Robinsonade of fictional objects and the poverty of arbitrary objects

Full Text

Introduction
The present paper is a study of the concept of fictional objects, according to which those are understood as epistemically finite with respect to their special properties. Fictional objects play a key role in fiction, and they should be distinguished from real-life prototypes even when they seem to be practically the same. The mentioned concept of fictional objects is contrasted with the concept of arbitrary objects. This material was presented in the form of a talk at the 7th Lem’s Readings (Samara University, 28-30 March 2024), after which the participants asked a number of interesting questions and made comments, which are taken into account in the present paper. The paper has the following structure: (i) the conception of fictional objects as epistemically finite with respect to their special properties; (ii) the conception of arbitrary objects in its comparison with the mentioned concept of fictional objects; (iii) a comparative table for the two mentioned conceptions; (iv) answers to the questions and comments made after the talk at the 7th Lem’s Readings.

Fictional objects
The hero of Marcel Coscat's fictional ‘Les Robinsonades’ from ‘A Perfect Vacuum’ (Lem 1979), Serguis N., or the New Robinson, cannot create the Supreme Being, but becomes a creator himself. Disappointed by the fact that the people around him in his life before the shipwreck were rather ‘murky and turbid water’ (while ‘a man without Others is a fish without water’), he exclaims the first word of Genesis – ‘Away with this clutter!’ and begins to create. First he creates the fat man Snibbins, his faithful servant, a butlet, valet and footman, whose existence is not to be doubted, for ‘to doubt it is to doubt that the trees stand and the clouds float when no one is watching them’. Of course, Snibbins eats nothing, but cooks his master's own food while he gathers a collection of interesting stones (‘how easy on the budget and how convenient’). However, the New Robinson gradually grows dissatisfied with Snibbins's over-executiveness, Snibbins cannot be changed, ‘the wat he is, is the way he is’. So he creates a cook named Boomer. Due to the need to keep an eye on Boomer, Snibbins begins to perform his duties somewhat worse, so Robinson has to kick him out after paying him three months' wages (the money is still stored on the reef-wrecked Patricia). Snibbins leaves, but he does not take the money. Finally, Robinson creates for himself a three-legged servant girl, Wendy Mae, who would become his Beatrice. The unhappy and extremely confused love for Wendy Mae forces the New Robinson in the end to create ‘those throngs which Robinson calls into existence off the cuff, carelessly, writing names, first and last, and nicknames on whatever comes to hand’ (Lem 1979, p. 22).
The new Robinson surrounds himself with fictional objects, creates Others who are his boundary: ‘If there is no God and if, moreover, there are neither Others nor the hope for their return, one must save oneself through the construction of a system of some faith, a system that, with respect to the one creating it, must be external’ (Lem 1979, p. 24). Inaccessible beings (‘beyond reach’), whose existence, however, is not in doubt, are more appealing. Inaccessibility is required for Robinson in order not to be able to realise their non-existence. It is important for him not to be able to be convinced of the non-existence of his beloved Wendy Mae, so he invents a whole society on the island that ‘will stand between him and the girl; this that will throw up a system of obstacles and thus provide that impassable distance from which he will be able to love her, to desire her continually [...] He realizes – he must – that [...] if he attempts to feel her, the whole world that he has created will, in the bat of an eye, crumble’ (Lem 1979, p. 25). Robinson even ‘forgets’ about Wendy Mae's third leg, which initially served as an insurmountable obstacle to their intimate relationship.
As a creator, New Robinson is, of course, somewhat free. He creates fictional characters (and Wendy Mae in particular) with the sole requirement that he must not disbelieve their (more importantly, Wendy Mae's) existence. However, these characters ‘live’ their own lives, they are like Snibbins, the way they are, is the way they are. Although New Robinson may make adjustments to individual characters in his fictional society (and these changes are unlikely to be entirely conscious) or simply forget some of their peculiarities, the characters remain themselves. Adjustments are rather detrimental to them.

