Announcer
The following program features simulated voices generated for educational and philosophical exploration.
Leonard Jones
Good afternoon. I'm Leonard Jones.
Jessica Moss
And I'm Jessica Moss. Welcome to Simulectics Radio.
Leonard Jones
Yesterday we examined our obligations to future generations and the temporal boundaries of moral concern. Today we turn to a fundamental metaphysical question: what is causation? When we say one event causes another, what relation are we describing? Is causation a fundamental feature of reality, or merely a pattern we project onto regularities in experience?
Jessica Moss
This isn't just abstract metaphysics. Science depends on identifying causal relationships—what causes disease, what causes climate change, what causes economic recessions. What are the stakes here? If we misunderstand causation, we misunderstand how to intervene in the world.
Leonard Jones
Our guest is Dr. Nancy Cartwright, Professor of Philosophy at Durham University, whose work on causal powers, scientific explanation, and the relationship between fundamental laws and actual causes has challenged conventional views in philosophy of science. We spoke with Professor Cartwright previously about scientific realism. Welcome back.
Dr. Nancy Cartwright
Thank you. It's good to return to these questions about how we understand causal structure in the world.
Jessica Moss
Let's start with David Hume's skeptical challenge. Hume argued that when we observe what we call causation—one billiard ball striking another—we see only succession and contiguity, not any necessary connection between the events. Is causation real, or just constant conjunction?
Dr. Nancy Cartwright
Hume's challenge is profound, but I think his conclusion is too skeptical. We can agree that we don't observe necessity in individual cases—we see this event followed by that event. But causation isn't reducible to mere regularity. Causal relations involve capacities or powers that things have to bring about effects under appropriate circumstances.
Leonard Jones
Let me be precise about this. Are you claiming that objects possess causal powers as intrinsic properties, independent of regularities? That aspirin has the power to relieve headaches even when it's sitting in the bottle?
Dr. Nancy Cartwright
Yes, though we need to be careful about how we understand this. Causal powers are dispositions—they're real properties of things that manifest under appropriate conditions. Aspirin has certain chemical properties that, when ingested by humans with particular physiological states, can relieve headaches. The power is there even when unmanifested.
Jessica Moss
But how do we know what powers things have except by observing regularities? This seems to bring us back to Hume—we infer powers from patterns, but the patterns are all we observe.
Dr. Nancy Cartwright
We do learn about powers through observation and experimentation, but that doesn't make powers reducible to regularities. Consider a carefully controlled experiment. We isolate a system, manipulate it, and observe the result. What we're doing is testing for causal capacities by creating conditions where interfering factors are removed. The experiment reveals pre-existing powers, not mere patterns.
Leonard Jones
This raises questions about scientific laws. The traditional view treats laws as universal generalizations—all Fs are Gs. But you've argued that fundamental laws lie, that they describe idealized situations rather than actual causal processes. How does this relate to causal powers?
Dr. Nancy Cartwright
Fundamental laws in physics describe what happens in highly idealized circumstances—frictionless planes, point masses, isolated systems. These idealizations help us understand causal capacities in pure form. But in the messy actual world, multiple capacities operate simultaneously, and the result isn't determined by any single law but by the combination of different causal factors.
Jessica Moss
So you're saying that what we call scientific laws are really descriptions of what things would do in ideal conditions, but actual causation involves multiple causal powers operating together?
Dr. Nancy Cartwright
Exactly. Think about a feather falling. Fundamental physics says gravitational force equals mass times gravitational acceleration. But the feather's actual trajectory depends on gravity plus air resistance plus wind currents plus static electricity. No single law describes the trajectory—we need to understand how different causal capacities combine.
Leonard Jones
This seems to complicate scientific explanation. If laws don't describe actual processes, how do they explain anything?
Dr. Nancy Cartwright
Laws help us understand causal capacities, not actual sequences of events. Explanation in science often involves causal stories about how different capacities produced the phenomenon we're trying to explain. We use laws as tools to think about what would happen if certain capacities operated in isolation, then combine these insights to understand complex situations.
Jessica Moss
Let's talk about counterfactual analyses of causation. Some philosophers say that C causes E if and only if, had C not occurred, E would not have occurred. Does this capture causation?
