Navigazione – Mappa del sito

HomeNumeri56variaFictionalism in Ontology: The 201...


Fictionalism in Ontology: The 2011 Paolo Bozzi Lecture

Achille C. Varzi
p. 253-270


Fictionalism in ontology is a mixed bag. Here I focus on three main variants – which I label after the names of Pascal, Berkeley, and Hume – and consider their relative strengths and weaknesses. The first variant is just a version of the epistemic Wager, applied across the board. The second variant builds instead on the fact that ordinary language is not ontologically transparent; we speak with the vulgar, but deep down we think with the learned. Finally, on the Humean variant it’s the structure of the ontological inventory, not its content, that may turn out to involve fictional elements. That is, for the Humean the fiction lies, not in the reality of common-sense ontology, but in the laws – of unity, identity, causation, etc. – in terms of which we articulate our experience of that reality. In the end, this is the kind of fictionalism that I find most interesting, sensible, and tenable. And I argue that it is even compatible with the sort of “naive” realism we have all come to appreciate in the work of Paolo Bozzi, to whom the paper is dedicated.

Torna su

Note dell’autore

This is the text of the Lecture delivered at the Universty of Turin, Faculty of Letters and Philosophy, on May 26, 2011, on the occasion of the award cerimony for the “2011 Paolo Bozzi Prize in Ontology” (May 26). The cerimony was held within the framework of the conference Fictional Entities and Other Social Objects and the text of the Lecture appears also in the follow-up volume, From Fictionalism to Realism, edited by Carola Barbero, Maurizio Ferraris, and Alberto Voltolini (Newcastle, Cambridge Scholars Publishing, 2013).

Testo integrale

1At first, I thought that there was something ironic in the idea of honoring Paolo Bozzi within the context of a conference on fictionalism. Paolo was a champion of realism. Fictionalism, at least on some understanding of the term, is naturally viewed as a tool for the antirealist. On second thought, however, I’ve become convinced that there is a good way of honoring Paolo by speaking about fictionalism, especially fictionalism in ontology.

1. The ontological wager

2When it comes to ontology, factionalism is a mixed bag. It’s not just that there are different ways of construing ontological fictionalism. In ontology, there are different views that may be construed as fictionalist.

3Here is a familiar view. One could say: Look, for all we know, we may be brains in a vat. Maybe there really are no such things as cats and dogs; maybe they are just illusions. Maybe the whole world as we experience it, including our own bodies, is just a product of our own minds – a fiction. Maybe we are disembodied minds and there is nothing else. And even if there is something, it may well be that it is not what we thought there was. (And so on.)

4Still – our philosopher may continue – even if everything is just a fiction of sorts, what goes on in that fiction matters a lot to us. Indeed, that’s all that matters, for we feel as though it were real. We do have the impression of having a material body that lives in a world inhabited by cats and dogs, and our happiness, which is what matters the most, depends on how successfully we manage to interact (in whatever sense of “interact”) with such things, such illusions. Everything in our lives depends on this, illusory as its reality may be. So let us just pretend that it is not an illusion. Let’s continue to pretend that it isn’t. Let us pretend that there really are cats and dogs, and that we have a body, and so on, and let us take it from there. Perhaps such things do not, in fact, exist. But they do seem to exist and they certainly could exist, so let’s just pretend we live in a world in which they do exist. Just like when we play cowboys and Indians. Or when we watch a movie or read a novel. We step into a state of pretense and we feel joy or sorrow depending on how things go in the pretense, for what goes on in that pretense is what matters.

5What about knowledge? What about truth and falsity? Don’t worry about that, says our philosopher. All we really care about is truth and falsity in the pretense, in the fiction. Forget about the truth value of P. What matters is the truth value of:

According to the fiction, P

exactly as in:

According to the Sherlock Holmes stories, there is a brilliant detective who lives in 221b Baker Street.

6Of course we all hope that the fiction is true (not the Holmes stories, but the fiction corresponding to the world as we seem to experience it), in which case the operator is redundant and we are back to truth and falsity simpliciter. But it may well not be, and we shouldn’t waste time pondering whether it is. We would not gain anything and we would lose everything.

  • 2 *Pascal 1670: III, §233.

7Let us call this view the skeptic’s fictionalism, for obvious reasons. Another good label may be Pascalian fictionalism. After all, the position is reminiscent of Pascal’s wager in his Pensées: although we are unsure whether there is a God or not, we still ought to believe in God on the basis of expected reward2. Likewise, although we are irremediably unsure whether there are cats and dogs and so on, we still ought to act as though they existed on the basis of expected reward, fictitious as it might be.

2. Speaking with the vulgar

8The view that I have just outlined is a natural candidate for what may seem like a plausible (if not attractive) fictionalist position in ontology. But it isn’t the only one. Another view, which is becoming increasingly popular, is this.

9Take a nominalist (for example) – a philosopher who believes in a world composed entirely of particulars. There are no universals, on her view, and when we say that Sam is wise, or that the apple is red, the truth or falsity of what we say is a primitive and irreducible fact about Sam and about the apple. For years and years, if not for centuries, our nominalist philosopher had to face a constant challenge: even assuming such primitive and irreducible facts, how can we account for the truth conditions of sentences that appear to involve explicit reference to universals, or explicit quantification over them, as in

Wisdom is a virtue.
Mary prefers red to blue.
Sam and Bob have some virtues in common.

