havel@cts.cuni.cz

Human beings have a special gift of being able to invent stories that never happened, imagine things that do not exist, consider facts that are not valid, and easily talk about situations that are next to impossible. We do it all the time—and I even do not know why I call it a special gift. Perhaps because we don’t expect this ability from animals and stones.

There is no sharp dividing line between states-of-affairs (stories, things, facts, and situations) that we are inclined to qualify as real or actual, and states-of-affairs possibly invented by us, which I shall call conceivable. So far as our mental inventions are constructed in order to say something relevant to the world in which we live, and because of the mentioned lack of a dividing line we can think about the conceivable and the real blended together into one conceivable world. We should not be surprised that such a conceivable world is almost certainly prone to be inconsistent. (However, we may be able, at least conjecturally, to dissect it into various fuzzy domains of consistency).

The term ‘world’ is used in modern philosophy with various meanings. In one it is the aggregation of everything that exists. Another meaning is derived from a certain life philosophy, certain understanding of experience, and may be related to a particular environment, culture, or epoch (thus we can talk, e.g., about the world of science, world of myths, world of Eskimos, and with a certain deliberation, about the world of ants). In this paper I prefer more or less the latter sense (but taking into account the specific use of the word 'world' in the phrase 'possible worlds').

Even if the concept of conceivability is central in this essay I do not intend to explicate it in a formal way. Instead, I would like to point to some of its cognitive aspects that can be illustrated (mainly) in three different examples of conceivability with intended or established diversion from (putative) reality, namely thought experiments in physics, a hypothetical world devoid of consciousness (zombie world), and virtual reality.

I believe that what is most interesting concerning the concept of conceivable worlds, and conceivability as such, is neither what is explicitly conceived as differing from our real world, nor what is (not always explicitly) taken from our world, but what is ignored. Related to this is the so-called coherence gap problem which I will somewhat vaguely formulate below.

Furthermore, I shall touch upon the issue of the participation of the conceiver of a hypothetical world in the world conceived. I will call it the spectator problem. I am in favor of the participatory thesis even if it calls for a "softer" concept of the real world which would admit replacement of certain of its fragments with imaginary situations.

In science thought experiments play a significant heuristic, inferential and speculative role. There are a number of famous thought experiments in physics, for instance Galileo’s argument on falling bodies, Newton’s rotating backet, Huygens’ proof that equal elastic masses exchange their velocities on impact, Maxwell’s demon, Einstein’s elevator, Schrödinger’s cat. (We can find thought experiments in other fields, last but not least in philosophy, as we shall comment upon soon).

Thought experiments, as means of acquiring knowledge, make, in a way, a bridge between logic and empiricism, between deduction and induction, between testing reason and testing the world. They are of various types. Some peer out of the known towards the unknown, others play with the unknown to peculiarize the known. Some are destructive, others constructive—the former usually lead to a paradox or contradiction in order to call into question a proclaimed hypothesis, the latter illustrate or support an advanced hypothesis, or stimulate us to a discovery of an entirely new hypothesis. Many thought experiments serve just as a substitute for prohibitively expensive, lengthy, or painful real experiments.

All thought experiments are, in a sense, of the same making: from the material of our real world, i.e. from our theories, concepts, and facts, a new hypothetical conceivable world is constructed (which may, of course, sooner or later turn out to be incoherent or inconsistent). During the construction, some concepts are given a new meaning, some theories are replaced by others, and some facts from our world are omitted or added or modified. What appears to be a crucial issue is the word "some" in the previous sentence, as I shall indicate soon.

Galileo’s thought experiment is destructive (for Aristotle’s hypothesis) and at the same time constructive (for Galileo’s hypothesis). That is why it is so famous. Its conclusion is evident and seemingly nothing crucial was omitted. Yet something important was forgotten. There may be other ways how to escape from the paradox. What if we lived, say, in a world where two or more bodies after being physically attached change their properties in a certain strange way, for instance by increasing their joint speed of falling. Or what if we lived in a world where attaching bodies together is forbidden by unknown laws of nature? It appears that many hypotheses about our world are taken for granted and transferred to the constructed hypothetical setting without explicitly noticing it.

Again, once allowing ourselves such an absurd world of a solitary backet, we are free to invent scores of odd explanations. For instance, what if the backet is, say, in a quantum superposition of two different states that collapses into one of the states whenever we look at it (from god knows where)?

