I propose a dynamic causal approach to characterizing the notion of a mechanism. Levy and Bechtel, among others, have pointed out several critical limitations of the new mechanical philosophy, and pointed in a new direction to extend this philosophy. Nevertheless, they have not fully fleshed out what that extended philosophy would look like. Based on a closer look at neuroscientific practice, I propose that a mechanism is a dynamic causal system that involves various components interacting, (...) typically nonlinearly, with one another to produce a phenomenon of interest. (shrink)

Perfect adaptation in a dynamical system is the phenomenon that one or more variables have an initial transient response to a persistent change in an external stimulus but revert to their original value as the system converges to equilibrium. With the help of the causal ordering algorithm, one can construct graphical representations of dynamical systems that represent the causal relations between the variables and the conditional independences in the equilibrium distribution. We apply these tools to formulate (...) sufficient graphical conditions for identifying perfect adaptation from a set of first-order differential equations. Furthermore, we give sufficient conditions to test for the presence of perfect adaptation in experimental equilibrium data. We apply this method to a simple model for a protein signalling pathway and test its predictions in both simulations and using real-world protein expression data. We demonstrate that perfect adaptation can lead to misleading orientation of edges in the output of causal discovery algorithms. (shrink)

In recent debates mechanisms are often discussed in the context of ‘complex systems’ which are understood as having a complicated compositional structure. I want to draw the attention to another, radically different kind of complex system, in fact one that many scientists regard as the only genuine kind of complex system. Instead of being compositionally complex these systems rather exhibit highly non-trivial dynamical patterns on the basis of structurally simple arrangements of large numbers of non-linearly interacting constituents. The (...) characteristic dynamical patterns in what I call “dynamically complex systems” arise from the interaction of the system’s parts largely irrespective of many properties of these parts. Dynamically complex systems can exhibit surprising statistical characteristics, the robustness of which calls for an explanation in terms of underlying generating mechanisms. However, I want to argue, dynamically complex systems are not sufficiently covered by the available conceptions of mechanisms. I will explore how the notion of a mechanism has to be modified to accommodate this case. Moreover, I will show under which conditions the widespread, if not inflationary talk about mechanisms in complex systems stretches the notion of mechanisms beyond its reasonable limits and is no longer legitimate. (shrink)

Real causal systems are complicated. Despite this, causal learning research has traditionally emphasized how causal relations can be induced on the basis of idealized events, i.e. those that have been mapped to binary variables and abstracted from time. For example, participants may be asked to assess the efficacy of a headache-relief pill on the basis of multiple patients who take the pill (or not) and find their headache relieved (or not). In contrast, the current study examines learning (...) via interactions with continuous dynamic systems, systems that include continuous variables that interact over time (and that can be continuously observed in real time by the learner). To explore such systems, we develop a new framework that represents a causal system as a network of stationary Gauss--Markov (``\OU'') processes and show how such \textit{OU networks} can express complex dynamic phenomena such as feedback loops and oscillations. To assess adult's abilities to learn such systems, we conducted an experiment in which participants were asked to identify the causal relationships of a number of OU networks, potentially carrying out multiple, temporally-extended interventions. We compared their judgments to a normative model for learning OU networks as well as a range of alternative and heuristic learning models from the literature. We found that, although participants exhibited substantial learning of such systems, they committed certain systematic errors. These successes and failures were best accounted for by a model that describes people as focusing on pairs of variables, rather than evaluating the evidence with respect to the full space of possible structural models. We argue that our approach provides both a principled framework for exploring the space of dynamic learning environments as well as new algorithmic insights into how people interact successfully with a continuous causal world. (shrink)

Both natural and engineered systems are fundamentally dynamical in nature: their defining properties are causal, and their functional capacities are causally grounded. Among dynamical systems, an interesting and important sub-class are those that are autonomous, anticipative and adaptive (AAA). Living systems, intelligent systems, sophisticated robots and social systems belong to this class, and the use of these terms has recently spread rapidly through the scientific literature. Central to understanding these dynamical systems is their complicated organisation and their consequent capacities (...) for re- and self- organisation. But there is at present no general analysis of these capacities or of the requisite organisation involved. We define what distinguishes AAA systems from other kinds of systems by characterising their central properties in a dynamically interpreted information theory. (shrink)

This paper proposes an account of neurocognitive activity without leveraging the notion of neural representation. Neural representation is a concept that results from assuming that the properties of the models used in computational cognitive neuroscience must literally exist the system being modelled. Computational models are important tools to test a theory about how the collected data has been generated. While the usefulness of computational models is unquestionable, it does not follow that neurocognitive activity should literally entail the properties construed (...) in the model. While this is an assumption present in computationalist accounts, it is not held across the board in neuroscience. In the last section, the paper offers a dynamical account of neurocognitive activity with Dynamical Causal Modelling that combines dynamical systems theory mathematical formalisms with the theoretical contextualisation provided by Embodied and Enactive Cognitive Science. (shrink)