Fictional vs. real objects
From a philosophical point of view, when talking about fictional objects in order to contrast them with real ones, one should pay attention to the distinction (Rescher 2003, p. 341) between generic and non-generic properties of an object. Generic properties are properties possessed by all members of some (often natural) class. In other words, these are properties that constitute the essence of some kind of object. Non-generic properties are properties possessed by the unique object, that is, these are properties that it does not share with other objects of its class. To illustrate this separation, N. Rescher considers a single snowflake (Rescher 2003, p. 341). The generic properties of some particular snowflake are the hexagonality of its shape, its general chemical composition, its melting point, etc. Among the non-generic properties of a certain snowflake are its particular shape, angular momentum of its fall, etc.
Actually, the difference between fictional and real objects, when interpreted through generic and non-generic properties, is that fictional objects (provided that the ‘act of creation’ is already completed and no more changes are made to the created fictional object) have a limited number of non-generic properties, while real objects turn out to be ‘cognitively inexhaustible’ (Rescher 2003, pp. 341–343) or possessing immeasurable cognitive depth. For example, it is possible to make a detailed description of A. C. Doyle's Sherlock Holmes, the source of which will be stories, drafts, correspondence and other contents of the author's entire archive. However, this detailed description will be finite; at some point in the compilation of such a description, it will be so ‘complete’ in the context of a particular set of sources that one can add to it only by appealing to the power of imagination. In fact, we do not know very much about Sherlock Holmes. There is no even a clue as to his date of birth (the probable year is 1854), no information about his parents (not even their names). We don't know, after all, how many atoms were in Sherlock Holmes' body at a certain moment of time (for example, at the moment when he saw Professor James Moriarty for the first time), and we will never know it. We have no other sources of information about Sherlock Holmes (namely, A.C. Doyle's Sherlock Holmes) besides the writer's works and archive (perhaps some recollections of A.C. Doyle's relatives and friends could be added here, if we could consider them reliable enough). We obviously cannot meet Sherlock Holmes, and therefore we cannot study him as any really existing object.
Of course, we should admit that our knowledge about some existing object at any moment of time is finite, but we have a principal possibility to extend this knowledge. And it concerns both knowledge of generic and non-generic properties of this object. It may not be said about Sherlock Holmes that he has all the properties that characterise human beings, those properties that all people share, but if in the process of research it turns out that all human beings have some property unknown till now (for example, at some moment we have found out that all people have somewhat similar DNA structure), then we will conclude from it that Sherlock Holmes, as a human being, albeit a fictitious one, also has this property. However, our knowledge of the non-generic properties of a fictional object is fundamentally incomplete, no matter what studies of existing objects we conduct. For real objects, not only is our knowledge of generic properties fundamentally extensible, but so is our knowledge of non-generic properties, and it does not require a recourse to the aid of our imagination. In other words, real objects are cognitively inexhaustible, and what is ‘exhaustible by linguistic characterization would thereby be marked as fictional rather than real.’ (Rescher 2003, p. 341).

Arbitrary objects
In addition to real and fictional objects in the context of generic and non-generic properties, one may consider arbitrary objects. The concept of arbitrary objects in modern metaphysics has been developed since the 1980s by authors such as Kit Fine (Fine 1985), Marco Santambrogio (Santambrogio 1987), Leon Horsten (Horsten 2019) and others. An arbitrary object is an object that possesses all generic properties of some class, but does not possess any non-generic properties.
A similar idea concerning variables was held by some 19th century mathematicians who interpreted statements of the form ‘Let x be a natural number’ as statements about an arbitrary number. However, this view was criticised in the second half of the 19th century, and the concept of arbitrary numbers (respectively, the concept of arbitrary objects in general) was abandoned. In particular, we may note the criticisms made by Gottlob Frege in his article ‘What is a function?’ (Frege 1904), in which he criticises the idea of arbitrary numbers in Emanuel Czuber's ‘Lectures on Differential and Integral Calculus’ (1898). From Frege's point of view, there are no indefinite or variable numbers, and expressions like ‘Let x be a natural number’ should be interpreted in such a way that the variable x here means some definite natural number, but at the moment we simply do not know which one. In other words, according to Gottlob Frege, the arbitrariness in this case is rather a characteristic of the relation of reference, not of the object. Be that as it may, since the 1980s the concept of arbitrary objects has proved fruitful, for example, for the development of structuralist philosophy of mathematics, the concept of reasoning with indefinite objects, the semantics of natural language, epistemic justification for infinitary reasoning, etc.
It has been stated above that arbitrary objects possess all generic properties of some class and do not possess any non-generic properties. At the same time, they also possess special properties of arbitrary objects – for example, the property of being some definite object in some particular situation (more precisely: being in the state of being some definite object in some particular situation) or being a diagonal arbitrary object (that is, an arbitrary object which is a different definite object in different situations, and at the same time it is not the same definite object in any two different situations).

