(i) Introduction
(ii) Aristotle: Substance and Essence
(iii) Rovelli on Nāgārjuna and Western Philosophy
(iv) Carlo Rovelli’s Relationalism

Firstly, let’s ignore Aristotle for a moment, and see what the word “substance” means when taken much more broadly. Here’s one account:
“In its first sense, ‘substance’ refers to those things that are object-like, rather that property-like. For example, an elephant is a substance in this sense, whereas the height or colour of the elephant is not. In its second sense, ‘substance’ refers to the fundamental building blocks of reality.”
Now let’s take this account of Aristotle’s own ontology of substances:
“Aristotle says that substances do not depend, for their existence, on any other being of which they must be predicated or in which they must inhere.”
Some readers will be able to see why the theoretical physicist Carlo Rovelli — as a relationalist — has a problem with Aristotle’s ontology.
This Aristotelean stance on substance is why the term “particular” came to be used. Relevantly, it was widely believed that a particular does not depend, for its existence, on any other being.
Relationalism, on the other hand, inverts this binary opposition by arguing that all substances (or all particulars) are — entirely?- dependent for their existence on their relations to other beings…
This means that such ?’s can’t really be substances (or particulars) at all!
Rovelli himself expresses all the above in a more technical way when he allows Aristotle (384–322 BC) to put a position (as paraphrased by Rovelli himself) which seems to be the exact antithesis of his own. Rovelli writes:
“For Aristotle the relation is a property of the substance. It is that belonging to the substance that is towards something else. Among all the categories, for Aristotle, relationality is the one that has ‘least being and reality’. Can we think differently?”
So, according to Rovelli, relationality has the most being and reality. However, is Rovelli’s strong stress on relations (or relationality) automatically to factor out substances, objects or particulars? (It’s not clear that Rovelli himself is differentiating these three different things.)
In any case, it can be supposed that most laypersons would probably sympathise more with Aristotle’s position than with Rovelli’s. Indeed, isn’t a relation something that has the “least being and reality” when compared to an object (i.e., if not to a substance)? What’s more, intuitively, it’s hard to see a relation as being “a property of the substance” at all.
Yet, of course, this issue is partly (or largely) dependent on how the words “relation” and “object” are defined.
Again, relations do indeed seem to have the least being and reality in the simple sense that they can’t be observed, touched, smelled, kicked, tested, or experimented upon. The objects which have (or bear) such relations, on the other hand, are physical: therefore they can be observed, touched, smelled, kicked, tested, or experimented upon.
Does all that make Rovelli’s relations abstract?
If not abstract, then… what?
To move on with Rovelli’s Aristotle.
Aristotle: Substance and Essence

Where there’s an Aristotelian substance, there’s also an essence — i.e., at least that’s often been the case in much Western philosophy. And Rovelli picks up on this too when he writes the following:
“If every metaphysics seeks a primary substance, an essence on which everything may depends [ ].”
It’s not clear if Rovelli seeing the words “substance” and essence” as being (virtual?) synonyms here. Or, if these words aren’t synonyms, then Rovelli seems to fuse the ontological categories substance and essence together. However, according to the following account, substance and essence aren’t the same thing:
“Aristotle examines the concepts of substance (ousia) and essence (to ti ên einai, ‘the what it was to be’) in his Metaphysics (Book VII), and he concludes that a particular substance is a combination of both matter and form, a philosophical theory called hylomorphism.”
That means that a substance has (or bears) an essence. In other words, a substance isn’t the same thing as its (own) essence. Moreover, the essence characterises “the what it was to be” of a substance. All that said, the essence must somehow still be instantiated by (or in) the substance…
Rovelli on Nāgārjuna and Western Philosophy

As already stated, the Aristotelian stress on substances seems to go alongside the stress on their ontological independence — at least according to Rovelli. Indeed, Rovelli uses the ideas of the Indian Buddhist philosopher Nāgārjuna to put his point across. He tells us that Nāgārjuna argued that
“it is possible to think of the manifestations of objects without having to ask what the object is in itself, independent from its manifestations”.
Rovelli uses the words “the manifestations of objects”. (More specifically, he uses the two words “of objects”.) Clearly, and perhaps obviously, then, objects aren’t really factored out of this alternative relationalist picture. Yet if Rovelli’s ontology is literally all about “manifestations” (as he makes it seem), then why mention objects at all?
What’s more, it’s not even clear that many people (Westerners or otherwise) do believe in the arcane notion of an object-in-itself. Sure, a few (perhaps more than a few) philosophers have done so in the past. However, virtually no laypersons or scientists do so today. Yet Rovelli’s argument is that this way of thinking has been central to all Western thought — not only to the thought of a small group of philosophers.
All that said, elsewhere Rovelli does state the following:
“In the history of Western philosophy there is a recurrent critique of the notion that ‘entities’ are the foundation of reality. It can be found in widely different philosophical traditions, from the ‘everything flows’ of Heraclitus to the contemporary metaphysics of relations.”
If Rovelli isn’t setting up a straw target here, then we can agree with him and say that it would be difficult — if not impossible — to describe any given object as it is in itself. (Kantian things-in-themselves, for example, are things-in-themselves largely because they can’t be described.)
Rovelli then says something similar about the history of physics when he writes:
“The discovery that quantities we had thought of as absolute are in fact relative instead is a theme that runs throughout the history of physics. Beyond physics, relational thinking can be found in all the sciences.”
It’s not clear how the words “absolute” and “relative” are being used here. If Rovelli (as a follower of Albert Einstein) is talking about space and time, then he has a strong case. As it is, though, Rovelli is not talking exclusively about space and time (or Einstein’s spacetime): he’s talking about literally everything.
But what of Rovelli’s relations?
Carlo Rovelli’s Relationalism

Intuitively, surely there must firstly be an object in order for it to have relations — of any kind whatsoever.
Inversely, could it be the case that firstly there’s relations, and only then do we have an object?
Unless, that is, we believe that object O and its relations have always (as it were) belonged together. Alternatively put, object O may actually be constituted by its relations. In other words, object O wouldn’t actually be object O if it weren’t for its relations.
That latter possibility is backed up by Rovelli’s own paraphrase of Nāgārjuna’s position. Rovelli put the following words into Nāgārjuna’s mouth:
“[Structures, not relations] are neither precedent to objects; nor not precedent to objects; neither are they both things; nor, ultimately, neither one nor the other thing.”
Perhaps there never was a time when object O didn’t have any relations. Yet wouldn’t that simply mean that we’d need to take into account all of O’s relations throughout its entire existence?
(This is similar to Leibniz’s position on essence, in which all the properties of an object are essential to that object — see here.)
So how could we distinguish important or relevant relations from unimportant or irrelevant ones when it comes to any given object? …
Isn’t it here that Aristotle’s substance and essence come to the rescue?
In detail.
Any given object O will have different relations at different times.
So what about object O at all these different times?
What accounts for the very same object having many different relations at many different times?
How do we know, for example, that the very same object O at t¹ had another — different — set of relations at t²?
Again, isn’t it here that Aristotle’s substance and essence come to the rescue?
Of course, we can bite the bullet and argue that object O at t² simply can’t be the same object as ? at t¹. And surely that would mean that we’d need to dispense with the object entirely. Yet taking this eliminativist view would also mean that we can’t speak of “O’s relations” either. In other words, relations would also need to be factored out if there are no (Aristotelian) objects in the first place.