The subject of this chapter is the identity of individual dynamical objects and properties. Two problems have dominated the literature: transtemporal identity and the relation between composition and identity. Most traditional approaches to identity rely on some version of classification via essential or typical properties, whether nominal or real. Nominal properties have the disadvantage of producing unnatural classifications, and have several other problems. Real properties, however, are often inaccessible or hard to define (strict definition would make them nominal). I suggest (...) that classification should be in terms of dynamical properties of systems, starting with individual systems rather than classes, and working up by abstractions that fit causal generalities. The advantage of this approach is that individuality is testable and revisable as we come to know more about systems. Another advantage is that if anything is real, then it is the dynamical. Once I have presented this approach in general, I will show that the central concept of dynamical cohesion (the "dividing glue") is amenable to giving a principled account of individuation as a process, at the same time explaining the origin of diversity. Some other advantages of this approach are presented, including how it can be used as a basis for testable classifications. This last has moral implications, since cohesion at the individual and the social levels, and their interactions, can impinge on proper moral decisions. (shrink)

This paper concerns design thinking (Lawson, 1980), system thinking (systems theory) (von Bertalanffy, 1968), and system dynamics modeling as methodological platforms for analyzing large amounts of qualitative data and transforming it into quantitative mode. The aims of this article are to present an integral (mixed) research process including the design thinking process—a solution oriented approach applicable in the social sciences and humanities which enables to reveal causality in research on societal and behavioral issues. This integral approach is illustrated (...) by an empirical pilot study from art/design-educational environment. (shrink)

A notion of delegated causality is introduced here. This subtle kind of causality is dual to interventional causality. Delegated causality elucidates the causal role of dynamical systems at the “edge of chaos”, explicates evident cases of downward causation, and relates emergent phenomena to Gödel’s incompleteness theorem. Apparently rich implications are noticed in biology and Chinese philosophy. The perspective of delegated causality supports cognitive interpretations of self-organization and evolution.

Dynamical systems play a central role in explanations in cognitive neuroscience. The grounds for these explanations are hotly debated and generally fall under two approaches: non-mechanistic and mechanistic. In this paper, I first outline a neurodynamical explanatory schema that highlights the role of dynamical systems in cognitive phenomena. I next explore the mechanistic status of such neurodynamical explanations. I argue that these explanations satisfy only some of the constraints on mechanistic explanation and should be considered pseudomechanistic explanations. I defend this (...) argument against three alternative interpretations of the neurodynamical explanatory schema. The independent interpretation holds that neurodynamical explanations and mechanisms are independent. The constitutive interpretation holds that neurodynamical explanations are constitutive but otherwise non-mechanistic. Both the independent and constitutive interpretations fail to account for all the features of neurodynamical explanations. The partial interpretation assumes that the targets of dynamical systems models are mechanisms and so holds that neurodynamical explanations are incomplete because they lack mechanistic details. I contend instead that the targets of those models are dynamical systems distinct from mechanisms and defend this claim against several objections. I conclude with a defense of the pseudomechanistic interpretation and a discussion of the source of their explanatory power in relation to a causal-mechanical description of the world. (shrink)

Systems involving many interacting variables are at the heart of the natural and social sciences. Causal language is pervasive in the analysis of such systems, especially when insight into their behavior is translated into policy decisions. This is exemplified by economics, but to an increasing extent also by biology, due to the advent of sophisticated tools to identify the genetic basis of many diseases. It is argued here that a regularity notion of causality can only be meaningfully defined for (...) systems with linear interactions among their variables. For the vastly more important class of nonlinear systems, no such notion is likely to exist. This thesis is developed with examples of dynamical systems taken mostly from mathematical biology. It is discussed with particular reference to the problem of causal inference in complex genetic systems, systems for which often only statistical characterizations exist. (shrink)

Dynamical systems promise to elucidate a notion of top–down causation without violating the causal closure of physical events. This approach is particularly useful for the problem of mental causation. Since dynamical systems seek out, appropriate, and replace physical substrata needed to continue their structural pattern, the system is autonomous with respect to its components, yet the components constitute closed causal chains. But how can systems have causal power over their substrates, if each component is sufficiently caused (...) by other components? Suppose every causal relation requires background conditions, without which it is insufficient. The dynamical system is structured with a tendency to change background conditions for causal relations anytime needed substrates for the pattern's maintenance are missing; under the changed background conditions, alternative causal relations become sufficient to maintain the pattern. The system controls the background conditions under which one or another causal relation can subserve the system's overall pattern, while the components remain causally closed under their given background conditions. (shrink)

When people make judgments about the effects of a perturbation on populations of species in a food web, their judgments exhibit the dissipation effect: a tendency to judge that effects of the perturbation weaken or dissipate as they spread out through the food web from the locus of the perturbation. In the present research evidence for two more phenomena is reported. Terminal locations are points in the food web with just a single connection to the rest of the web. Judged (...) changes tended to be higher for species at terminal locations than for species the same distance from the perturbation but at nonterminal locations. Branches are points in the web where a route splits into two or more routes. Judged changes tended to be lower for species following branching points than for species the same distance from the perturbation but not following branching points. It is proposed that the findings can be explained as effects of a mental model employing concepts of influence and resistance. Under this model a perturbation is a change in energy level at a point in the system that acts as an influence affecting the rest of the system. The basic concepts in this model are domain‐general and on that basis it is predicted that the dissipation effect should be found in judgments of any physical system to which notions of influence and resistance can be applied. (shrink)