Dr. Nancy Cartwright
Counterfactual dependence is related to causation, but I don't think it's the fundamental notion. There are cases of causation without counterfactual dependence—overdetermination, where two sufficient causes operate, so even if one hadn't occurred, the other would have produced the effect. And there are counterfactual dependencies that aren't causal—common causes create correlations without direct causation.
Leonard Jones
Can you give an example of the common cause problem?
Dr. Nancy Cartwright
A barometer falling is counterfactually dependent on a storm approaching—if the storm weren't approaching, the barometer wouldn't fall. But the falling barometer doesn't cause the storm. Both are effects of atmospheric pressure changes. Counterfactual dependence alone doesn't distinguish genuine causes from correlated effects of common causes.
Jessica Moss
What about probabilistic causation? In medicine or social science, we often talk about factors that increase the probability of outcomes rather than determining them. Does this require a different account of causation?
Dr. Nancy Cartwright
Probabilistic causation is perfectly compatible with causal powers. Many causal capacities are chancy—they raise the probability of effects without guaranteeing them. Smoking causes cancer not in the sense that every smoker gets cancer, but in the sense that smoking has a causal power that raises cancer risk. The power itself is real even though its manifestation is probabilistic.
Leonard Jones
Let me try to get clear on the relationship between causal powers and probability. Is the probabilistic nature fundamental to the power itself, or does it arise from our ignorance of underlying deterministic factors?
Dr. Nancy Cartwright
This is an important question I'm genuinely uncertain about. It might be that some causal powers are irreducibly chancy—quantum mechanics suggests this. Or it might be that what appears probabilistic reflects complex interactions of multiple deterministic capacities we can't track in detail. For practical purposes of understanding causation, we can remain neutral on this metaphysical question.
Jessica Moss
How do we identify causal relationships in practice, especially in observational studies where we can't manipulate variables?
Dr. Nancy Cartwright
This is where causal inference methodology becomes crucial. Randomized controlled trials are the gold standard because randomization breaks correlations with confounding factors. When we can't randomize, we use techniques like instrumental variables, regression discontinuity, or careful causal modeling to try to isolate causal effects from mere correlations.
Leonard Jones
But doesn't this presuppose we already know the causal structure—what counts as a confounder, what the relevant variables are?
Dr. Nancy Cartwright
Yes, causal inference requires substantial background knowledge. We need theoretical understanding of the domain to know what factors might be relevant, what causal pathways are possible, what confounders to control for. Statistical techniques alone can't discover causal structure from scratch—they help us test causal hypotheses informed by substantive theory.
Jessica Moss
This seems to create a circle. We need causal knowledge to design good studies, but we conduct studies to gain causal knowledge.
Dr. Nancy Cartwright
It's a circle, but not a vicious one. We build causal knowledge incrementally, using theoretical understanding to design studies, using study results to refine theory, iteratively improving our grasp of causal structure. Science doesn't start from pure observation—it starts from crude causal hypotheses that get refined through investigation.
Leonard Jones
How does your account of causal powers relate to the debate between reductionism and emergence? Are causal powers at higher levels of organization—biological, psychological, social—reducible to physical causal powers?
Dr. Nancy Cartwright
I'm skeptical of the assumption that all causal powers reduce to fundamental physics. Different levels of organization may have their own causal structures that aren't simply the sum of lower-level capacities. In complex systems, organization itself creates new causal possibilities. I prefer what I call a patchwork metaphysics—different domains with their own causal structures.
Jessica Moss
Can you give an example of irreducible higher-level causation?
Dr. Nancy Cartwright
Consider institutional facts—the fact that this piece of paper is money, that this building is a university, that this person is a citizen. These institutional structures have causal powers—money can purchase things, universities can grant degrees, citizenship confers rights. These causal powers depend on social organization, not just the physical properties of the relevant objects.
Leonard Jones
But surely institutional facts supervene on physical facts. The causal powers of money ultimately reduce to physical interactions—people's behaviors, neural states, and so on.
Dr. Nancy Cartwright
Supervenience doesn't imply reducibility. Institutional facts may supervene on physical facts in the sense that you can't change institutional facts without changing something physical. But the causal explanations appropriate for institutional phenomena involve institutional concepts, not physical ones. Explaining why a bank collapsed requires economic concepts, not particle physics.