10The traditional answer was that the nominalist can always, at least in principle, come up with suitable ways of paraphrasing such sentences that are ontologically innocent, i.e., make no genuine reference to or quantification over universals – just particulars. For example:

Wise persons are virtuous.
Mary prefers red things to blue things.
Sam and Bob are both wise, or both honest, or…

11The idea was that we may use those initial sentences when we speak, but deep down this is what we mean. We may speak with the vulgar but think with the learned and, if pressed, we know how to put things, we know how to rephrase. And the paraphrases are nominalistically unblemished.

12Alas, this traditional answer turns out to be problematic in many ways. One problem is simply that the approach is too piecemeal for the nominalist to ever feel comfortable. The language we speak is rich and various. Every day the platonist may come up with a new challenge, a new recalcitrant example, and the nominalist is constantly under pressure to produce a suitable paraphrase. Moreover – and this is a bigger problem – it often turns out that what looks like a suitable paraphrase is not, in fact, entirely adequate. Take the first of our three sentences. Surely “Wisdom is a virtue” doesn’t say that wisdom is sufficient for virtuousness; it takes more to be virtuous than being wise. (Nor does it say that wisdom is necessary for virtuousness, for presumably one may be virtuous even if one lacks that particular virtue.) So perhaps what the nominalist really means with that sentence is something like this:

Wise persons are more virtuous than unwise ones.

13Yet even this paraphrase is inadequate. Surely, a wise person who lacks tolerance, patience, honesty, courage, good temper, etc. is less virtuous than someone who lacks wisdom but has all those other qualities (speaking with the vulgar). Really, if we want to reduce putative talk about wisdom to genuine talk about wise people, the right way of putting things would be something like this:

Other things being equal, wise persons are more virtuous than unwise ones.

14Yet this generates a new challenge. For on the face of it, the ceteris paribus condition expresses a quantification over qualities, i.e., universals, whish is not nominalistically acceptable. And so on. A third problem bites even deeper. For how can we assess the adequacy of a paraphrase? The intuition is that, generally speaking, A is an adequate paraphrase of B only if A has the same truth conditions as B. But this presupposes that we can determine the truth conditions of B in the first place, hence that B admits of a direct interpretation in its own right. The nominalist is not willing to concede that much when B is nominalistically unacceptable. Should the nominalist defer to the platonist on the adequacy issue? Should the nominalist say instead that the sentence she is offering, A, is not really a “paraphrase” of B, but rather the only (acceptable) way of expressing the proposition B is supposed to express? On what grounds? And so on and so forth.

  • 3 Berkeley 1710: I, § 51.

15One day, the nominalist has a brilliant idea: Look, forget about all this rigmarole. It’s a trap. To provide an ontologically transparent paraphrase of each and every possible sentence that seems to require reference to or quantification over universals? That’s ridiculous. No one speaks in a perfectly ontologically transparent way. No one should feel under pressure to speak that way. Remember Berkeley: If someone actually talked like that, wouldn’t she be laughed at, and rightly so? That’s the whole idea of the motto, think with the learned and speak with the vulgar3. There is nothing disreputable about that, for language is a tool for communication and communication is ontologically innocent. Indeed, all sorts of learned people are perfectly happy to speak with the vulgar when it comes to saying things. When they engage in communication, astronomers still say that the sun rises, the sun sets, or the sun is high in the sky, even if they are fully convinced of the truth of the Copernican theory. Physicists go shopping or talk about sport, love, and the financial crisis without feeling any pressure to rephrase everything they say in terms of the sparse ontology they believe in. Why should philosophers feel any different? Let everybody carry on and speak with the vulgar. That way we can communicate, which is what language is meant for. We just have to realize that it’s all a fiction, a game of make belief. And if we really need to make that clear, we can just say so:

According to the fiction of common sense, P.

Or perhaps:

According to the fiction of the platonist, P.

No need to tinker with the logical form of P itself. Even the semantics of P can be perfectly standard. It’s enough to preface P with the fiction operator. Ditto for all other cases where philosophers may disagree:

According to the fiction of mathematics, there are even numbers and odd numbers.
According to the fiction of semantics, the meaning of a complex expression is a function of the meanings of its constituents.
According to the fiction of modal realism, “Possibly A” is true iff there exists a possible world at which A is true.

And so onexactly as in:

According to the Sherlock Holmes stories, there is a brilliant detective who lives in 221b Baker Street.

Just as we understand Conan Doyle even if we do not believe in the existence of his fictional characters, likewise we understand mathematicians, semanticists, and modal realists – and successfully partake in mutual communication – even if we do not share the same ontology. We do understand and successfully communicate because we know how to engage in a game of make-belief.

16Let’s call this sort of fictionalism Berkeleyan fictionalism. Its great advantage, for the nominalist as for any other philosopher whose ontology doesn’t quite fit the way we ordinarily speak, is that it leaves everything pretty much as is. To repeat: No need to fiddle with the logical form of P. And even the semantics of P can be perfectly standard.