The logic of genetic discovery has changed little over time, but the focus of biology is shifting from simple genotype–phenotype relationships to complex metabolic, physiological, developmental, and behavioral traits. In light of this, the traditional reductionist view of individual genes as privileged difference‐making causes of phenotypes is re‐examined. The scope and nature of genetic effects in complex regulatory systems, in which dynamics are driven by regulatory feedback and hierarchical interactions across levels of organization are considered. This review argues that it (...) is appropriate to treat genes as specific actual difference‐makers for the molecular regulation of gene expression. However, they are often neither stable, proportional, nor specific as causes of the overall dynamic behavior of regulatory networks. Dynamical models, properly formulated and validated, provide the tools to probe cause‐and‐effect relationships in complex biological systems, allowing to go beyond the limitations of genetic reductionism to gain an integrative understanding of the causal processes underlying complex phenotypes. (shrink)

Our approach aims at accounting for causal claims in terms of how the physical states of the underlying dynamical system evolve with time. Causal claims assert connections between two sets of physicals states—their truth depends on whether the two sets in question are genuinely connected by time evolution such that physical states from one set evolve with time into the states of the other set. We demonstrate the virtues of our approach by showing how it is able (...) to account for typical causes, causally relevant factors, being ‘the’ cause, and cases of overdetermination and causation by absences. (shrink)

Efforts to bridge emotion theory with neurobiology can be facilitated by dynamic systems (DS) modeling. DS principles stipulate higher-order wholes emerging from lower-order constituents through bidirectional causal processes cognition relations. I then present a psychological model based on this reconceptualization, identifying trigger, self-amplification, and self-stabilization phases of emotion-appraisal states, leading to consolidating traits. The article goes on to describe neural structures and functions involved in appraisal and emotion, as well as DS mechanisms of integration by which they interact. (...) These mechanisms include nested feedback interactions, global effects of neuromodulation, vertical integration, action-monitoring, and synaptic plasticity, and they are modeled in terms of both functional integration and temporal synchronization. I end by elaborating the psychological model of emotion–appraisal states with reference to neural processes. (shrink)

So-called ‘distinctively mathematical explanations’ (DMEs) are said to explain physical phenomena, not in terms of contingent causal laws, but rather in terms of mathematical necessities that constrain the physical system in question. Lange argues that the existence of four or more equilibrium positions of any double pendulum has a DME. Here we refute both Lange’s claim itself and a strengthened and extended version of the claim that would pertain to any n-tuple pendulum system on the ground that (...) such explanations are actually causal explanations in disguise and their associated modal conditionals are not general enough to explain the said features of such dynamical systems. We argue and show that if circumscribing the antecedent for a necessarily true conditional in such explanations involves making a causal analysis of the problem, then the resulting explanation is not distinctively mathematical or non-causal. Our argument generalises to other dynamical systems that may have purported DMEs analogous to the one proposed by Lange, and even to some other counterfactual accounts of non-causal explanation given by Reutlinger and Rice. (shrink)

Action theory has given rise to some perplexing puzzles in the past half century. The most prominent one can be summarized as follows: What distinguishes intentional from unintentional acts? Thanks to the ingenuity of philosophers and their thought experiments, we know better than to assume that the difference lies in the mere presence of an intention, or in its causal efficacy in generating the action. The intention might be present and may also cause the intended behavior, yet the behavior (...) may not be an intentional action; it may not be an action at all. The classic example is that of the nervous nephew who intends to kill his uncle to inherit his estate, and whose intention makes him so nervous as to drive recklessly, thereby running over a pedestrian... who happens to be his very uncle. The intention to kill is present and it causes the killing, yet the killing is not an intentional action. Rather, what appears to distinguish intentional from non-intentional action is voluntary control of the proper sort, and what distinguishes action from non-action is behavior caused in a particular manner. But spelling out the sort and specifying the manner have proven vexing tasks. (shrink)

Philosophical accounts of the constitution relation have been explicated in terms of synchronic relations between higher‐ and lower‐level entities. Such accounts, I argue, are temporally austere or impoverished, and are consequently unable to make sense of the diachronic and dynamic character of constitution in dynamical systems generally and dynamically extended cognitive processes in particular. In this paper, my target domain is extended cognition based on insights from nonlinear dynamics. Contrariwise to the mainstream literature in both analytical metaphysics and extended (...) cognition, I develop a nonstandard, alternative conception of constitution, which I call “diachronic process constitution”. It will be argued that only a diachronic and dynamical conception of constitution is consistent with the nature of constitution in distributed cognitive processes. (shrink)

