a pure or mixed view, assume—explicitly and/or implicitly—that the exertion of powers is a matter of unidirectional action of an active object on a passive object. The concept of mutual manifestation is thus potentially biased by the Aristotelian idea of unidirectional action.
The purpose of this project is to engage with the debate and, with explicit reference to the outputs of the projects mentioned above, examine the potential fallacies deriving from the agent-patient distinction, and to compare how my positive account of causation would deal with the issues mentioned above. Below I will address each research question in a little more detail.
- Active vs. Passive Powers, and Agents and Patients.
If one looks at the range of concrete examples that philosophers use to illustrate the nature of causation—including philosophers that reject anything to do with powers and Aristotelian metaphysics generally—they all fit snugly to the Aristotelian model of agent acting in virtue of its active powers on the patient that subsequently changes in virtue of its passive powers: the ball in motion towards a ball at rest, which, on contact, is followed by a motion away by the ball previously at rest (Hume, Treatise: Part III, sect. XIV); the leaden ball dropped on a pillow, thus making a hollow (Kant 1787: A203); the locomotive pulling a truck (Taylor 1973: 35); the short circuit in a wire causing a fire in the attic (Mackie 1974); the baseball breaking the window (Salmon 1984); water dissolving salt (Lowe 2010; Marmodoro forthcoming). Typically these examples, explicitly or implicitly, illustrate causation as the result of a relation between the active powers of one thing and the passive powers of another. It is the active power of the water to dissolve that combines with the passive power of the salt to be dissolved.
However, when the very same examples are described in the language of science, their characteristics are very different. In all cases of collision (between billiard balls, leaden ball and pillow, locomotive and truck, baseball and window…) the two objects concerned both have equal measures of active power to exert a force on each other, and they both affect each others state of motion equally in opposite directions. Every type of mechanical collision plainly involves an equal measure of active powers of all objects involved and the same kind of passive power to change their state of motion. Physics simply leaves no conceptual space to single out one object as the agent and the other as patient.
Other kinds of changes, e.g. chemical changes like water dissolving salt, appear at first blush to be trickier; water dissolves salt, but salt does not dissolve water. However, again, chemical changes of that kind appear only to involve unidirectional action of one substance on another, when considered in the realm of common sense. In common sense we treat the water into which salt is flung as a single entity acting as a unity on the salt, which again is treated as a unity. In the langue of science, the dissolving power of water does not belong to water as a unified body, but to the individual molecules of H2O. Each individual water molecule interacts reciprocally with either the Na+ or Cl– components of salt, creating a mixture of mainly water, sodium, and chloride molecules that are continuously bonding and breaking up in a series of interactions, all of which are reciprocal. One has to be careful to consider where the appropriate power resides.
One aim of the project is to call attention to the problem of unidirectional actions, how it undermines the distinction between active and passive powers as well as the distinction between agents and patients, and explore whether the conception of reciprocal powers of the same kind is a more viable option.
- Mutual Manifestation
Due to the realisation that powers never individually manifest anything, but only in combination with other powers, power-ontologists increasingly accept that manifestations (as used in the sense of ‘effect’) are always the result of a combination of powers. C. B. Martin was perhaps the first to argue forcefully against a conditional analysis of powers—which try to define the nature of single powers in terms of what they can do in certain conditions—and in favour of the view that powers act together to mutually manifest an outcome. Martin argued that we must understand the nature of powers in terms of the multitude of outcomes that a single power can manifest together with an infinity of so-called ‘reciprocal disposition partners’ (Martin 2008). However, when Martin illustrates dispositional partners and what they mutually manifest, he typically (but not consistently) chooses an active power and its putative passive power, say, the power to dissolve and the power to be dissolved. I have not come across any exceptions in the field to this general pattern. The lack of consistency, I suspect, is simply that people use the examples that are popularly used, some deriving from common sense and which are invariably active-passive combinations, others deriving from science and which are more often examples of purely reciprocal powers, say inertia; the power to resist a change in the state of motion, and at the same time the power to change the state of motion in other objects. The inconsistency then only reflects that the authors are unaware of the difference.
The way I have depicted mutual manifestations, is in terms of mutually reciprocal powers of the same kind; two objects interact in particular ways when both share the same kind of power. If both objects have inertia, both will be able to act and be acted upon in the same way. The same seems to hold for all fundamental kinds of interactions, such as between charged particles.
One aim of the project is to call attention to the problem of unidirectional actions, how it undermines the distinction between active and passive powers, its consequences for current conceptions of mutual manifestations, and to explore whether the conception of reciprocal powers of the same kind is a more viable option.
I am aware of a potential problem here. Although all fundamental interactions are reciprocal, the interactions between more complex entities, say between chemical compounds and/or biological systems, may not be. The problem relates to the issue of emergence, which is being investigated within the Durham Emergence Project and A Process Ontology for Contemporary Biology (Exeter)—also to some extent the Power-structuralism in Ancient Ontologies (Oxford), and it is my hope that collaboration with these projects may bring to light implications neither were aware of before, and hopefully lead to innovative solutions to deal with them. Collaboration is already agreed upon with the Oxford and Durham groups, but I am invited to Exeter in the beginning of 2015 and will strive to initiate collaboration at that time.
- The problem of ‘fit’
As already described, it is increasingly admitted that powers never produce a manifestation except in conjunction with some other power. However, particular powers seem only to manifest with a select few other powers, to which they appear to ‘fit’. If water is to successfully exert its dissolving powers, the substance on which it acts better possess the appropriate power to dissolve (it better be ’soluble’). It appears then that the exertion of a power, and its ability to produce a manifestation, is dependent on some other coinciding power that ‘fits’ the other, i.e. is able to receive the influence. This fact gives rise to the question of why and how powers combine in particular ways, and to the question of whether and in what way exactly we can regard powers as intrinsic properties of their bearers.