3. What’s the difference?

17Now, these two sorts of fictionalism – the skeptic’s and Berkeleyan fictionalism – are different, and surely enough the latter is much more powerful, for it does not reflect any specific epistemological stance. On the contrary, it corresponds to a strategy that is available to anyone regardless of whether she has clear and distinct ideas about what there is (and what there isn’t). Unfortunately, this greater effectiveness comes with a price – and a big one, in my view. For while the first brand of fictionalism admits of a clear semantics for the fiction operator, the second does not.

  • 4 From Lewis 1978.

18Consider David Lewis’s classical account4:

A sentence of the form “According to fiction F, P” is true iff P is true at every world where F is told as known fact rather than fiction.

More precisely, since there may be truths in a fiction that are not explicitly mentioned (i.e., background truths that the community of origin of the fiction takes for granted):

A sentence of the form “According to fiction F, P” is (non-vacuously) true iff some world where F is told as known fact and P is true differs less from our actual world, or rather from any one of the collective belief worlds of the community of origin of F, than does any world in which F is told as known fact and P is not true. (The sentence is vacuously true if there are no possible worlds where F is told as known fact.)

Thus, given Lewis’s account of counterfactuals, “According to fiction F, P” is essentially equivalent to the counterfactual

If F were to be told as known fact, it would be the case that P

though in general we are interested in whether this counterfactual is true, not only (or not necessarily) at the actual world, but at every collective belief world, i.e., every possible world that verifies all the overt beliefs of the community of origin of F. For example, the statement

According to the Sherlock Holmes stories, there is a brilliant detective who lives in 221b Baker Street.

is non-vacuously true because the corresponding counterfactual

If the Sherlock Holmes stories were told as a known fact, there would be a brilliant detective who lives in 221b Baker Street

is true at every possible world that verifies all the beliefs that we implicitly share with the author of the Holmes stories, such as that people have kidneys, that Tuesday comes after Monday, and so on. (The actual world may very well be such a world, though we would be extremely lucky if it were so. We would be extremely lucky if all of our overt beliefs turned out to be true.)

19This account is not, of course, unproblematic. Several variants and refinements have been put forward since Lewis’s original formulation, but this need not concern us. The basic idea still holds and is enough to make the point. And the point is that while this sort of account is perfectly fine for the skeptic fictionalist, it is not fine when it comes to Berkeleyan fictionalism.

20It is fine for the skeptic fictionalist, because the skeptic says that although such things as cats and dogs may not exist, they do seem to exist and they certainly could exist. The fiction could be true – and we may even hope that it is. Thus, when we use the fiction operator, we are really engaging in the sort of counterfactual thinking that Lewis’s semantics associates with that operator. We imagine a possible world where the fiction holds true – we pretend that our world is such a world – and then we see whether our sentence, P, holds at that world. Perhaps there are no cats and dogs. But if there were, they would have four legs. No special effort is needed to conceive of worlds in which the antecedent of this counterfactual is true just as no special effort is needed to evaluate the consequent at such worlds. Indeed, the actual world seems to be one of them and the skeptic still hopes that it is one of them. It’s just that she has doubts.

21Not so for the Berkeleyan fictionalist. To stick to our example, if you are a nominalist concerning universals, you are not just saying that universals do not or may not exist as a matter of contingent fact – that they are or may be fictional creatures on a par with Sherlock Holmes, with the characters of a dream, with the illusions of a brain in a vat. Even the most radical of the skeptics is willing to admit that such fictional creatures may exist. But no serious nominalist is willing to concede that much regarding universals. For a serious nominalist, universals do not belong to the furniture of this world just as they do not belong to the furniture of any possible world. Her ontological stance has the modal force of necessity. Accordingly, she cannot sincerely engage in the pretense. And if she cannot sincerely engage in the pretense, then the Lewisian account of the fiction operator cannot even get off the ground; the space of possible worlds is just too narrow for the nominalist to play that game – on pain of certifying every P as vacuously true.

  • 5 This worry is addressed in Lewis 1983. The proposed solution, however, is based on the idea that an (...)

22Ditto in all other cases where a philosopher may wish to endorse a fictionalist stance of the Berkeleyan sort. It’s not that numbers do not exist but could exist. It’s not that meanings do not exist but could exist. Serious anti-realist claims have the modal force of necessity. Lewis’s semantics, however, as every other plausible semantics for the fiction operator I can think of, works well for literary fiction precisely insofar as literary fiction comes with a sense of possibility. (Indeed, one of its most controversial limits is that such semantics doesn’t fare well with “impossible” stories5.) It also works well for epistemological fictions, and for the same reason, which is why the worry does not apply to the skeptic fictionalist. When it comes to genuine ontological fiction, the semantics is simply helpless.

4. Humean fictionalism

23Of course, Paolo Bozzi was not a Berkeleyan fictionalist. He was as ontologically honest and true to his commitments as a philosopher can be, and that brand of fictionalism would have struck him as a cheat, a swindle, an intellectual fraud. He would have been happy, I think, to hear the bad news. But neither was Bozzi a Pascalian fictionalist, for he had no liking whatsoever for the sort of skepticism that instigates it. Thus, whatever the limits and merits of those two views, so far we would be entitled to conclude that ontological fictionalism is indeed a bad topic to square with Bozzi’s overall philosophical attitude. Cats and dogs – a fiction? Wisdom – a fiction? There is, however, a third way in which fictionalism enters ontology, and here the story gets interesting even for someone like Bozzi. I am going to call it Humean fictionalism, for it is best illustrated in relation to Hume’s views concerning all sorts of fundamental ontological issues. (The word “fiction” occurs al least fifty times in the Treatise.)