In the last 20 years, a stream of research emerged under the label of „complex problem solving“ (CPS). This research was intended to describe the way people deal with complex, dynamic, and intransparent situations. Complex computer-simulated scenarios were as stimulus material in psychological experiments. This line of research lead to subtle insights into the way how people deal with complexity and uncertainty. Besides these knowledge-rich, realistic, intransparent, complex, dynamic scenarios with many variables, a second line of research used (...) more simple, knowledge-lean scenarios with a low number of variables („minimal complex systems“, MCS) that have been proposed recently in problem-solving research for the purpose of educational assessment. In both cases, the idea behind the use of microworlds is to increase validity of problem solving tasks by presenting interactive environments that can be explored and controlled by participants while pursuing certain action goals. The main argument presented here is: both types of systems - CPS and MCS – can only be dealt with successfully if causal dependencies between input and output variables are identified and used for system control. System knowledge is necessary for control and intervention. But CPS and MCS differ in their way of how causal dependencies are identified and how the mental model is constructed; therefore, they cannot be compared directly to each other with respect to the cognitive processes that are necessary for solving the tasks. Knowledge-poor MCS tasks address only a small fraction of the cognitive processes and structures needed for knowledge-rich CPS situations. (shrink)

With the advent of evolutionary developmental research, or EvoDevo, there is hope of discovering the roles that the genetic bases of development play in morphological evolution. Studies in EvoDevo span several levels of organismal organization. Low-level studies identify the ultimate genetic changes responsible for morphological variation and diversity. High-level studies of development focus on how genetic differences affect the dynamics of gene networks and epigenetic interactions to modify morphology. Whereas an increasing number of studies link independent acquisition of homoplastic or (...) convergent morphologies to similar changes in the genomes, homoplasies are not always found to have identical low-level genetic underpinnings. This suggests that a combination of low- and high-level approaches may be useful in understanding the relationship between genetic and morphological variation. Therefore, as an empirical and conceptual framework, we propose the causality horizon to signify the lowest level that allows linking homoplastic morphologies to similar changes in the development. A change in a system below the causality horizon cannot be generalized. In more concrete terms, homoplastic morphologies cannot be reduced to the same change in gene regulation when that change occurs below the causality horizon; rather, a higher-level mechanism should be identified. (shrink)

The concepts and powerful mathematical tools of Dynamical Systems Theory (DST) yield illuminating methods of studying cognitive processes, and are even claimed by some to enable us to bridge the notorious explanatory gap separating mind and matter. This article includes an analysis of some of the conceptual and empirical progress Dynamical Systems Theory is claimed to accomodate. While sympathetic to the dynamicist program in principle, this article will attempt to formulate a series of problems the proponents of the approach in (...) question will need to face if they wish to prolong their optimism. The main points to be addressed involve the reductive tendencies inherent in Dynamical Systems Theory, its somewhat muddled position relative to connectionism, the metaphorical nature DST-C exhibits which hinders its explanatory potential, and DST-C's problematic account of causality. Brief discussions of the mathematical and philosophical backgrounds of DST, seminal experimental work and possible adaptations of the theory or alternative suggestions (dynamicist connectionism, neurophenomenology, R&D theory) are included. (shrink)

The principle of causality has traditionally been formulated in terms of the dynamic equations of classical physics. It has been believed that only stable (conservative) systems can be meaningfully studied in this way. The evolutionary systems studied by social scientists and biologists are non-conservative, and so are excluded from the traditional formulation of causality. The author argues that causal laws for non-conservative systems can be meaningfully formulated. An example based on Malthus'' population theory is given.

Evolutionary systems biology (ESB) aims to integrate methods from systems biology and evolutionary biology to go beyond the current limitations in both fields. This article clarifies some conceptual difficulties of this integration project, and shows how they can be overcome. The main challenge we consider involves the integration of evolutionary biology with developmental dynamics, illustrated with two examples. First, we examine historical tensions between efforts to define general evolutionary principles and articulation of detailed mechanistic explanations of specific traits. Next, these (...) tensions are further clarified by considering a recent case from another field focused on developmental dynamics: stem cell biology. In the stem cell case, incompatible explanatory aims block integration. Experimental approaches aim at mechanistic explanation while dynamical system models offer explanation in terms of general principles. We then discuss an ESB case in which integration succeeds: search for general attractors using a dynamical systems framework synergizes with the experimental search for detailed mechanisms. Contrasts between the positive and negative cases suggest general lessons for achieving an integrated understanding of developmental and evolutionary dynamics. The key integrative move is to acknowledge two complementary aims, both relevant to explanation: identifying the space of possible dynamic states and trajectories, and mechanistic understanding of causal interactions underlying a specific phenomenon of interest. These two aims can support one another in a joint project characterizing dynamic aspects of evolving lineages. This more inclusive project can lead to insights that cannot be reached by either approach in isolation. (shrink)

A standard platitude about the function of causal knowledge or theories is that they are valuable because they support prediction, explanation, and control. Knowledge of predator-prey relations enables us to predict future animal populations, as well as design policies or interventions that help influence those populations. If we understand the underlying biochemical mechanisms of some disease, then we can predict who is at risk for it, explain why it produces particular symptoms, and develop interventions to try to reduce its (...) prevalence or the symptom severity. Of course, there are many situations in which one has, for practical reasons, only some of these desiderata; for example, control might be infeasible for technical or ethical reasons. But these remain, for many researchers, the ideal for why causal knowledge is a valuable end of scientific inquiry, including biological inquiry. There are, however, certain types of systems—in particular, chaotic systems—in which it appears that these ends are unattainable, and these systems appear to be widespread in the biological domain, broadly construed. In this paper, we will show why it is natural to think that causal models of chaotic systems cannot satisfy any of the three functions. But we will also show why this natural thought is wrong: we can have usable causal knowledge about even chaotic systems. Moreover, the ways in which we can have such knowledge lead us naturally to rethink a standard understanding of how causal learning and modeling proceed. In particular, just as we often must find the appropriate variables for a causal system, we also must determine the proper level or granularity of description for the dynamics of that system. (shrink)