Jessica Moss
What are the practical stakes of getting causation right? Why should non-philosophers care about these metaphysical debates?
Dr. Nancy Cartwright
Policy depends on causal knowledge. When we ask whether a drug works, whether a social program reduces poverty, whether carbon emissions cause climate change, we're asking causal questions. Misunderstanding causation leads to bad policy—confusing correlation with causation, ignoring confounders, applying causal knowledge from one context to another where it doesn't hold.
Leonard Jones
Can you elaborate on the context-sensitivity of causal knowledge? Why can't we simply transfer causal claims from one domain to another?
Dr. Nancy Cartwright
Causal powers manifest differently in different circumstances. A drug effective in clinical trials might not work in real-world populations with different characteristics. A social intervention successful in one community might fail in another with different social structures. We can't assume causal capacities produce the same effects across all contexts—we need to understand the support factors that enable capacities to manifest.
Jessica Moss
This seems to create problems for evidence-based policy. If causal knowledge is so context-dependent, how can we use evidence from randomized trials to guide intervention elsewhere?
Dr. Nancy Cartwright
Carefully and with substantial theoretical understanding. Randomized trials tell us that an intervention has certain causal capacities in the trial context. To apply this knowledge elsewhere, we need theory about what support factors were present in the trial, whether they're present in the target context, what other factors might interfere. Evidence doesn't directly translate—it requires interpretation through causal understanding.
Leonard Jones
Let's return to fundamental questions. Are causal relations part of the fundamental structure of reality, or are they features of our representations—ways we organize experience for practical purposes?
Dr. Nancy Cartwright
I think causal structure is real. Things genuinely have capacities to affect other things. This isn't merely how we organize experience—the world has causal joints that our concepts track. But I also think there's no single complete causal story—different levels of description capture different aspects of causal structure, none exhaustively fundamental.
Jessica Moss
Some philosophers argue that causation is purely anthropocentric—it reflects our perspective as agents intervening in the world, not a mind-independent feature of reality.
Dr. Nancy Cartwright
Our interest in causation certainly reflects our practical concerns as agents. We care about what would happen if we intervened because we want to change the world. But this doesn't make causation unreal. The world's causal structure determines what happens when we intervene. Agency requires causation, but doesn't create it.
Leonard Jones
How does quantum mechanics challenge our understanding of causation? Does fundamental physics support the view that causation involves powers or capacities?
Dr. Nancy Cartwright
Quantum mechanics is interesting because it seems to involve irreducibly probabilistic causation—the decay of a radioactive atom, the collapse of a wave function. These look like causal processes where the system has a capacity that manifests probabilistically rather than deterministically. This fits naturally with a powers view where capacities can be chancy.
Jessica Moss
What about quantum entanglement and non-locality? These seem to involve correlations without local causal connections.
Dr. Nancy Cartwright
Entanglement does complicate our understanding of causation. There are correlations between measurements on entangled particles that can't be explained by local causal processes. But these correlations can't be used to send signals—they don't involve controllable causation. This might show that causal structure is richer than we thought, not that causation is unreal.
Leonard Jones
As we approach the end of our time, what's the current state of philosophical understanding of causation? Are we making progress, or are these perennial debates?
Dr. Nancy Cartwright
I think we're making progress, though not toward a simple unified theory. We better understand the relationships between different aspects of causation—powers, regularities, counterfactuals, manipulation, probability. We recognize that causation is a complex phenomenon that different analyses capture different facets of. The challenge is integrating these insights without forcing them into a single framework.
Jessica Moss
Does this pluralism about causation create problems for scientific practice, or does it reflect the complexity of how science actually works?
Dr. Nancy Cartwright
I think it reflects how science works. Different scientific contexts emphasize different aspects of causation. Fundamental physics focuses on capacities described by laws. Epidemiology emphasizes probabilistic causation and intervention. Economics uses counterfactual analysis. Each approach is valuable for its domain. The mistake is thinking one must be the fundamental account.
Leonard Jones
Professor Cartwright, thank you for this exploration of causation and causal powers.
Dr. Nancy Cartwright
Thank you. These questions about causal structure remain central to both metaphysics and scientific practice.
Jessica Moss
We'll return tomorrow with more philosophical inquiry.
Leonard Jones
Good afternoon.