24Consider causation. For Hume, causation is a typical example of a concept that does not correspond to a genuine feature of reality. There is nothing, in reality, necessarily connecting what we call “cause” and what we call “effect”. Or rather – since Hume’s claim is epistemological-semantic rather than strictly ontological – there is nothing we can observe in reality except for certain relations of succession, contiguity, and constant conjunction. Hence we cannot form any philosophically respectable concept of causation over and above that of a constant conjunction of like objects in like relations of succession and contiguity, pace our natural “propensity” to go for something bigger:

  • 6 Hume 1739-1740: I.iii.6.

Such particular objects, in all past instances, have been constantly conjoined with each other: and as an object similar to one of these is supposed to be immediately present in its impression, we thence presume on the existence of one similar to its usual attendant6.

“Presume” is not quite the same as “pretend”, but it’s close enough to justify the label “fictionalism” in the present context. We do have a propensity to presume the existence of a necessary connexion between cause and effect, to suppose that their conjunction depends upon an efficacy, an energy, a metaphysical oomph with which they themselves are endowed. And we do think and act as though such a connection were real. But it isn’t, or at least it need not be. All we have is a mental construct, a fiction of the imagination that does not reflect the structure of reality but rather helps us make sense of reality by structuring the scattered multitude of our perceptions.

25Similarly, consider Hume on identity, that is, diachronic identity. On the face of it, the thought that things persist through time underlies much of our everyday interaction with the world of ordinary experience. We readily suppose that an object may continue numerically the same, in spite of the fact that it may undergo several qualitative changes and that for most of the times it is absent from the senses. Bananas ripen, ships deteriorate, people lose hairs and acquire new body cells. In this world of flux, persisting things are the only anchor we have, but the source of their persistence is a genuine puzzle – a puzzle that has been with us since the Presocratics. What grounds our belief that the things around us (and ourselves, too) may survive from day to day, in spite of the many changes that affect them? How can we say that they are the same things, if they are no longer the same? The answer, for Hume, is that we can’t. For Hume, the identity relation applies in its strictest sense only to constant and unchangeable objects, and it is only “the smooth passage of the imagination” along the ideas of resembling perceptions that makes us ascribe identity even to variable or interrupted objects, it is our propensity to unite broken appearances of resembling perceptions that produces the fiction of a continued existence:

  • 7 Hume 1739-1740: I.iv.6.

That action of the imagination, by which we consider the uninterrupted and invariable object, and that by which we reflect on the succession of related objects, are almost the same to the feeling. […] The relation facilitates the transition of the mind from one object to another, and renders its passage as smooth as if it contemplated one continued object. This resemblance is the cause of the confusion and mistake, and makes us substitute the notion of identity, instead of that of related objects7.

Ditto for personal identity, where again Hume is quite explicit in using the language of fiction:

  • 8 Ibidem.

The identity, which we ascribe to the mind of man, is only a fictitious one, and of a like kind with that which we ascribe to vegetables and animal bodies. It cannot, therefore, have a different origin, but must proceed from a like operation of the imagination upon like objects8.

26Finally, consider unity. Here, too, Hume famously argued that we have a propensity to attribute existence to multiplicities, such as a group of twenty men, when strictly speaking existence in itself belongs only to unity – a man. Strictly speaking, unity is never applicable to a multiplicity except on account of the “unites” of which that multiplicity is composed. Thus, again, when we give way to our propensity to say more, strictly speaking we engage in a mental construction, a pretense, a fiction:

  • 9 Hume 1739-1740: I.ii.2.

These twenty men may be considered as an unite. The whole globe of the earth, nay, the whole universe may be considered as an unite. That term of unity is merely a fictitious denomination, which the mind may apply to any quantity of objects it collects together9.

  • 10 Bozzi 1969.

27I mention these three central Humean topics – unity, identity, causation – because they all feature in the title of one of Paolo Bozzi’s books: Unità, identità, causalità10. That is one of his most genuinely philosophical books, even though (or precisely because) the subtitle reads: Una introduzione allo studio della percezione, an introduction to the study of perception. I’ll try to make the connection more explicit shortly. First, however, let me try to clarify what I take to be the distinguishing feature of the sort of fictionalism that I see at work in Hume’s treatment of these topics, and how it differs from the two brands of fictionalism discussed previously.