There are two well-developed formalizations of discrete time dynamic systems that evidently share many concerns but suffer from a lack of mutual awareness. One formalization is classical systems and automata theory. The other is the logic of actions in which the situation and event calculi are the strongest representatives. Researchers in artificial intelligence are likely to be familiar with the latter but not the former. This is unfortunate, for systems and automata theory have much to offer by way of (...) insight into problems raised in the logics of action. This paper is an outline of how the input-output view of systems and its associated solution of state realization may be applied to the formalization of dynamics that uses a situation calculus approach. In particular, because the latter usually admits incompletely specified dynamics, which induces a non-deterministic input-output system behavior, we first show that classical state realization can still be achieved if the behavior is causal. This is a novel systems-theoretic result. Then we proceed to indicate how situation calculi dynamic specifications can be understood in systems-theoretic terms, and how automata can be viewed as models of such specifications. As techniques for reasoning about automata are abundant, this will provide yet more tools for reasoning about actions. (shrink)

The causal closure of physics is usually discussed in a context free way. Here I discuss it in the context of engineering systems and biology, where strong emergence takes place due to a combination of upwards emergence and downwards causation. Firstly, I show that causal closure is strictly limited in terms of spatial interactions because these are cases that are of necessity strongly interacting with the environment. Effective Spatial Closure holds ceteris parabus, and can be violated by Black (...) Swan Events. Secondly, I show that causal closure in the hierarchy of emergence is a strictly interlevel affair, and in the cases of engineering and biology encompasses all levels from the social level to the particle physics level. However Effective Causal Closure can usefully be defined for a restricted set of levels, and one can experimentally determine Effective Theories that hold at each level. This does not however imply those effective theories are causally complete by themselves. In particular, the particle physics level is not causally complete by itself in the contexts of solid state physics, digital computers, or biology. Furthermore Inextricably Intertwined Levels occur in all these contexts. (shrink)

The notion of cognitive system is widely used in explanations in cognitive psychology and neuroscience. Traditional approaches define cognitive systems in an agent-relative way, that is, via top-down functional decomposition that assumes a cognitive agent as starting point. The extended cognition movement challenged that approach by questioning the primacy of the notion of cognitive agent. In response, [Adams, F., and K. Aizawa. 2001. The Bounds of Cognition. Oxford, UK: Wiley-Blackwell.] suggested that to have a clear understanding of what a (...) cognitive system is we may need to solve “the demarcation challenge”: the problem of identifying a reliable way to determine which mechanisms that are causally responsible for the production of a certain cognitive process constitute a cognitive system responsible for such process and which ones do not. Recently, [Rupert, R. 2009. Cognitive Systems and the Extended Mind. Oxford: Oxford University Press.] offered a solution based on the idea that the mechanisms that constitute a cognitive system are integrated in a particular sense. In this paper I critically review Rupert’s solution and argue against it. Additionally, I argue that a successful account of cognitive system must accommodate the fact that the neural mechanisms causally responsible for the production of a cognitive process are diachronically dynamic and yet functionally stable. At the end, I offer a suggestion as to how to accommodate this diachronic dynamicity without losing functional stability. I conclude by drawing some implications for the discussion on cognitive ontologies. (shrink)

Although we applaud the interactivist approach to language and communication taken in the target article, we notice that Shanker & King (S&K) give little attention to the theoretical frameworks developed by dynamical system theorists. We point out how the dynamical idea of causality, viewed as multidirectional across multiple scales of organization, could further strengthen the position taken in the target article.

This article assesses how people reason with categories whose features are related in causal cycles. Whereas models based on causal graphical models have enjoyed success modeling category-based judgments as well as a number of other cognitive phenomena, CGMs are only able to represent causal structures that are acyclic. A number of new formalisms that allow cycles are introduced and evaluated. Dynamic Bayesian networks represent cycles by unfolding them over time. Chain graphs augment CGMs by allowing the (...) presence of undirected links that model feedback relations between variables. Unfolded chain graphs are chain graphs that unfold over time. An existing model of causal cycles is also evaluated. Four experiments in which subjects reason about categories with cyclically related features provided evidence against DBNs and alpha centrality and for the two types of chain graphs. Chain graphs—a mechanism for representing the equilibrium distribution of a dynamic system—may thus be good candidates for modeling how people reason causally with complex systems. Applications of chain graphs to areas of cognition other than category-based judgments are discussed. (shrink)

