Home > Determinism, Philosophy, Philosophy of Mind, Philosophy of Science > Nonreductive Agent Causation Part I: A Dialogue Between O’Connor and Pereboom

Nonreductive Agent Causation Part I: A Dialogue Between O’Connor and Pereboom

February 19, 2012 Leave a comment Go to comments

"To reduce, or not to reduce - that is the question."

I have recently come to believe that the crux of disagreements in contemporary discussions on physicalism and agency is the seemingly impassable divide between reductionist and non-reductionist positions. Perhaps one of the clearest examples of this disconnect can be seen in a dialogue between Derk Pereboom and Timothy O’Connor regarding the plausibility of a certain type of physicalist agency theory. The conversation is multi-faceted and invokes emergent agent causal powers (which I have mentioned here before, though only in passing) as well as quantum indeterminism. In this post I would like to introduce the reduction/non-reduction divide by unfolding the conversation between Pereboom and O’Connor. Part I will be heavily exegetical, but in Part II I offer up four points of analysis on the dialogue at large and the theories therein.

Agent Causation and Emergent Macroproperties

In Agent Causation, O’Connor seeks to give a robust libertarian account of agent causation as distinct from event causation. He insists that, contrary to traditional objections, this theory is coherent and consistent with what is known currently about the natural world. Of this he says:

I will contend that the commonsense view of ourselves as fundamental causal agents – for which some have used the term “unmoved movers” but which I think might more accurately be expressed as “not wholly moved movers” – is theoretically understandable, internally consistent, and consistent with what we have thus far come to know about the nature and workings of the natural world.[1]

Before treating the emergence of agent causality from event causes, O’Connor iterates the fundamental tenet of the agency theory he espouses, viz. that (1) there are two types of causal properties, and (2) that one such property “applies uniquely to intelligent, purposive agents.”[2] By fleshing out the agent, O’Connor seeks to avoid the pitfalls of substance dualism. In this effort O’Connor qualifies a human agent as, “a wholly biological organism, whose macroproperties are either constituted by or dependent on the properties of certain elementary physical particles, organized into complex subsystems at a number of levels.”[3] Thus he presents the human agent as composed of purely natural substances (biological), with macroproperties that are dependent upon lower level arrangements of particles and structures[4]. In this way he avoids substance dualism and apparently fulfills his original criteria of naming the agent as consistent with what is known about nature.

Prima facie, this agent and his macroproperties would still appear to be determined by “the behavior of microelements” and systems of microelements. So too it would seem that, while the organization O’Connor describes would allow for top-down agent causation as a macroproperty of the biological agent, this causation would only exist because it was caused by these microstructures to begin with.[5] O’Connor’s response to this objection is that it is quite plausible that such an organization of matter could potentially give rise to an emergent macroproperty that can “exert (in certain circumstances) an irreducibly ‘downward’ form of causal influence, or ones that enable the objects that bear them to do so ‘at will’.”[6] O’Connor does not fully describe this emergent property, but he does comment that, “Suffice it to say that an emergent property is a macroproperty that is generated by the properties of an object’s microstructure, but whose role in the causal processes involving that object are not reducible to those of the microproperties.”[7] In effect this emergent macroproperty, specifically agent causation, is derived from microstructures but cannot be reduced down to them. Once this macroproperty emerges it can then exercise top-down causation, and actions following this top-down causation cannot be said to be determined by the microstructures that gave rise to the macroproperty itself.[8] When applied, this macroproperty enables “its possessor directly to effect changes at will.”[9] As we will now see, Derk Pereboom interprets this as an incoherent sense of the macroproperty deriving from but sitting outside of the laws of physics or quantum physics.