28Both the Pascalian and the Berkeleyan brand of fictionalism are fundamentally ontological, in a strict Quinean sense. They concern what there is, or what there might be, and the pretense they engage in is a pretense concerning the putative existence of certain entities, or certain types of entity. With Hume it’s different. Here the fiction lies, not in the prima facie ontological make-up of reality, but in the laws – of unity, identity, causation – in terms of which we attribute a structure to that reality. It’s the structure of the ontological inventory, not the content of the inventory, that turns out to be fictitious. Our propensity to give way to the unifying act of our intellect, of our imagination, makes us speak as though there were a unity, an identity, a causal nexus when in fact all we have is patterns of broken appearances. In an important way, this has a direct impact on the ontology. For if we deny the existence of composite units (for instance), then obviously our inventory of what there is will contain fewer entities than if we endowed such composites with bona fide existence. But what distinguishes the fictionalist stance, here, is not the need to follow up on our ontological commitments, to provide adequate truth-conditions for our linguistic practices vis-à-vis our skeptical or parsimonious views on what we are willing to admit into our inventory of the world. It is, rather, the emphasis on the “confusions and mistakes” that drive our structuring activity, which is to say our impulse to always provide the complex system of concepts and principles through which we represent the world of experience with an objective foundation in the nature of things – over and above any specific view concerning what those things might actually be.

  • 11 In Lewis 1986: ix.
  • 12 Goodman 1983: 104.
  • 13 See Sider 1996.
  • 14 Russell 1913: 1.

29Of course, in Hume all this follows from his empiricism. But you don’t need to be an eighteenth-century empiricist to see the bite of his perspective on such matters, and the idea that all there is to the world is “a vast mosaic of local matters of particular fact, just one little thing and then another” – to use Lewis’s popular characterization 11 – has indeed been driving much contemporary philosophy on independent grounds. Think of the increasingly popular view on unity known as compositional nihilism. There are, strictly speaking, no composite objects on that view; only mereological simples. No chairs and tables; only simples arranged chairwise and simples arranged tablewise. No world and universe; only gazillions of simples frantically dancing and interacting together in an otherwise empty space. We have a tendency to “connect the dots” and to articulate reality in terms of continuous boundaries even when such boundaries are not genuinely present, and such a tendency may well be grounded in the perceptual apparatus through which we experience the herds of simples that float around us, as in a Seurat painting. But it’s all a fiction, a natural and convenient way of worldmaking. As Nelson Goodman once put it, “as we make constellations by picking out and putting together certain stars rather than others, so we make stars by drawing certain boundaries rather than others”12. Or think of the so-called stage view about diachronic identity. There are, strictly speaking, no persisting objects on that view; only processions of momentary entities following one another in time. No enduring bananas, ships, people; only sequences of instantaneous banana-stages, ship-stages, person-stages popping in and out of existence one after the other, though suitably related to one another so as to give rise to the fiction of a continued existence. As Ted Sider famously put it, “all the world’s a stage”, and when we say e.g. that this banana was green, what we say is true if, and only if, the current referent of “this banana” – a momentary stage – has a past temporal “counterpart” – another momentary stage – that is green13. Think, finally, of how the Humean view on causation has made its way into contemporary philosophical naturalism through Russell’s famous indictment: the law of causality is “a relic of a bygone age, surviving, like the monarchy, only because it is erroneously supposed to do no harm”14. True, Russell was himself an empiricist of sorts. But Hume was far behind him, as he is far behind those contemporary naturalists who claim that physics and other advanced sciences do not and should not engage in cause-seeking as a quest for objective “laws” of constant conjunction.

30In each of these cases, then, as on Hume’s original view, much of the structure that we tend to attribute to the world out there is a fiction. But this is not to say that it is a bad fiction, or merely a fiction erroneously supposed to do no harm. On the contrary, in each case one might very well think that the fiction is to be taken seriously, for it is the best fiction we could think of. We would not be able to plan our lives and to carry on with our everyday commerce with the world, let alone to pursue progress in science, if we didn’t pretend that the fiction were true. It would be irrational not to pretend that it is true. It would be foolish not to bet on the sun rising tomorrow, not to talk about chairs, ships, people, not to act as though the things we encounter today were already there yesterday and will still be there tomorrow. Still, a pretense it is, and as philosophers we have to keep that in mind. This is why the view in question is a genuine brand of fictionalism. And this is also why the fictionalist machinery described earlier applies holus-bolus. For we do care about the truth or falsity of statements of the form

According to the fiction, P

we care because the truth or falsity of such statements matters a lot in the Lebenswelt. And these statements can be given a perfectly Lewisian semantics. The world could be as the fiction says; it’s just that it is isn’t. Perhaps the world might be that way; it’s just that we cannot honestly think it is.

5. Fictionalism and Bozzian realism

  • 15 For example, in Varzi 2010 e 2011.

31I don’t know how you feel about this view. For my part, I have a great deal of sympathy for Humean fictionalism and I have occasionally tried to give my reasons15. But it is not my purpose to defend that view here. Rather, let me conclude by trying to explain why I think this sort of fictionalism is not as incompatible with Paolo Bozzi’s philosophical views as the other two brands I’ve discussed above, and as one might initially think.

32In fact, it is no secret that Bozzi admired Hume greatly, because of his clarity and, perhaps more importantly, his intellectual honesty. So much so that the epigraph at the beginning of Unità, identità, causalità has, next to a line from the ninth book of Aristotle’s Metaphysics (the enigmatic τὸ μὲν θιγεῖν καὶ φάναι ἀληθές), a quotation from Hume’s Treatise, and precisely from the section on probability and the idea of cause and effect:

  • 16 Hume 1739-1740: I.iii.2.

