The more radical parts of agential realism—the ontological prioritizing of phenomena and the questioning of prominent epistemic dualism—are at tension with Bohr’s writings. However, Barad’s primary aim is after all not to interpret Bohr but to show what implications quantum mechanics must have in social theorizing. As she emphasizes, “I am not interested in drawing analogies between particles and people, the micro and the macro, the scientific and the social, nature and culture; rather, I am interested in understanding the epistemological and ontological issues that quantum physics forces us to confront” (Barad, 2007, p. 24). Agential realism is Barad’s answer to this question. This section, however, argues that the adoption of agential realism is by no means forced upon us by quantum mechanics. It is well beyond the scope of this paper to discuss every detail of quantum mechanics up against Barad’s interpretation. Rather, we shall provide a two-fold indirect argument to the effect that at best agential realism is consistent with quantum mechanics according to our current best understanding of the quantum world.
First, we argue by the example of Bohmian mechanics that there are interpretations of quantum mechanics that share very few features in common with agential realism. This suggests, as one might have expected, that agential realism is not entailed by quantum mechanics in any strict sense of entailment. Second, we compare agential realism to Rovelli’s relational quantum mechanics which Barad (2007, p. 333) describes as similar to agential realism. The point of the comparison is to signify that agential realism has subtle differences even from relational quantum mechanics and that these differences importantly manifest themselves in the measurement problem where agential realism cannot help itself to the same solution as relational quantum mechanics. Indeed, the ontological inseparability of object and agency of observation indicates a tension in the solution to the measurement problem within agential realism. Thus, not only are there many other and very different interpretations of quantum mechanics, but it also remains to be clarified whether agential realism is even consistent with quantum mechanics, or so we argue.
The differences between agential realism and Bohmian mechanics are also noticed but not further developed by Pinch (2011). As Pinch observes, Bohm dissented from Bohr’s view in two important respects: He rejected the inevitable role of the classical terminology advocating the development of a new quantum language. We only mention it to emphasize that Bohr’s contextualism with its central role for the classical terminology is not generally accepted. Furthermore, it is in any case a philosophical question whose relevance is amplified by quantum mechanics, but which quantum mechanics does nothing to answer. What we shall discuss in a bit more detail is Bohm’s view—inherited from Einstein, de Broglie, and others—that what we described earlier as indeterminacy, for instance in Heisenberg’s relations, should (in a particular sense that we return to below) be interpreted as mere uncertainty. These so-called hidden variable theories argue that quantum objects do have determinate trajectories through space, and that the wave function therefore gives an incomplete description of the system. Indeed, Bohm (1952) finds that the results of paradigmatic quantum experiments can be accounted for by the introduction of an additional dynamical equation—a guiding law—that relates the wave function to the change of the position of the particle, i.e. to the velocity. As is well known, the violation of Bell’s (1964) inequalities entails that any such hidden variables theory must be non-local, meaning that it must feature action-at-a-distance.Footnote 30 Such non-locality appears in Bohmian mechanics when the velocity of a particle in entangled many-particle states is affected or guided instantaneously by the other particles irrespective of their mutual distance.
The interested reader can consult Goldstein (2017) and references therein for more on the technical details of Bohmian mechanics. We shall for present purposes simply observe that Bohmian mechanics reproduces all the predictions of quantum mechanics, while promoting a metaphysics that is significantly different from that of agential realism.Footnote 31 Recall that according to agential realism, there are no individual objects with determinate properties that are simply revealed by measurement and relatedly, complementarity signifies the metaphysical indeterminacy of the property not being measured. Phenomena are the fundamentally real from which the separation into object of study, apparatus and observer are derived with different intra-actions enacting different such boundaries. In comparison, the ontology of Bohmian mechanicsFootnote 32 comprises of individual particles with determinate position and (instantaneous) velocity,Footnote 33 which provides for pre-determined trajectories through space and time.Footnote 34 Bohmian particles are individual “local beables” in the terminology of Bell (2001). Any event is accounted for in terms of these trajectories of the particles that constitute all elements involved in the event (Dürr et al., 2004). In a measurement, for instance, both the object of study and the measurement apparatus must be included in the Bohmian description to yield the correct results. As such, the entire universe should ideally be included in the Bohmian description.