Comparison
Arbitrary, fictional and real objects can, using the apparatus of the semantics of possible worlds, be compared also by which possible worlds they exist in and how those objects are realised. It should be clarified that possible worlds are understood here not just as parts of a model, but in a broader metaphysical sense (whatever it may be). Thus, (1) real objects necessarily exist in the actual world and can exist in an indefinite set of possible worlds (this indefinite set of worlds is half of the whole set of possible worlds, since an object in a possible world either exists or does not exist); (2) fictional objects do not exist in the actual world, but exist in an indefinite set of merely possible worlds (fictional objects also exist in half of all possible worlds and do not exist in the other half of them); (3) arbitrary objects exist in any possible worlds where there is at least one definite object of the corresponding class, but no arbitrary object in any possible world is any definite object (it does not take the ‘value’ of any definite object), since for arbitrary objects the essential modality is the afthairetic modality, according to which an arbitrary object in some situation can be some definite object, but at the same time it is not actually a definite object in any situation, an arbitrary object always remains arbitrary.
The above considerations can be summarised using the following Table 1. 

If we return to the title of this paper, it should become clear that the ‘robinsonade’ of fictional objects in this case means existence only in purely possible worlds and being limited in their non-generic properties by the creator's imagination, and the ‘poverty’ of arbitrary objects means the absence of any non-generic property and non-realization in any possible world in the form of any definite object.

Conclusion
One important remark on the concept of fictional objects as epistemically finite with respect to their non-generic properties is to point out that objects that have ceased to exist, like fictional objects, are epistemically finite. Indeed, it must be recognised that, for example, the ancient Greek historian Thucydides cannot be directly accessible to us to conduct a research, and all we know about him relies on a finite set of written sources. Accordingly, Thucydides, must be recognised as a fictional object. This is a rather strong observation, to which it seems to be possible to respond in two ways. First, the metaphysical position of realism regarding the past is certainly problematic (at least I am not aware of any consistent realist position regarding the past). (What is meant here by realism about the past is the conception that the past exists really and independently of cognitive agents, alongside the present or in the present. At the same time, the supposed anti-realist position on the past does not imply that the past never existed, it just does not exist in the present.) The past is inaccessible to us directly in epistemic terms and is known only indirectly. Within metaphysics, the most consistent anti-realist position on the past (e.g., in the fashion of the verificationists) seems to be one that assumes that past objects do not exist independently of cognitive agents. Secondly, if the anti-realist position on the past seems unsatisfactory, then the concept in question itself can be reformulated so that it is not a concept of fictional objects, but an explication of at least one aspect of what it means to really exist (provided, of course, that past objects are recognised as non-existent) or really exist in the present. In other words, it is not only fictional objects that are epistemically finite in their non-generic properties, but all non-existent objects in general. In this case it is required to separate non-fictional and non-existent objects from fictional (but also non-existent) objects in some other way.
Another remark is that fictional objects may have a normative dimension, which should also be taken into account in the concept of fictional objects. It seems that the concept of fictional objects as epistemically finite with respect to their non-generic properties does not contradict normativity. Fictional objects, as they are described at the beginning of the article, quite possess the required normative dimension; after all, it is the fictional society (in the hero's fantasy) that interferes with the love relationship between New Robinson and Wendy Mae. Since fictional objects may or may not play a normative role, it seems that the concept of normativity of fictional objects is external to the concept of fictional objects proper. The concept of fictional objects under consideration is quite compatible with a suitable concept of normativity. At first glance, only those conceptions of normativity that assume the real existence of objects playing normative roles are not suitable.
Furthermore, in the discussion of the concept in question, one may raise a question of how to interpret situations in which interaction with fictional objects takes place. For example, based on the film ‘Swiss Army Man’, we can assume a situation in which a real person interacts with a real corpse, but believes that it is not a corpse, but another living person. It seems that such a situation in the context of the considered concept of fictional objects can be interpreted as follows. Firstly, there is a person interacting with the corpse in the actual world. Thus we have two real objects. Secondly, this person imagines (or hallucinates) communication and other variants of interaction with what in the (other) possible world ‘takes’ the place of the real corpse. Thus, we can speak of a possible world in which practically everything happens in the same way as in the real world, but in which our real person interacts not with a real corpse, but with another person who ‘takes’ the place of a real corpse in that possible world. To correctly predicate relations between the person in actual world and the other person (in the role of the corpse in actual world) one needs a sematic framework compatible with a cross-world predication (e. g., (Borisov 2023)). The concept of fictional objects under consideration is quite consistent with this interpretation.

About the authors

Lev D. Lamberov

Ural Federal University named after the first President of Russia Boris Yeltsin

Author for correspondence.
Email: lev.lamberov@urfu.ru
ORCID iD: 0000-0001-9228-4909

Candidate of philosophy, associate professor, associate professor of the Ontology and Theory of Knowledge Chair

Russian Federation, 51, Lenin av., Yekaterinburg, 620002, Russian Federation

References

Supplementary files