You don’t like the idea of forced change, of intervention, being so integral to such a seemingly basic notion as causality.
You may not, but indeed, according to Judea Pearl, interventions are integral to the idea of causation. Eliezer’s post is not reliable on this point. A causal model and a Bayes net are not the same thing, a point Ilya Shpitser makes in the comments to that post.
The data don’t tell you what causal graph to draw. You choose a hypothetical causal graph, not based on the data. You can then see whether the observational data are consistent with the causal graph you have chosen. However, the data may be consistent with many causal graphs.
What is an intervention? To call something an intervention still involves causal assumptions: an intervention is something we can do to set some variable to any chosen value, in a way that we have reason to believe screens off that variable from all other influences on it, and does not have a causal influence on any other variable except via the one we are setting. That “reason to believe” does not come from the data (which we have yet to collect), but from previous knowledge. For example, the purpose of randomisation in controlled trials is to ensure the independence of the value we set the variable to from everything else.
>You don’t like the idea of forced change, of intervention, being so integral to such a seemingly basic notion as causality. It feels almost anthropomorphic: you want the notion of cause and effect within a system to make sense without reference to the intervention of some outside agent—for there’s nothing outside of the universe.
RK:
>You may not, but indeed, according to Judea Pearl, interventions are integral to the idea of causation.
Indeed, from the Epilogue of the second edition of Judea Pearl’s book Causality (emphasis added):
The equations of physics are indeed symmetrical, but when we compare the phrases “A causes B” versus “B causes A,” we are not talking about a single set of equations. Rather, we are comparing two world models, represented by two different sets of equations: one in which the equation for A is surgically removed; the other where the equation for B is removed. Russell would probably stop us at this point and ask: “How can you talk about two world models when in fact there is only one world model, given by all the equations of physics put together?” The answer is: yes. If you wish to include the entire universe in the model, causality disappears because interventions disappear – the manipulator and the manipulated lose their distinction. However, scientists rarely consider the entirety of the universe as an object of investigation. In most cases the scientist carves a piece from the universe and proclaims that piece in – namely, the focus of investigation. The rest of the universe is then considered out or background and is summarized by what we call boundary conditions. This choice of ins and outs creates asymmetry in the way we look at things, and it is this asymmetry that permits us to talk about “outside intervention” and hence about causality and cause-effect directionality.
You may not, but indeed, according to Judea Pearl, interventions are integral to the idea of causation. Eliezer’s post is not reliable on this point. A causal model and a Bayes net are not the same thing, a point Ilya Shpitser makes in the comments to that post.
The data don’t tell you what causal graph to draw. You choose a hypothetical causal graph, not based on the data. You can then see whether the observational data are consistent with the causal graph you have chosen. However, the data may be consistent with many causal graphs.
What is an intervention? To call something an intervention still involves causal assumptions: an intervention is something we can do to set some variable to any chosen value, in a way that we have reason to believe screens off that variable from all other influences on it, and does not have a causal influence on any other variable except via the one we are setting. That “reason to believe” does not come from the data (which we have yet to collect), but from previous knowledge. For example, the purpose of randomisation in controlled trials is to ensure the independence of the value we set the variable to from everything else.
OP:
>You don’t like the idea of forced change, of intervention, being so integral to such a seemingly basic notion as causality. It feels almost anthropomorphic: you want the notion of cause and effect within a system to make sense without reference to the intervention of some outside agent—for there’s nothing outside of the universe.
RK:
>You may not, but indeed, according to Judea Pearl, interventions are integral to the idea of causation.
Indeed, from the Epilogue of the second edition of Judea Pearl’s book Causality (emphasis added):