This descriptive feature provides for two similarities to agential realism, but ultimately for many more differences. First, Bohmian mechanics erases the ontological division between the quantum objects and the experiments studying them, which also follows by their merging into phenomena in agential realism. Second, in order to reproduce quantum phenomenology, all properties apart from position and velocity are produced from the interaction between the quantum object and the apparatus (or more generally, the environment).Footnote 35 These properties—including energy, momentum, and spin—do not pre-exist the measurement; again, a thesis that is also central to agential realism. The import, however, is very different. In Bohmian mechanics, an experimental context does not project—or enact, as Barad would have it—these properties from indeterminate to determinate. Rather, these properties can simply be regarded as non-existing in Bohmian mechanics. They function as convenient book-keeping devices for the changing positions of particles which are ultimately realFootnote 36: “the state of a physical system is completely and precisely determined, at any moment in time, by the actual particle positions and the wave function, fixing how the positions change in time” (Lazarovici et al., 2018, p. 7).
The Bohmian ontology is in other words very different from the ontological relational holism of agential realism where “[b]oundaries, properties, and meanings are differentially enacted through the intra-activity of mattering” (Barad, 2007, p. 392). In Bohmian mechanics, quantum object, apparatus, and observer are all part of the universe’s determinate configuration of individual interacting particles in space that evolves entirely deterministically by specified laws of motion. The non-locality of the Bohmian guiding law entails that there is an intrinsic interconnectedness in this ontology; however, it is an interconnectedness between individuals. Again, our point here is not to promote Bohmian mechanics on behalf of Barad’s agential realism. Instead, the case of Bohmian mechanics shows that agential realism is not exposing “the epistemological and ontological issues that quantum physics forces us to confront” (Barad, 2007, p. 24). At best, agential realism provides us with the epistemological and ontological implications of one particular interpretation of quantum mechanics, whereas the same sorts of implications are very different in other interpretations.
That the implications are different in crucial respects in other interpretations is even the case for those interpretations that Barad finds to “have important features in common with each other and with the view [agential realism] presented here” (Barad, 2007, p. 333). These (allegedly) associated interpretations are Mermin’s (1998) Ithaca interpretation and Rovelli’s (1996) relational quantum mechanics. The latter in particular is superficially similar to agential realism in its account of measurement, properties, and observers. Relational quantum mechanics rejects “the notion of observer-independent values of physical quantities” (Rovelli, 1996, p. 1637). This immediately entails that one cannot ascribe determinate values of properties to quantum objects as they are in themselves. Rovelli, however, goes further and argues that such values are observer-dependent in the sense that they are never intrinsic to the quantum object: “Value actualization is a relational notion like velocity” (Rovelli, 2018, p. 6). The velocity of something is always relative to some observerFootnote 37 and the same is the case for the value of all properties in relational quantum mechanics. While Rovelli’s mode of presentation remains one where the value is ascribed to the observed object or system, it seems equivalent to adopt a more relational mode where velocity, for instance, is the relative movement of observer and observed. This comes very close to Barad’s ascription of properties to the relational whole of phenomena.
Furthermore, as is also the case in agential realism, there is no inherent boundary between object and observer in relational quantum mechanics: “Standard quantum mechanics requires us to distinguish system from observer, but it allows us freedom in drawing the line that distinguishes the two” (Rovelli, 1996, p. 1643). This split—or agential cut in Barad’s terminology—is required, according to Rovelli, since the observer cannot be part of the quantum mechanical description in terms of wave function, if the observer is to determine the value of some observable related to the object. This feature is central to Rovelli’s response to the measurement problem where O is an observer measuring and thereby observing a determinate value for some initially superposed observable of the system, S.
The unitary evolution does not break down for mysterious physical quantum jumps, due to unknown effects, but simply because O is not giving a full dynamical description of the interaction. O cannot have a full description of the interaction of S with himself (O), because his information is correlation, and there is no meaning in being correlated with oneself (Rovelli, 1996, p. 1666).