A common strategy for simplifying complex systems involves partitioning them into subsystems whose behaviors are roughly independent of one another at shorter timescales. Dynamic causal models clarify how doing so reveals a system’s nonequilibrium causal relationships. Here I use these models to elucidate the idealizations and abstractions involved in representing a system at a timescale. The models reveal that key features of causal representations—such as which variables are exogenous—may vary with the timescale at which (...) a system is considered. This has implications for debates regarding which systems can be represented causally. (shrink)

Mechanistic explanations are often said to explain because they reveal the causal structure of the world. Conversely, dynamical models supposedly lack explanatory power because they do not describe causal structure. The only way for dynamical models to produce causal explanations is via the 3M criterion: the model must be mapped onto a mechanism. This framing of the situation has become the received view around the viability of dynamical explanation. In this paper, I argue against this position and (...) show that dynamical models can themselves reveal causal structure and consequently produce nonmechanistic, dynamical explanations. Taking the example of cell fates from systems biology, I show how dynamical models, and specifically the attractor landscapes they describe, identify the causes of cell differentiation and explain why cells select particular fates. These dynamical features of the system better fit Woodward’s (2010, 2018) criteria of specificity and proportionality and make them the best candidate causes of cell fates than mechanisms. I also show how these causes are irreducible and inaccessible to mechanistic models, making 3M unworkable and counterproductive in this case. Dynamical models can reveal dynamical causes and thereby provide causal explanations. (shrink)

Since the origins of the notion of emergence in attempts to recover the content of vitalistic anti-reductionism without its questionable metaphysics, emergence has been treated in terms of logical properties. This approach was doomed to failure, because logical properties are either sui generis or they are constructions from other logical properties. If the former, they do not explain on their own and are inevitably somewhat arbitrary (the problem with the related concept of supervenience, Collier, 1988a), but if the latter, reducibility (...) is assured because logical constructs are reducible, by definition, to their logical components. A satisfactory account of emergence must recognise that it is a dynamical, not a logical property of property of natural systems, and that its basis is dynamical rather than logical composition. Collier (1988a) introduced the concept of cohesion as the closure of the causal relations among the dynamical parts of a dynamical particular that determine its resistance to external and internal fluctuations that might disrupt its integrity. Cohesion is an equivalence relation that partitions a set of dynamical particulars into unified and distinct entities, providing the identity conditions for such particulars. Cohesion blocks reduction of dynamical particulars, and is necessary for dynamical emergence. We will give reasons for thinking that cohesion might be sufficient for emergence as well. (shrink)

Even if one can experiment on relevant factors, learning the causal structure of a dynamical system can be quite difficult if the relevant measurement processes occur at a much slower sampling rate than the “true” underlying dynamics. This problem is exacerbated if the degree of mismatch is unknown. This paper gives a formal characterization of this learning problem, and then provides two sets of results. First, we prove a set of theorems characterizing how causal structures change under (...) undersampling. Second, we develop an algorithm for inferring aspects of the causal structure at the “true” timescale from the causal structure learned from the undersampled data. Research on causal learning in dynamical contexts has largely ignored the challenges of undersampling, but this paper provides a framework and foundation for learning causal structure from this type of complex time series data. (shrink)

Throughout the Covid-19 pandemic, people have been eager to learn what factors, and especially what public health policies, cause infection rates to wax and wane. But figuring out conclusively what causes what is difficult in complex systems with nonlinear dynamics, such as pandemics. We review some of the challenges that scientists have faced in answering quantitative causal questions during the Covid-19 pandemic, and suggest that these challenges are a reason to augment the moral dimension of conversations about causal (...) inference. We take a lesson from Martha Nussbaum—who cautions us not to think we have just one question on our hands when we have at least two—and apply it to modeling for causal inference in the context of cost-benefit analysis. (shrink)

We show that, given standard assumptions about classical dynamical systems, Lewis' conception of possible worlds is incompatible with classical physics in that it would imply that all dynamical systems were integrable.

A formulation of probabilistic causality is given in terms of the theory of abstract dynamical systems. Causal factors are identified as invariants of motion of a system. Repetition of an experiment leads to the notion of stationarity, and causal factors yield a decomposition of the stationary probability law of the experiment into ergodic components. In these, statistical behaviour is uniform. Control of identified causal factors leads to a corresponding statistical law for the events, which is offered (...) as a notion of probabilistic causality. After a suggestion by Feller, randomization is identified as mixing, formulated in above terms. (shrink)

This dissertation challenges two prevalent views on the topic of scientific explanation: that science explains by revealing causal mechanisms, and that science explains by unifying our knowledge of the world. ;My methodological strategy is to compare our best current philosophical accounts of scientific explanation with evidence from contemporary scientific research. In particular, I focus on evidence from dynamical explanations, that is, explanations which appeal to nonlinear dynamical modeling for their force. Nonlinear dynamical modeling is a type of mathematical modeling (...) which is used by scientists in many different disciplines, including population biology, ecology, physics, and psychology. ;Chapter 1 argues that dynamical explanations are a philosophically important class of explanations, based on their prevalence and their close relationship with the semantic view of theories. ;Chapters 2 and 3 conclude that extant causal and unification accounts encounter serious difficulties when applied to dynamical explanations. Chapter 2 argues that causal relevance is neither a necessary nor a sufficient condition for explanatory relevance. Chapter 3 clarifies and evaluates different claims about the unifying power of science. Here, I argue that dynamical explanations provide some support for the thesis that unification is constitutive of explanation. However, dynamical explanations also indicate that some of the intuitions embodied in extant unification accounts are untenable. ;Chapter 4 concludes by advancing diagnostic criticisms of existing causal and unification accounts. I explore the relationships between dynamical explanations and three conceptions of scientific explanation, as well as the relationships between different types of inference and scientific understanding. I suggest that philosophers look to cognitive science and the semantic view of theories for aid in constructing a viable positive account of explanation. ;The technical appendix introduces several key definitions and concepts of dynamical systems theory. (shrink)