Objections to Agent Causal Libertarianism

Pereboom is skeptical that the nonreductive materialist approach towards agent causation, as championed by O’Connor above, can provide libertarians with free decisions that are replete with moral responsibility. His first order of business is to demonstrate this to the libertarians. Even if we posit that nonreductive materialism is true, and that all microsystems are wholly governed by physical laws, this does not provide agent-caused free decisions. As Pereboom says, “as long as the microphysical level is governed by deterministic laws, all of our decisions will be rendered inevitable by virtue of previous states of the universe, just as their microphysical realizations are.”[10] These decisions, then, would be determined by factors beyond the control of the agent, and therefore no decision would constitute an agent-caused free decision. Though adroitly aimed, Pereboom’s first salvo does not knock O’Connor out of the race. Recall that O’Connor deftly rejects this maneuver in his own article, insisting that even though these macroproperties would emerge from determined microsystems, having emerged they would be free from determination.[11]
Knowing this, Pereboom constructs an argument that highlights the empirical implausibility and alleged incoherence of agent causal libertarianism. He begins by asserting that, even if the microsystems are governed by statistical quantum mechanics rather than determined physics, it would still be the case that all microsystems would be determined in the same way: “If everything is wholly constituted of microphysical entities governed by such laws of quantum mechanics, then all of our decisions will be wholly constituted of events on the continuum we discussed earlier [regarding a deterministic system].”[12] In this way, because the microsystems would be governed completely by laws of quantum mechanics, so too would our decisions be overwhelmingly governed by factors outside of the agent. In this way, although it would not be purely determined, statistically the outcomes would seem to be outside of the control of the agent. This quantum picture does not seem to offer agent-caused free decisions either, and emergentism similar to O’Connor’s appears to be one of the few options remaining. [13]

If O’Connor’s emergentism is correct then, despite determined microsystems, the macroproperties would be free from the governance of statistical quantum laws or deterministic laws. If this is this case then agents would need to have the distinct ability to ignore or trump these laws as they will, either in the predictive capacity of quantum mechanics or in the determined capacity of the laws of physics. Of this Pereboom says that, “if agent-causes are to be capable of such free decisions, they would require the power to produce deviations from the physical laws – deviations from what these laws would predict and from what we would expect given these laws.”[14] Over the period of time after these macroproperties emerged, we would expect to see at least a slight pattern of divergence from the outcomes predicted by the physical laws. Pereboom offers the following analogy:

Another way of seeing this is that if we were agents making transcendentally free choices, one would expect, in the long run, that these choices be evident in our bodies as patterns of divergence from the deterministic physical laws. Kant’s proposal that there are no such divergences, although it involves no logical contradiction, would run so sharply counter to what we would expect to occur as to render the proposal incredible.[15]

Here lies the rub for emergent agent causation. If, as O’Connor argues, agent causal power is an emergent macroproperty free from determination or prediction by the laws of physics or quantum mechanics, then some divergence must necessarily be present. But this does not afford with our observations of the laws of nature, since no divergence is apparent. As Pereboom points out above, Kant proposes that there are no divergences and that “every transcendentally free choice ever made dovetails precisely with the way the physical components of actions are causally determined to be.”[16] Though not logically impossible, this alignment would seem incredibly implausible in either a deterministic or a statistical view of the universal laws, and as Pereboom says, “the wild coincidences implied by this proposal make it incredible.”[17] Given this attack, Pereboom shows that agent causal libertarianism derived from emergentism does not accomplish what it sets out to accomplish.

O’Connor’s Rebuttal

In a review of Living Without Free Will, O’Connor responds to Pereboom’s assault on agent causal accounts of free action. Prior to outlining the two weaknesses he sees in Pereboom’s argument, O’Connor gives a brief iteration of Pereboom’s argument. This 5 point summation is reproduced below to both clarify the argument up to this point and set up the weaknesses O’Connor perceives:

(1) the statistical laws of quantum mechanics extend to all complex physical systems, including human organisms, and give complete explanations for the behavior of all such systems.

(2) Hence the physical aspects of all human actions should individually fall within the range of permitted possibilities and collectively converge upon the frequencies of outcome types predicted by these laws [from (1)].

(3) However, if human organisms are agent causes, there is a causal factor that is independent of all statistical physical laws.

(4) So the pattern of outcomes of agent-caused actions would probably diverge, in the long run, from the frequencies predicted by quantum mechanical laws [from (3)].

(5) Therefore human organisms are not agent causes.[18]

This appears to be a faithful account of Pereboom’s argument as presented in LWFW. The first weakness that O’Connor tackles is the truth of point (3) and its implication, point (4).

O’Connor’s first tactic is to play off of the role of indeterminacy in a quantum mechanical view of the universe. He cites (3) as not only vague but questionable as well. Instead of the view presented by Pereboom, he asks us to “Imagine that some conscious reasons-guided systems ‘magnify’ microphysical indeterminacies in such a way that several significantly different outcomes are physically possible. Then further suppose that agent-causal power emerges when conscious reasons-guided systems achieve a requisite threshold of complexity.”[19] This agent-causal power would then be “shaped” by states that reflected these magnified indeterminacies, resulting in the agent-causal action physically correlating with the possible outcomes. This would mean that, contrary to (4), no divergence would be necessary. On this, O’Connor says, “Agent-causal theorists have not typically thought of agent causation as being shaped by determinate probabilities that ultimately stem from impersonal factors, but I do not see why they cannot accept this.”[20] So while this move is a departure from the norm of agent-causal libertarians, O’Connor does not see it as a particularly contentious shift.