Another observer, P, who describes the combined system of S and O prior to a measurement, will describe it as a superposed state where S and O are correlated (or entangled, in technical terms). There is no contradiction here, as Rovelli observes, since the determinate value ascribed by O to S is relative to O, whereas the superposed state is relative to P. In parallel, Barad emphasizes that “measuring agencies” cannot determine their entanglement with the measured system and remarks in reply to the measurement problem that:
we are either describing a mark on the ‘measuring agency’ […], in which case what it measures is its correlation with the system with which it intra-acts, constituting a particular phenomenon; or we make a different placement of the agential cut, using a different experimental arrangement such that the complete ‘original’ phenomenon, this time including what was previously marked as the ‘measuring agency,’ is being measured by the ‘new’ ‘measuring agency,’ in which case it is possible to characterize the existing entanglement” (Barad, 2007, p. 348).
Analogous to relational quantum mechanics, the phenomenon constituted by the measuring agency and system has a determinate value for the measurement, whereas an outside observer has another description.
However, there are also important differences between these two solutions to the measurement problem. For Barad, there is no collapse or rather no destruction of the entanglement between observer and observed upon measurement. The appearance to the contrary arises from the fact that the measuring apparatus cannot measure itself; however, the entanglement is there all along: “we should not conclude from the fact that the entanglement is not made explicit by this measurement that the entanglement has become ontologically ‘disentangled’” (Barad, 2007, p. 348). As such, Barad seems to suggest that the description obtained by a measuring agency within a phenomenon is incomplete, since it does not include this “extension of entanglements that take place through measurement intra-actions” (Barad, 2007, p. 350). In contrast, Rovelli insists that the wave function with its superposition—and thus possible entanglement—is merely a bookkeeping device. Describing a quantum event as “the actualization of the value of a variable in an interaction”, Rovelli argues that “[t]he proper ontology for [relational quantum mechanics] is a sparse ontology of (relational) quantum events happening at interactions between physical systems” (2018, p. 7; see also Smerlak & Rovelli, 2007). The world consists of a sequence of interactions between systems where physical quantities take determinate values, which means, of course, that the systems precede the interactions.
While the values remain relational, this ontology of real systems and interactions is profoundly different from the relational holism entailed by the treatment of phenomena as the ontological primitive in agential realism. In this case, however, this ontological difference signifies more than the distinctiveness of agential realism compared to other interpretations of quantum mechanics. Any theory featuring collapse upon measurement faces the difficulty of explaining the collapse (or at least indicate when a collapse takes place and when contact between systems follows the unitary evolution given by the Schrödinger equation).Footnote 38
However, in rejecting the reality of the wave function, there is no ontological collapse in relational quantum mechanics (see Dorato (2016, sec. 3.2) for a discussion). It is simply a brute fact in relational quantum mechanics that the quantum events follow the pattern specified by the quantum formalism (in a relative state formulation (Everett, 1957)).Footnote 39 Agential realism cannot help itself to a similar solution: The intra-actions within phenomena enact boundaries of system and measuring agency that produces determinate values for observables, but at the same time these two systems become (ontologically) entangled (though this is not explicit in the measurement and can only be measured by a new external measuring agency.) It is true, as Barad remarks, that the system “appears as a mixture if the degrees of freedom of the instruments are bracketed” (2007, p. 346), but this does not explain why a measurement finds one rather than another value associated with the eigenstates of the (improper) mixture. Rovelli’s strategy of relational values and states does nothing to resolve this issue.
On Rovelli’s view, there is no contradiction in assigning a determinate value to the phenomenon comprising of system, S, and observer, O,—in Rovelli’s terms, a determinate value to S relative to O—and ascribing an entangled state to S + O relative to an external observer, P. This, however, provides no explanation why O finds a determinate value when Barad insists that S and O are not ontologically disentangled by the measurement. Barad can of course resort to Rovelli’s solution, but this would compromise the ontological relational holism that drives many of the other alleged consequences of agential realism detailed in Sect. 2. There might be other more subtle ways to address the measurement problem within the framework of agential realism, but if nothing else, the present discussion exposes the need to supplement the many explorations of uses of agential realism in interdisciplinary work with a study of how it fares when faced with the standard foundational questions of quantum mechanics. In the absence of such studies, even saying that agential realism is consistent with quantum mechanics involves a leap of faith.