’Tis impossible to reason justly, without understanding perfectly the idea concerning which we reason; and “tis impossible perfectly to understand any idea, without tracing it up to its origin, and examining that primary impression, from which it arises. The examination of the impression bestows a clearness on the idea; and the examination of the idea bestows a like clearness on all our reasoning16.

  • 17 Helmholtz 1879: 23, 36 (Eng. trans.: 353, 360).
  • 18 Bozzi 1969: 86.

33But Bozzi also thought that Hume got a lot of things wrong. Specifically, he thought that Hume’s philosophy of perception was seriously mistaken, because grounded on a wrong understanding of the working of our visual system. Thus, with regard to Hume’s account of unity, Bozzi does not hesitate to take issue with the claim that every visible surface is really just an aggregate emerging from the juxtaposition of a number of minima visibilia (a claim that, after Hume, we can still find in Helmholtz, who famously wrote that “our eye sees all that it sees as an aggregate of color surfaces in the visual field” and “everything that is added in the intuition to the raw material of sensations can be resolved in thought”17). He takes issue with Hume’s (and Helmholtz’s) claim on empirical grounds. True, under normal stimulus conditions, different points of our retina’s surface are affected by different photochemical processes, one next to the other. But that is not to say that we see points (and unify them through the intellect). That the points are on the retina, or travel along the fibers of the optical nerve, does not make them visible events, immediate experiential data. They are just “physiological notions”18. Thus, for Bozzi, as for anyone who relies on a more advanced understanding of our physiology in the tradition of Gestalt theory (Max Wertheimer, but especially Wolfgang Köhler and Kurt Koffka), that is the wrong starting point for a good phenomenological analysis. A good analysis should first of all aim at a systematic understanding of the conditions through which the facts of direct experience present themselves to observation – conditions whose variation goes hand in hand with a (measurable) alteration in the very facts being observed. Perceptual experience is not merely a retinal business and must be taken in its full complexity. And when it is, Hume’s fictionalism about unity just does not follow (and what Helmholtz viewed as the result of unconscious inferences from the raw material of sensations turn out to be the very objects of experience).

  • 19 Ibidem: 140.

The job of a psychologist is to study the properties of events that are present in experience, as they manifest themselves, and to analyze their conditions so as to attain an ever more complete picture of the laws that connect specific constellations of stimuli to specific varieties of experience; it is not to derive, from physical and physiological premises, a picture of experience as it should be19.

34So, as a philosophy of perception, Hume’s was simply on the wrong track. And for Bozzi this applies to the issue of unity as it applies to identity and causation. But what about the metaphysics? Would Hume’s fictionalism – or any of in its recent incarnations – be equally unacceptable as a metaphysical stance? As I said at the beginning, Bozzi was a champion of realism, indeed a champion of so-called naive realism, so there is a strong temptation to answer in the affirmative and leave it at that. For a naive realist, fictionalism is just on the wrong track, not only as a philosophy of perception but as a metaphysical stance, and not only in its skeptic or Berkeleyan variant but in the Humean variant as well.

  • 20 Bozzi’s realism is most explicit in his 1990 book and in the essays collected in Bozzi 2007.

35Nonetheless, I want to resist that temptation. Bozzi’s realism was grounded in his unshakeable commitment to the view that the external world exists and is what it is independently of our conceptual schemes, our organizing activity, our conventions, and our intentions, whether individual or collective20. Those who knew him personally know how serious he was about this – serious and straightforward, as in the following often-quoted passage from his “Frammenti da opere perdute”, which Maurizio Ferraris and The Laboratory for Ontology chose as a slogan for the first edition of the Paolo Bozzi Lectures on Ontology:

  • 21 Bozzi 1999: 17 (reprint: 287).

If there is a black rock on an island, and if all people on the island have come to believe – through elaborate experiences and intensive use of persuasion – that the rock is white, the rock is still black and those people are idiots21.

There is, I think, no question that this sort of realism is incompatible with any metaphysics that rests on an ontology of white rocks, so to speak. And I think it’s fair to say that it is at odds, too, with those theories that doubt the existence of rocks altogether. That is why Bozzi would never have acquiesced in a fictionalist stance of the Pascalian or of the Berkeleyan variety. But what does this have to do with ontological fictionalism of the Humean variety? Surely, for the Humean the world exists and is what it is regardless of our conceptual schemes, our organizing activity, etc. Indeed, Humean fictionalism about unity, identity, and causation stems precisely from the realization that the “laws” governing our conceptual schemes and our organizing activity need not and arguably do not reflect the way the world is in and of itself. Isn’t that perfectly compatible with a realist metaphysics?

36It is not clear to me whether Bozzi’s views on unity are especially problematic in this connection. Granted, Hume’s theory of perception was seriously mistaken and the idea of minima visibilia betrays a bad understanding of the working of our visual system. But fix that. Would it be wrong to say that the term of unity is merely a “fictitious denomination” in those cases where the better theory of perception delivers verdicts to that effect? Would it be wrong to say that working out the better theory is to figure out more precisely what the fiction is in the first place? As a scientist of perception, Bozzi endorsed Berkeley’s esse est percipi wholeheartedly:

  • 22 Bozzi 1969: 179.

When it comes to perception, the reality with which we have to work is precisely the object as it appears in a given moment, under certain conditions22.