O’Connor’s second critique of Pereboom’s argument calls into question the supporting premise (1) of the entire argument, namely that it is wrong to assume that just because laws have a specific effect on microsystems they must have identical effects on systems of increasing complexity. That is to say, O’Connor takes offense at Pereboom’s utilization of:

a reductionist article of faith, blithe acceptance of which is at odds with the cautious skepticism regarding over-generalization that is characteristic of mature physical science. Well established physical theory leaves it an open question whether human mentality and agency involve ontologically emergent capacities that interplay with fundamental physical forces.[21]

For example, simply because the laws of quantum mechanics completely explain microsystems and the relationships between particles and other lower-level elements does not automatically mean that an emergent macroproperty (a far more complex system) must also be explainable by such laws. This contention appears to center chiefly around a reductionist-nonreductionist stalemate. Is O’Connor’s claim that hard line reductionism is not only contrary to a scientific attitude but also an unexamined and arbitrarily entrenched position? In part II of this post I offer four points of analysis on non/reduction as well as on emergent agent causal theory and quantum indeterminism.


[1] Timothy O’Connor, “Agent Causation,” , 258.

[2] O’Connor, 262. O’Connor derives the thesis of two types of causation from a prior thesis that, “objects have causal powers in virtue of their properties, so that objects sharing the same properties share the same causal capacities,” (pg 261). This thesis in turn derives from an assumption O’Connor makes that causality cannot be reduced along Humean lines (pg 259).

[3] Ibid., 262.

[4] From this point on these lower level arrangements will be referred to as microelements.

[5] O’Connor presents this objection in Agent Causation as embodied by John Searle’s critique, pg. 262.

[6] O’Connor, 263.

[7] Ibid.

[8] Since this is, in effect, what makes it an emergent macroproperty.

[9] O’Connor, 264.

[10] Derk Pereboom, Living Without Free Will, 70.

[11] I have elsewhere argued that supervenient relationships of the type employed by O’Connor to protect emergent agents from upward causation, and therefore determination, result in over determination by way of  a slight modification of Jaegwon Kim’s exclusion argument. Though I cannot go into detail here, this holds due to the downward causal influence the emergent macroproperty exerts upon its microproperties.

[12] Pereboom, 71.

[13] Ibid., 71-3. “To simplify, the causal history of all of the constituents of any of our decisions will be exhausted by the contribution made by factors beyond the agent’s control, and nothing else. But if this is so, then the causal history of the decision itself will also be exhausted by the contribution made by factors beyond the agent’s control, and nothing else. This picture also admits of no agent-caused free decisions.”

[14] Ibid., 79.

[15] Ibid., 81.

[16] Ibid., 80.

[17] Ibid.

[18] Timothy O’Connor, review of Living Without Free Will, 309.

[19] O’Connor, Review, 309.

[20] Ibid.

[21] Ibid.

  1. February 20, 2012 at 9:18 AM

    Another great post Jared. I am only now beginning to see how deep this redux/non-redux conflict can go. Your observation about the Pereboom/O’Connor dialogue seems to apply to various philosophical conflicts.

    In my brief experience with the University of Colorado, Boulder, I observe that some of the most salient disagreements in the philosophy department stem from a disagreement about materialism/non-materialism (a divide that I take to be similar, albeit non-identical to the redux/non-redux conflict).

    Oh, and be sure to submit something to March’s Philosopher’s Carnival over at CMT. If you don’t, I’ll just have to snag something myself.

  2. February 20, 2012 at 5:51 PM

    Hey Nick,

    Thank you! Congrats on YOUR great post on U.S. Consciousness being featured in the latest Carnival!

    I would agree, reduction plays a large role in many different fields of philosophy, even those one might not expect. The materialism/non-materialism divide does not surprise me too much – it seems to touch nearly ever theory, and even if one does not begin from that point, it often seems one must come down on either side eventually.

    I will be sure to submit something! Likely it will be this past two-part post.

    • February 20, 2012 at 10:02 PM

      Great. Those are the posts I would have snagged would you have failed to submit.

  1. February 20, 2012 at 2:29 PM
  2. August 17, 2012 at 1:58 PM
  3. September 6, 2012 at 7:31 PM

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: