Some philosophers of science have suggested that scientific discussions of “reductionism” and “emergentism” are merely rhetorical funding grabs. But drawing together my work in earlier parts of the book, in the final section, Part IV, I outline how we are in substantive, ongoing and empirically resolvable scientific debates about the structure of scientific cases of compositional explanation from superconductors to the eusocial insects shown in the picture above. In this post, I briefly highlight the live positions in our new debates, the empirically resolvable issues between them, and how and why the debate is ongoing in various concrete scientific cases. Once again the post is compressed, but I hope it may convince you they are interesting issues to explore.
Let me note some important tidying-up I begin by doing in Part IV. For one might ask why I am bothering with scientific reductionism if I take its most prominent argument, in the simple parsimony reasoning using the existence of compositional explanation, to be invalid? The answer is that I return to my findings about scientific reductionism to illuminate a valid argument for compositional reduction in a case of compositional explanation. This reasoning is based around the applicability of compositional concepts plus what I call the ‘determinative completeness’ of components in a case – basically the claim that the relevant components are only determined by other components. I term this the ‘Argument from Composition and Completeness’, but I highlight how this new reasoning has far, far higher evidential requirements than the invalid Argument from Composition.
Secondly, I use what I term the ‘Argument from Exhaustion’ to formally establish that machretic determination is required, whether alone or in concert with other forms of determination, in order for a composed entity to be determinative. I consequently confirm that in order to exist a compositional relation must always be accompanied by a machretic relation. And I also thus establish that Mutualism is plausibly the only viable ontologically non-reductive position about the structure of a case of compositional explanation or about nature itself.
With these two arguments in hand, as well as earlier findings, I then offer a brief survey of which positions about a case of compositional explanation are live and which are dead. To begin, I show that the most popular positions in science and philosophy, in positions built around the Scientifically Manifest Image or standard forms of non-reductive physicalism, are dead positions that we can now see must collapse into other views. In contrast, I show that there are three live positions in the Mutualism of scientific emergentism and two versions of the Fundamentalism of scientific reductionism.
In our new debates, there are two broad layers of ontological disputes. The first, and more familiar, set of differences concerns the range of determinative entities, the kinds of determination we find in nature, and hence the overall structure of cases of compositional explanation. On one side of these disputes, there is the Fundamentalism of scientific reductionism that subtracts compositional relations and levels to leave us with a universe that has one layer of determinative entities, but alongside a rich macro-world involving their collectives and higher sciences that study them. On the other side of these disputes, one can embrace the Mutualism of scientific emergentism where we do have walls of compositional relations rising vertically upwards in the house of nature, but only alongside the addition of a solid roof of machretic relations pressing downwards from composed entities to components. The Scientifically Manifest Image is not amongst the viable positions, nor are the standard versions of non-reductive physicalism defended in philosophy, since I show that the only live options involve either adding to, or subtracting from, the ontological accounts of these popular positions.
On top of this first set of disputes, my work also illuminates that in the new era focused on collectives we have a second layer of ontological differences that has been largely overlooked by philosophers and which has been left implicit even in scientific debates. This second set of disputes focus on differences over the continuity/discontinuity of nature and, in particular, the continuity/discontinuity of aggregation across simpler and more complex collectives. Central to these disagreements are differing epochal assumptions in the opposing Simple and Conditioned views of aggregation, outlined in the last post, and held by scientific reductionists and emergentists.
My work shows this second set of disputes does not neatly line-up with the first set of issues and leaves us with a still more complex set of live positions. I highlight this situation by illuminating a final overlooked position in a species of Fundamentalism that follows Mutualism in embracing discontinuities in aggregation, and hence the Conditioned view of aggregation and the existence of differential powers, whilst still endorsing the determinative completeness of components and pressing parsimony arguments to show that component entities are the only determinative entities. I thus distinguish this overlooked position, in what I term ‘Conditioned’ Fundamentalism, from the Fundamentalism commonly held by scientific reductionists like Weinberg and many in philosophy that I label ‘Simple’ Fundamentalism. Crucially, Conditioned Fundamentalism offers a live, and often very plausible, account of the components in the complex collectives in actual cases of compositional explanation.