  • 23 Bozzi 1969: 180.
  • 24 Bozzi 1999: 8 (reprint: 276). The point is not, however, pressed in Bozzi 1969.

37But this amounts to a methodological use of Berkeley’s principle. Bozzi says so repeatedly in his writings. More importantly, he repeatedly warns his readers that the simplicity of the principle goes hand in hand with “the complexity of its possible applications”23. That’s where things get difficult, for a serious scientist of perception. And although Bozzi was explicit in claiming that metaphysical realism “follows” from the methodological esse est percipi (provided one takes into account “all of its consequences” and works through a “fine exercise of analysis and observation”24), it is not obvious to me that acknowledging the complexity of its applications is not a way of assenting to the spirit, if not the letter, of Humean fictionalism. You need to engage in a lot of psychology before you can draw any metaphysical inferences.

  • 25 All four quotations that follow are from Bozzi 1969: 264-266.

38But never mind that. When it comes to the metaphysic part of the story, the interesting question is not whether Bozzi would agree with Hume’s own way of cashing out his fictionalist stance with respect to each and every case (or with anyone else’s specific way). The interesting question is whether there is room for such a stance within Bozzi’s overall philosophy. And even if the above remarks will hardly suffice to point towards a positive answer with respect to unity, it seems to me that in other respects the answer is more definitely in the affirmative. By way of conclusion, let me try to substantiate this claim with reference to Bozzi’s views on identity. (Alas, the case of causation is too complex to be dealt with in a few paragraphs.) Indeed, let me do so by going step-by-step through Bozzi’s own recapitulation of the issue in Unità, identità, causalità25.

39The starting point is familiar enough:

Is it possible to assert that identity is a genuine property of the objects of experience, on a par with colors, geometric properties, weight, etc.? According to common sense, one can only answer in the affirmative.

Familiar, too, is the problem that arises immediately thereafter, though here Bozzi adds his own twist:

In order to truly say that A is identical to A, where A is some object, it would be necessary for the object in question to be one (si duo faciunt idem non est idem) and for it to be immersed in a reality in which time does not flow at all (for if time did flow, A would differ from one time to another for the simple reason that it would be associated with a different instant of that time). A rigid criterion of identity, therefore, makes it impossible to apply such a concept to the objects of experience. And if it is so, then it is true that we cannot step twice into the same river.

The twist is that for Bozzi the problem does not arise exclusively from the tension between numeric identity and qualitative diversity. (How can A at t be numerically the same as A at t’, if at t A is P and at t’ A is not P ?) The problem lies already in the fact that A is supposed to be an object, a particular, and particulars cannot be multiply located. They cannot, of course, be multiply located in space, which is how philosophers since Aristotle distinguish particulars from universals. But neither can they be multiply located in time. Otherwise they would not, with respect to time, be particulars at all; they would be universals – a hybrid metaphysical nature that flies in the face of intelligibility. So for Bozzi diachronic identity is not just hard to square with our common-sense intuitions. It is strictly and literally impossible. Which means that endurantism, the view according to which an object persists through time by being fully present at each time at which it exists, and which is supposed to embody the metaphysics of persistence implicit in common sense, is just not an option. Now,

As soon as we abandon the most rigorous definition of identity, which applies only to ideal objects, every other definition becomes more or less flexible and fluid. We have to acknowledge and accept this unfortunate condition, and we must study the methods that yield as little equivocation as possible. In practice, we have to see what sorts of transformations an object can undergo without losing its self-identity.

Obviously, the self-identity in question is not strict self-identity, for we have just seen that such a notion is practically useless. It applies only to ideal objects – not “ideal” as opposed to “imperfect”, such as a geometric circle vs. the actual circles we find in this à peu près world of ours, but “ideal” as opposed to “concrete”, i.e., embedded in space-time. Thus, what we have to do “in practice” concerns a notion of identity that is, strictly speaking, fictitious and up for grabs. Hence the conclusion:

In the end: either identity is accepted in the form of a logically pure principle, and then it is only valid for a world like Plato’s world of ideas. Or else it makes sense to speak of identity also in relation to a world like ours, and then the problem arises of accounting for the persistence of identity across transformations, or in the absence of transformations. And this problem belongs to the psychology of perception.

The last sentence is the crucial one. The problem belongs to the psychology of perception. What problem? Coming up with a good account of how we attribute identity to things that are, strictly speaking, not identical. Explaining under what conditions it makes sense to speak of identity in the presence of objects that identical are not. That is the problem. And it belongs to the psychology of perception because that is the field of research seriously devoted to the study of how we go about doing such things. Hume’s language was the generic language of a philosopher. He would speak of “propensity”, “inclination”, “tendency”, “disposition”. It is up to the psychologist of perception to tell us more and to explain exactly what goes on in such cases, and why it goes on that way. But note, that is the problem, not the answer. Bozzi is not saying that by studying the complex psychological mechanisms of human perception we can come up with an answer to the initial question, the question of whether it is possible to assert that identity is a genuine property of the objects of experience. We already know the answer to that question, and it is in the negative. Bozzi is not saying that a thorough investigation into the working of our perceptual system can deliver a way of vindicating the endurantist metaphysics of persistence that is implicit in common sense. That metaphysics is strictly speaking inconsistent, and that’s that. But there is a notion of identity that we are prone to use nonetheless, and we need to understand how and why we use it. It is a fictional notion of identity, but the fiction is important because we engage in it all the time.