Maxwell's deduction of the probability distribution over the velocity of gas molecules—one of the most important passages in physics (Truesdell)—presents a riddle: a physical discovery of the first importance was made in a single inferential leap without any apparent recourse to empirical evidence. -/- Tychomancy proposes that Maxwell's derivation was not made a priori; rather, he inferred his distribution from non-probabilistic facts about the dynamics of intermolecular collisions. Further, the inference is of the same sort as everyday reasoning about the (...) physical probabilities attached to such canonical chance setups as tossed coins or rolled dice. The structure of this reasoning is investigated and some simple rules for inferring physical probabilities from symmetries and other causally relevant properties of physical systems are proposed. -/- Not only physics but evolutionary biology and population ecology, the science of measurement error, and climate modeling have benefited enormously from the same kind of reasoning, the book goes on to argue. Inferences from dynamics to probability are so obvious to us, however, that their methodological importance has been largely overlooked. (shrink)

Many of our decisions pertain to causal systems. Nevertheless, only recently has it been claimed that people use causal models when making judgments, decisions and predictions, and that causal Bayes nets allow us to formally describe these inferences. Experimental research has been limited to simple, artificial problems, which are unrepresentative of the complex dynamic systems we successfully deal with in everyday life. For instance, in social interactions, we can explain the actions of other's on the fly (...) and we can generalize from limited observations to predict future actions and their consequences. Our main argument is that none of these inferences (i.e., induction, generalization, explanation, and prediction) can be achieved without causal reasoning. As a case in point we use the popular television series desperate housewives and show how causal Bayes nets are able to explain the inferences made in social contexts. Crucially, causal Bayes nets also allow us to understand why we can infer so much from so little when making sense of a protagonist's behavior. (shrink)

This chapter questions the causal-constitution fallacy raised against the extended mind. It does so by presenting our signature temporal thesis about how to understand constitutive relations in the context of the extended mind, and with respect to dynamical systems, more broadly. We call this thesis diachronic constitution. We will argue that temporalising the constitution relation is not as remarkable (nor problematic) as it might initially seem. It is (arguably) inevitable, given local interactions between microscale and macroscale states of (coupled) (...) dynamical systems. We focus primarily on the metaphysics of the extended mind in this paper. However, we also show how our account of diachronic constitution has important implications for the metaphysics of dependence relations more generally as well as an emerging literature on inter-level explanations in the mechanistic framework applied to the discussion over extended, enactive and embodied cognition. (shrink)

The idea that intentions make the difference between voluntary and non-voluntary behaviors is simple and intuitive. At the same time, we lack an understanding of how voluntary actions actually come about, and the unquestioned appeal to intentions as discrete causes of actions offers little if anything in the way of an answer. We cite evidence suggesting that the origin of actions varies depending on context and effector, and argue that actions emerge from a causal web in the brain, rather (...) than a central origin of intentional action. We argue that this causal web need not be confined to the central nervous system, and that proprioceptive feedback might play a counterintuitive role in the decision process. Finally we argue that the complex and dynamic origins of voluntary action and their interplay with the brain’s propensity to predict the immediate future are better studied using a dynamical systems approach. (shrink)

The relationship between machine learning and the philosophy of science can be classed as a dynamic interaction: a mutually beneficial connection between two autonomous fields that changes direction over time. I discuss the nature of this interaction and give a case study highlighting interactions between research on Bayesian networks in machine learning and research on causality and probability in the philosophy of science.

Many approaches to evidence amalgamation focus on relatively static information or evidence: the data to be amalgamated involve different variables, contexts, or experiments, but not measurements over extended periods of time. However, much of scientific inquiry focuses on dynamical systems; the system’s behavior over time is critical. Moreover, novel problems of evidence amalgamation arise in these contexts. First, data can be collected at different measurement timescales, where potentially none of them correspond to the underlying system’s causal timescale. (...) Second, missing variables have a significantly different impact on time series measurements than they do in the traditional static setting; in particular, they make causal and structural inference much more difficult. In this paper, we argue that amalgamation should proceed by integrating causal knowledge, rather than at the level of “raw” evidence. We defend this claim by first outlining both of these problems, and then showing that they can be solved only if we operate on causal structures. We therefore must use causal discovery methods that are reliable given these problems. Such methods do exist, but their successful application requires careful consideration of the problems that we highlight. (shrink)

This contribution discusses competencies needed for regulating systems with properties of multi-causality and non-linear dynamics (therapeutic, economical, organizational, socio-political, technical, ecological, etc.). Various research communities have contributed insights, but none has come forward with an inclusive framework. To advance the debate, I propose to draw from dynamic systems theory (DST) and “4E” (embodied, embedded, enactive, and extended), cognition approaches, which offer a set of perspectives to understand what expert regulators in real-life settings do. They define the regulator's agency as (...) skillfully imposing constraints on a target system and hereby creating context-sensitive openings for self-organizing dynamics, rather than “controlling” the system. Adept regulators apply multi-pronged and multi-timescale constraints to achieve nuanced effects. Among other things, their skill set includes scarcely noted enactive processual competencies for “emergence management”, which the intellectualistic and insufficiently ecologically situated accounts of the complex problem solving literature omit. To capture the nature of system regulation, I advocate treating regulation dynamics and target system dynamics “symmetrically” by grounding regulator competencies in concepts from complexity theory. (shrink)

Causal knowledge unquestionably provides useful means to describe, explain and predict both natural and daily phenomena. This article addresses whether causality as such may not be ontologically primary and looks for an alternative fundamental mechanism encapsulating the information load of the causal framework. A probabilistic process view of reality asserting the struggle of natural forces is considered along with lines quoted from Nietzsche’s posthumously published notebooks and published work. Examples from scientific discoveries, in particular neurosciences, echoing his ontology (...) are provided. Furthermore, I propose a basic number game algorithm to illustrate and model the struggle of forces leading to causal inferences for an external observer. Nietzschean dynamic ontology involving the ceaseless struggle of will-to-power units accords with the emergentist approach to causation, which has been principally employed by the scientists working under the fields of thermodynamics, quantum mechanics and complex dynamical systems. (shrink)

Parallel lives is an ontological model of nature in which quantum mechanics and special relativity are unified in a single universe with a single space-time. Point-like objects called lives are the only fundamental objects in this space-time, and they propagate at or below c, and interact with one another only locally at point-like events in space-time, very much like classical point particles. Lives are not alive in any sense, nor do they possess consciousness or any agency to make decisions—they are (...) simply point objects which encode memory at events in space-time. The only causes and effects in the universe occur when lives meet locally, and thus the causal structure of interaction events in space-time is Lorentz invariant. Each life traces a continuous world-line through space-time, and experiences its own relative world, fully defined by the outcomes of past events along its world-line, which are encoded in its external memory. A quantum field comprises a continuum of lives throughout space-time, and familiar physical systems like particles each comprise a sub-continuum of the lives of the field. Each life carries a hidden internal memory containing a local relative wavefunction, which is a local piece of a pure universal wavefunction, but it is the relative wavefunctions in the local memories throughout space-time which are physically real in PL, and not the universal wavefunction in configuration space. Furthermore, while the universal wavefunction tracks the average behavior of the lives of a system, it fails to track their individual dynamics and trajectories. There is always a preferred separable basis, and for an irreducible physical system, each orthogonal term in this basis is a different relative world—each containing some fraction of the lives of the system. The relative wavefunctions in the lives’ internal memories govern which lives of different systems can meet during future local interactions, and thereby enforce entanglement correlations—including Bell inequality violations. These, and many other details, are explored here, but several aspects of this framework are not yet fleshed out, and work is ongoing. (shrink)

The essence of the path integral method in quantum physics can be expressed in terms of two relations between unitary propagators, describing perturbations of the underlying system. They inherit the causal structure of the theory and its invariance properties under variations of the action. These relations determine a dynamical algebra of bounded operators which encodes all properties of the corresponding quantum theory. This novel approach is applied to non-relativistic particles, where quantum mechanics emerges from it. The method works (...) also in interacting quantum field theories and sheds new light on the foundations of quantum physics. (shrink)

We consider the question of whether thermodynamic macrostates are objective consequences of dynamics, or subjective reflections of our ignorance of a physical system. We argue that they are both; more specifically, that the set of macrostates forms the unique maximal partition of phase space which (1) is consistent with our observations (a subjective fact about our ability to observe the system) and (2) obeys a Markov process (an objective fact about the system’s dynamics). We review the ideas (...) of computational mechanics, an information-theoretic method for finding optimal causal models of stochastic processes, and argue that macrostates coincide with the “causal states” of computational mechanics. Defining a set of macrostates thus consists of an inductive process where we start with a given set of observables, and then refine our partition of phase space until we reach a set of states which predict their own future, i.e. which are Markovian. Macrostates arrived at in this way are provably optimal statistical predictors of the future values of our observables. (shrink)

Alicia Juarrero's Dynamics in action is an attempt to bring dynamical systems theory to bear on action theory. Juarrero claims that the vocabulary and examples from dynamical systems theory can help us resolve some pressing problems in action theory, including the problems of extended action and disrupted causal chains. In the process, Juarrero calls for a return to an Aristotelean plurality of causes and a serious scrutiny of modern Humean/positivist science as she sees it. Although her book has some (...) inaccuracies and fails to convince the skeptical reader, it provides an alternative point of view from which to launch further studies. (shrink)