In this second set of disputes about concrete scientific cases, we thus have Simple Fundamentalists on one side pressing the continuity of nature, the non-existence of differential powers of components, and the truth of the Simple view of aggregation. However, on the other side of this dispute, we have two positions in Mutualism and also Conditioned Fundamentalism that endorse discontinuities in aggregation, the existence of differential powers of components and hence the truth of the Conditioned view of aggregation. However, these two views then themselves diverge over other issues. Crucially, Mutualism and Conditioned Fundamentalism differ over what entities determine the novel behaviors and differential powers of components, where the Mutualist claims ‘emergent’ composed entities machretically determine these contributions and the Conditioned Fundamentalist claims other components determine these contributions.
This abstract survey begins to draw out the nature of the three live views and the more complex array of disputes between them over the nature of concrete scientific cases were we have compositional explanations. But I also directly focus on the import for cases of compositional explanation of the three live views.
The negative conclusions of this detailed work confirms that debates in the new era of collectives have moved on from the issues of the last generation of debates that too many scientists and philosophers often still focus upon. Writers on both sides, including those who have hit on one of the live views, continue to take their opponents to hold views from the last generation of debates. But my work throughout the book shows that our present disputes do not concern any of the following: whether we can supply a compositional explanation in the case; whether compositional concepts apply comprehensively in the case; whether there is multiple composition in the case; whether we may need to posit higher level scientific explanations to capture all the truths about the case; whether we may need to use non-linear dynamics or explanation by simulation; whether feedback loops exist, amongst other issues. My work shows that all the live views in the new debates, both reductionist and emergentist, are agreed about all of these phenomena and that our discussions in the sciences have moved on.
In contrast, my positive work clarifies how the three live views do differ over a finer grained array of ontological issues in any case of compositional explanation. Broadly put, the hypothesis of Simple Fundamentalism is that, underneath all the semantic and epistemic complexity, aggregation really is continuous and we have no differential powers of components in complex collectives, the Simple view of aggregation applies to the case, and we only have component entities and the determination between them. In contrast, the hypotheses of Mutualism and Conditioned Fundamentalism both assert that aggregation is discontinuous, that components have differential powers in the relevant complex collective and hence that the Conditioned view of aggregation applies in the case. However, Mutualism and Conditioned Fundamentalism differ over the entities that determine the differential powers of components, the nature of the determinative entities and the species of determination we find in the case.
Building on this work, I then articulate particular ways to empirically confirm or disconfirm the three live hypotheses about the structure of a case of compositional explanation. I show that the hypothesis of Simple Fundamentalism may be confirmed or disconfirmed by resolving whether components have differential powers using the usual scientific methods by which we illuminate causal relations and hence powers. Assuming that we have found there are differential powers in some example, I outline how we may confirm or disconfirm the hypotheses of Mutualism and Conditioned Fundamentalism by resolving whether the contribution of differential powers by components is determined by composed entities or by other components. Once again, I note that common scientific methods plausibly suffice to supply such findings. Overall, I therefore outline how the issues of the new debates are empirically resolvable and my work clarifies how to break the present stalemate in these discussions using scientific methods.
I use my findings, in Chapter 10, to provide an assessment of what has been established so far from our empirical evidence. I therefore survey extant defenses using empirical evidence from both sides of the debates, in philosophy and the sciences, and focusing on cases from across the higher and lower sciences ranging from quantum mechanical derivations of chemical phenomena and the nature of superconductors, at one extreme, through to the nature of the ants shown in the picture above, and other eusocial insects and their colonies, at the other.
On the positive side, my work shows that a range of philosophers, and scientists, have now made plausible cases that in some examples we have, or are close to getting, the kind of empirical evidence that satisfies the threshold of evidence I argue is needed to resolve the new debates. However, on the negative side, I show that each of the extant defenses is plagued by the theoretical problems I have identified and hence plausibly fails due to one or both of two problems.
The first kind of difficulty I highlight in extant defenses is that, contrary to the claims of their protagonists, once we understand the empirical evidence relevant to our ongoing disputes, then we often simply find that we do not yet have the kinds of evidence in examples passing the necessary threshold.
In addition, the second problem I identify with extant defenses is still more important because it bites even when we do have sufficiently detailed evidence of the right kind. Crucially, for the reasons highlighted in earlier Chapters, I show extant defenses fail to engage their strongest relevant rivals or the key issues with them. But scientific theory appraisal is comparative, so extant defenses are unsuccessful because they have so far failed to engage relevant rival hypotheses and the key disputes with them – even in cases where I suggest we may well already have evidence meeting the threshold for productively engaging these disputes.