In discussion of the problem of fit, Williams (2010) being a notable example, the problem is again typically described in terms of a fit between active and passive powers. And again, the exceptions indicate a lack of awareness of the difference of a fit between active-passive vs. reciprocal powers of the same kind.
A third aim of the project is to investigate whether a conception of reciprocal powers of the same kind may resolve the problem of fit; indeed, whether the latter model may not in fact depict mutual manifestations as they already are described by the natural sciences (visavi the fact that science describes the fundamental forces as acting reciprocally between entities with the same power).
- Causal Necessity
From Aristotle until Hobbes (1656) it was widely taken for granted that causation involved necessity. Today, proponents of powers tend to sceptical about necessitation claims. Instead they favour the weaker claim that although the connection between causes and their effects is not entirely contingent, it is not necessary (Mumford and Anjum 2011; Schrenk 2009). This project aims to investigate the reasons for giving up on necessity, and to explore the case for necessity in light of my causal model. The working hypothesis is that the rejection of necessity may originate from what is really a Humean idea—foreign to Aristotelian metaphysics—notably that causation is at rock bottom a relation between two temporally distinct entities (events, states, properties…).
It is clear that current proponents of what is roughly an Aristotelian view of causation represent the relationship between causes—understood as some form of exertion of powers—and effects, in two different ways. On the one hand it is treated simply as a relation between temporally distinct relata, the obtaining of a power, or set of powers, and the obtaining of a manifestation. The clearest proponents of this approach are Molnar (2003), Bird (2007), Schrenk (2009), and Mumford and Anjum (2011). Others assume that the relation itself needs to be explained in terms of some kind of process in which two or more objects continuously interact to produce a manifestation (Ingthorsson 2002a & 2007; Huemer and Kovitz 2003; Chakravartty 2005; Marmodoro 2007). In other words, on the latter view the relation is in turn grounded in a more fundamental process of production, and this process is revealed by a closer examination of the event popularly picked out as the cause.
Characteristically, philosophers who take a relational view of the relation between powers and their manifestations, tend to think of the connection between causes and their effects as susceptible to interference and prevention and thus not strictly necessary (Schrenk 2009; Mumford and Anjum 2011: ch. 3). They think the cause can obtain fully before the effect comes into being, and thus invite the possibility that some preventive factor can obtain after the cause is completed but before the effect can begin.
On the other hand, philosophers who think of causation as a process of production do not accept that a cause can be completed before the effect can begin, but instead characterise the process of production as a continuous temporally extended process in which one thing exerts an influence on another and that the object acted upon simultaneously (but not momentarily) undergoes change as the influence is exerted. That process can be interrupted at any time, but only by physically intervening to stop the exertion of influence. That however only involves another exertion of influence whose outcome is necessary and does not prevent the outcome of any already exerted outcome. Typically, philosophers who endorse a process of production view of causation tend to think causation is necessary (Hobbes 1656; Ingthorsson 2002a, 2007; Huemer and Kovitz 2003; Chakravartty 2005).
The fourth aim of this project is to explore whether the process view of causation really has any problem with interference and prevention, or whether that problem only arises on the assumption of a relational model. The problem of interference and prevention has two distinct forms. On the one there is Russell’s argument in ‘On the Notion of Cause’ (1912), which attacks the general model of causation as a relation between temporally distinct events. On the other hand we have a family of different kinds of counterexamples which profess to show with the use of particular concrete examples that a certain kind of cause can occur to completion and yet not be followed by the effect because of an intervening factor. The intervening factors come in roughly four types: (i) prevention, (ii) pre-emption, (iii) finks, and (iv) antidotes. Roughly, the working hypothesis is that these arguments rely on a relational model in which cause acts upon effect, and do not work on a process view in which action is exerted by an agent on a patient, or between reciprocally acting objects. In this respect, I am in agreement with Chakravartty (2005) and Huemer and Kovitz (2003), but they do not explicitly address the difference between an agent-patient vs. reciprocal interaction.
- Laws of Nature
Powers-based accounts of laws start from the assumption that objects behave like they do in virtue of the properties they possess, rather than because of some abstract entities—the laws—that somehow impose upon things certain manners of behaviour. The question, then, for proponents of powers is really whether or not nature is uniform in its behaviour in the way implied by the laws proposed by the natural sciences.
There are, very roughly, two schools of thought. There is, as in the case of causal necessity, a relational conception of laws, where the uniformity of nature is tied to the relationships that hold between powers and their manifestations (Mumford 2004; Bird 2007; Mumford and Anjum 2011). The proponents of this view of laws tend to see laws of nature as expressions of tendencies rather than necessities (but Bird disagrees), for much the same reasons as they regard causation to be not-quite-necessary.
On the other hand we have the view of Cartwright, according to which laws do not express relationships between the obtaining of a power and the obtaining of a manifestation, but instead expresses the magnitude of the influence being exerted by a power in each instance, in terms of what it would manifest if nothing else was acting also. Each individual law then will only deliver correct predictions in the cases where a single power (or pairs of powers) is exerted, but this is rarely the case in actuality.
With regards to the laws of nature, this project aims to investigate the extent to which a relational conception of powers may be the source to the concerns that lead friends of powers to reject the uniformity of nature and hence the validity of natural laws.