40So there you are. This is why I think Humean fictionalism is compatible with Bozzi’s naive realism. Indeed, not only is it compatible with Bozzi’s realism; as I see it, it is the driving motivation of his work as a psychologist – surely his work on identity and arguably his work on causation, too, if not (also) his work on unity. The Humean fictionalist says that when it comes to the world out there, we must beware not to confuse the structure it really has with the structure we pretend it has. As philosophers, we must beware not to assert P when all we can say is, strictly speaking,

According to the fiction, P

Paolo Bozzi invites us not to overlook the real difficulty that lies behind this formula when it comes to applying it in practice: What fiction, exactly?

Torna su


Berkeley, G.
– 1710, A Treatise Concerning the Principles of Human Knowledge, Dublin: Rhames

Bozzi, P.
– 1969, Unità, identità, causalità, Bologna, Cappelli
– 1990, Fisica ingenua, Milano, Garzanti
– 1999, Frammenti da opere perdute, “Rivista di estetica”, n.s., 10: 3-24 (reprinted in Bozzi 2007: 269-296)
– 2007, Un mondo sotto osservazione. Scritti sul realismo (ed by L. Taddio), Milano, Mimesis

Goodman, N.
– 1983, Notes on the Well-Made World, “Erkenntnis”, 19: 99-107

Helmholtz, H. von
– 1879, Die Thatsachen in der Wahrnehmung, Berlin, Hirschwald; Eng. trans. by D. Cahan, The Facts in Perception, in Helmholtz’s Science and Culture: Popular and Philosophical Essays, Chicago, University of Chicago Press, 1995: 342-380

Hume, D.
– 1739-1740, A Treatise of Human Nature, London: Noon

Lewis, D.K.
– 1978, Truth in Fiction, “American Philosophical Quarterly”, 15: 37-46
– 1983, Postscript to “Truth in Fiction”, in his Philosophical Papers. Volume I, Oxford, Oxford University Press: 276-280
– 1986, Introduction, in his Philosophical Papers, Volume II, Oxford, Oxford University Press

Pascal, B.
– 1670, Pensées sur la religion et sur quelques autres sujets, Paris, Guillaume Desprez

Russell, B.
– 1913, On the Notion of Cause, “Proceedings of the Aristotelian Society”, 13: 1-26

Sider, T.
– 1996, All the World’s a Stage, “Australasian Journal of Philosophy”, 74: 433-453

Varzi, A.C.
– 2010, Il mondo messo a fuoco, Roma-Bari, Laterza
– 2011, Boundaries, Conventions, and Realism, in J. K. Campbell et al. (eds), Carving Nature at Its Joints: Natural Kinds in Metaphysics and Science, Cambridge (Mass.), MIT Press, pp. 129-153

Torna su


2 *Pascal 1670: III, §233.

3 Berkeley 1710: I, § 51.

4 From Lewis 1978.

5 This worry is addressed in Lewis 1983. The proposed solution, however, is based on the idea that an impossible story is one that is inconsistent, and when a story is inconsistent, several maximally consistent fragments can be extracted from it. Obviously, that would be to no avail to the Berkeleyan fictionalist.

6 Hume 1739-1740: I.iii.6.

7 Hume 1739-1740: I.iv.6.

8 Ibidem.

9 Hume 1739-1740: I.ii.2.

10 Bozzi 1969.

11 In Lewis 1986: ix.

12 Goodman 1983: 104.

13 See Sider 1996.

14 Russell 1913: 1.

15 For example, in Varzi 2010 e 2011.

16 Hume 1739-1740: I.iii.2.

17 Helmholtz 1879: 23, 36 (Eng. trans.: 353, 360).

18 Bozzi 1969: 86.

19 Ibidem: 140.

20 Bozzi’s realism is most explicit in his 1990 book and in the essays collected in Bozzi 2007.

21 Bozzi 1999: 17 (reprint: 287).

22 Bozzi 1969: 179.

23 Bozzi 1969: 180.

24 Bozzi 1999: 8 (reprint: 276). The point is not, however, pressed in Bozzi 1969.

25 All four quotations that follow are from Bozzi 1969: 264-266.

Torna su

Per citare questo articolo

Notizia bibliografica

Achille C. Varzi, «Fictionalism in Ontology: The 2011 Paolo Bozzi Lecture»Rivista di estetica, 56 | 2014, 253-270.

Notizia bibliografica digitale

Achille C. Varzi, «Fictionalism in Ontology: The 2011 Paolo Bozzi Lecture»Rivista di estetica [Online], 56 | 2014, online dal 01 juin 2014, consultato il 15 juin 2024. URL:; DOI:

Torna su


Achille C. Varzi

Articoli dello stesso autore

Torna su

Diritti d’autore


Solamente il testo è utilizzabile con licenza CC BY-NC-ND 4.0. Salvo diversa indicazione, per tutti agli altri elementi (illustrazioni, allegati importati) la copia non è autorizzata ("Tutti i diritti riservati").

Torna su
Cerca su OpenEdition Search

Sarai reindirizzato su OpenEdition Search