My final conclusion is consequently that, at least as yet, it is still very much an open question whether we should accept either version of Fundamentalism, or instead endorse Mutualism, about any of the particular scientific examples where the proponents of such positions are presently battling across physics, chemistry, biology and the sciences of complexity. My work therefore shows that, far from being mere rhetoric, recent scientific battles over reduction and emergence involve competing, substantive scientific hypotheses that differ over the nature, and structure, of concrete cases of compositional explanation in the sciences in empirically resolvable ways.
Our new debates are thus plausibly important, and ongoing, scientific debates. And my final conclusion is that, just as Laughlin claims, we have indeed moved into a new era, and one focused on collectives and their components, but that it still remains to be established whether the scientific reductionist or emergentist is correct about cases of compositional explanation across an array of levels and sciences. Still more exciting, my work shows how to resolve these presently stalemated disputes and that we may already have the evidence to hand that allows us to do so once our better theory allow us to marshal it properly. I also identify an array of important theoretical questions that urgently need to be addressed. There is thus much still to do for scientists, and philosophers, interested in reduction, emergence and the structure of nature.
Hi Carl – thanks for blogging on this… sounds like the book will be a landmark volume.
We’ve talked before about how some causal powers require fairly complex conditions to manifest themselves. This raises the possibility that the phenomena these scientists have allegedly observed are not cases of higher-level entities “machretically determining the powers of their realizers”, but the manifestation of their components’ powers that were there all along… it’s just that their manifesting conditions are complex, where many components need to be in place in the right relations. (Dispositions can exist unmanifested, of course, just waiting for the manifesting conditions to be met.)
Is that the idea behind the Conditioned Fundamentalist explanation of differential powers?
Hi Dan, thank you for the comment which is spot on. The Conditioned Fundamentalist accepts that we have differential powers in some case, but denies that the composed entity is determining their contribution by the components, contra the Mutualist.
So, for example, to take a systems biology case, rather than the cell with its properties determining a protein has a differential power the Conditioned Fundamentalist might claim that other proteins, and as you rightly emphasize their relations as well as properties, determine the contribution of the differential power. So all the determination is still at the component level.
I highlight in the book how the failure to appreciate, and engage, Conditioned Fundamentalism has been a problem for scientific emergentists and their allies in philosophy (and I am one of those!). For even when they supply evidence making differential powers plausible, the Mutualist then needs to go on to show the composed entity determines the differential power, contra the Conditioned view. I claim no one does this, as yet, so even in the most promising emergentist cases (of which there are quite a few) Mutualism has not, as yet, been established.
However, I argue that we can empirically resolve which of the two hypotheses is correct. People often assume the issue is intractable, but I think that fails into a common error. I can talk about how one might resolve the Mutualist/Conditioned Fundamentalist dispute if you want. But you may be going somewhere else — and your comment is exactly right.
That’s exactly where I was going… I’d love to hear a bit more on how the mutualist/conditioned fundamentalist dispute could be resolved empirically.
Also: are there perhaps two versions of the conditioned fundamentalist position: 1) the differential powers are acquired by the fundamental components upon entering into the requisite formation; and 2) the differential powers are possessed beforehand (and always, as long as the individual fundamentals exist).
Hi Dan, great – let me take the first, main question, since this is a little long on its own.
The key point is that the Mutualist and Conditioned Fundamentalist hypotheses each make opposing claims about the determination relations between various entities and hence entail diverging predictions about the existence of differential powers when certain entities are present or absent.
The Conditioned Fundamentalist hypothesis claims the contribution of differential powers by a specific component is determined by a certain other components, call them X1-Xn. Crudely put, the Conditioned Fundamentalist hypothesis predicts that the differential powers of the component will be absent if X1-Xn is absent, and present when those components are present, all else being equal. In contrast, the Mutualist hypothesis contends that the contribution of differential powers by a component is determined machretically by a certain Strongly emergent composed entity, call it H. Thus, again putting things roughly and ignoring a range of complications, the Mutualist account predicts that the differential powers will be absent if H is absent, and that the differential powers will be contributed if H is present, all else being equal.
If we can secure quantitative accounts of the nature of components across simpler and complex collectives, then it appears plausible that we can potentially empirically resolve whether we have differential powers when X1-Xn, and H, are present or absent. Such quantitative accounts describe the powers of components in various situations, so it is plausible that the same methods can allow us to explore whether we have differential or other powers across such collective with permutations of the relevant components and composed entities being absent or present. Using this richer array of evidence, which is often supplied by scientific inquiry, we can consequently resolve whether the predictions of the Mutualist or Conditioned Fundamentalist hypotheses are true or false. (Please note that nothing I just laid out involves ideal interventions, in the sense of Woodward’s machinery, between composed entities and their own components, since I take there to be good reasons to think that ideal interventions on composed entities with respect to their components are impossible, and vice versa. But we can get evidence about presence and absence of entities without such ideal interventions.)
The latter conclusion flies in the face of many people’s initial judgment that the choice between Mutualist and Conditioned Fundamentalist hypotheses is empirically unresolvable. People are moved by a version of the “snapshot fallacy” and the associated assumption that if we can resolve disputes between the opposing hypotheses at all, then these disputes will be resolvable solely using the evidence of a compositional explanation alone. But, claims the objector, the Mutualist and Conditioned Fundamentalist hypotheses each account for the features of the compositional explanation equally well. So, the objector concludes, we cannot empirically resolve whether we should endorse Mutualism or Conditioned Fundamentalism when we have differential powers.
But this intuitive judgment is badly mistaken. Not only evidence from the compositional explanation, and hence about the complex collective with all its components, is relevant to assessing the Mutualist and Conditioned Fundamentalist hypotheses. In fact, evidence about components, and their powers, is required from across both simpler and complex collectives, for only then can we resolve whether some power is a differential power or not. The intuition driving many people’s initial judgment is therefore mistaken and we can potentially resolve whether to accept the Mutualist or Conditioned Fundamentalist hypothesis if we have sufficiently detailed empirical evidence of the right kind.
Or so I claim. That is the bare bones account I outline.
I sort of get it… but I think I might need a worked out example. No doubt I just need to read the book! Which I will most certainly do.
Well maybe this will help. Imagine the case is again a systems biology one where the emergent property G is a property of a cell, and the component is a protein having a differential power CD contributed to it by one of its properties P1.
Assume the Conditioned Fundamentalist says this ONE other component property, call it Px, of some other kind of protein determines the contribution of the differential power — remember I am just trying to show in practice we could ever confirm/disconfirm the relevant hypotheses. The Mutualist says G determines the differential power — G could be a shape, metabolizing at a certain rate, or a combination of properties of the cell.
Ok, now consider the following.
First, note that multiple composition is pretty ubiquitous in nature and especially in biological systems given the way evolution works. So assume that the cell can be composed without any protein with property Px amongst its components. Now see if the protein still contributes CD — if it DOES then the relevant Conditioned Fundamentalist hypothesis is falsified, right?
Second, notice that we could well have the same protein with property P1 in much simpler systems than a cell, say a petri dish in the lab, where the system of components does not compose a cell with property G. Now look at the protein with P1 in the simpler and see if it has the differential power CD. If it DOES, then the Mutualist hypothesis is false, right?
Building on this base, and adding in the various nuances we need to make the inferences safe, I was suggesting we can use such cases of presence and absence, and observations of whether we have the differential power CD, so we can potentially often falsify one hypothesis but not the other.
That was crude, but does that help?
Actually, no, I don’t see how this is going to work. It seems to me that there will be a way to make P1 & Px (related by some relation R) equivalent to G, at least for all empirically testable purposes, and where you don’t get the differential power without P1 & Px & R, or, what amounts to the same thing for empirical purposes, without G. (Or P1 & Px & R could be replaced by a disjunction that covers all the multiple composition cases. Or Px can be given a purely dispositional/power formulation rather than a property formulation.)
I’ll check out the book for some of the actual scientific examples… I think that will help. Thanks!!
On the second point. I kind of remember talking about this before? You are outlining what I take to be a foundational issue for powers theorists with regard to conditional powers. On the first option conditional powers are only ever contributed to an individual under a certain condition #. On the second option, the conditional power is always contributed by a property but not switched “on” and open to manifestation except under the relevant structural condition # (one still needs further manifestation and triggering conditions for it to manifest).
I think the Mutualist or Conditioned Fundamentalist views come out the same under either option. As far as I can see what individuals produce, under all conditions, is the same under the two options — but the ontology *appears* to be different. Since different options do not lead to different behaviors under any condition, I think in this context the choice between the options does not matter. That is a quick pass…
The foundational issue an interesting one though, but I do not know of anyone in print who engages it. If you or anyone else has engaged it, I would be interested in reading the paper.
I wasn’t think of the second option quite like that… more like an ordinary power that has complex manifesting/triggering conditions. The structural condition doesn’t switch it on in any sense. The structural condition is just one part of a complex manifesting condition for an ordinary (fundamental) power.
This has, at least, the advantage of parsimony, since we need ordinary fundamental powers in any case.