In conceivable worlds of thought experiments, some states-of-affairs are, by design, the same as they are in our world, while other states-of-affairs are deliberately different. The crucial but often neglected feature of these worlds is that we seldom know what is the extent of the domain of "the same" and what is the extent of the domain of "the different", besides what is explicitly mentioned or used in the construction. Moreover, besides these two domains there is an inexhaustible realm of states-of-affairs that are omitted because they are believed to be irrelevant or because they are forgotten, obscured, or entirely beyond the reach of human knowledge.

The real or a conceivable world or some of its fractions is coherent if all of its parts fit smoothly together in a natural or imposed agreement and with no internal incompatibilities or deficiencies. Thus by coherence I mean a property of a certain totality of things and states-of-affairs, that expresses our more or less intuitive presumption that those things and states-of-affairs are blended into a harmonious whole. The real world may be considered coherent by its own nature, i.e., its coherence is inherent and a possible contingent inconsistency in its description would be a mere consequence of the imperfection of our language of description or of our understanding of the world.

Indeed, such coherence cannot be taken for granted in conceivable worlds due to the mentioned realm of omitted states-of-affairs, which may be a source of various unexpected inconsistencies. Let us call this omitted realm the coherence gap (which does not mean incoherence yet!) and the question of its possible impact on our reasoning the coherence gap problem.

Due to this coherence gap, we have to be very cautious before we start making inferences in (and with the help of) conceivable worlds. For instance: does it make sense to talk about rest or uniform motion of bodies in the absence of forces impressed upon it if we live in a world in which everything interacts with everything else? Is it possible to talk about falling bodies in a void space if there is no void? Or about a rotating body with the rest of the physical universe gone? Is it possible to talk about anything at all when the observer is gazing from nowhere?

Philosophers of an analytic breed often use thought experiments in their arguments, for instance, when they discuss the nature of the human mind. Thus we have the well-known brain-in-the-vat experiment, Searle’s Chinese room, the zombie world, and many others. Whenever human consciousness is in question the coherence gap should be particularly taken into account. Let us mention for example the world of zombies—a conceivable world devoid of consciousness currently discussed by (western) philosophers of mind.

David Chalmers uses a logical possibility of such a zombie world as one of the arguments against reductive explainability of consciousness. But can we really imagine a world inhabited by creatures physically indistinguishable from us, creatures who put on our intelligent faces and chat about what we would consider intelligent topics (including chats about "inner feelings" and "conscious experiences", even about "conceivability of a zombie world"), yet creatures who themselves lack conscious experience entirely? To Chalmers the logical possibility of zombies seems obvious, similarly as, say, the logical possibility of a mile-high unicycle. He claims that "a coherent situation is described", he discerns "no contradiction in the description", and does it with reference to a "brute intuition".

I have no problem with a description of a conceivable world in ordinary language (with vague and context-dependent meaning of words), nor with Chalmers’ reference to a brute intuition—after all, thought experiments are intended precisely to address our intuition. However, I see the problem of the coherence gap lurking around, especially when the idea of a (single) zombie is expanded to include all inhabitants of the hypothetical world.

Such a gap is, first of all, a natural consequence of the vagueness of words (e.g., of the word ‘unicycle’ above which may or may not entail usability as a real vehicle, it may be meant as a vehicle for cyclists of our size or of a corresponding size, it may or may not be assumed to be made from the same materials as is used for cycles in our world, etc.). The coherence gap is also a consequence of our intended or unintended omission of seemingly less relevant details in the constructed situation. Lastly, the gap is the outcome of our insufficient knowledge of our own world.

The zombie world is certainly conceivable but if we want to use it, for instance, in the argument about the logical independence of consciousness on the physical world, we would have to say something more to address the coherence gap. In particular, we should inspect all aspects in which the conceivable world may differ from the real world in addition to what was explicitly stipulated (that no conscious experiences exist). Otherwise some hidden inherent inconsistencies might be overlooked.

Consider, e.g., the sound waves caused by vocal chords of our zombie twins in the zombie world. The waves either have or do not have a linguistic meaning (to decide this would require certain assumptions about the nature of language in the zombie world and/or about the reality of linguistic meanings in general). In the first case, there would be a difference from our world in the semantic domain since in the world without consciousness there would be no distinction between sincerity and insincerity, nor between an intended lie and mere falsehood. In the second case we might face certain problems with the concept of causal efficacy of conscious decisions.

The main difference is, however, the fact that the one poor zombie who happens to be my own twin (to the last quark) in the zombie world cannot possibly be me.

This brings me to the important issue of the presence or absence of the observer in the conceived situation (of myself, if I am the conceiver). Am I a participant in the conceived world or am I merely a spectator from elsewhere? (If so then from where?). I believe that this issue—let us call it the spectator problem—seems to be philosophically even more essential than the coherence gap problem. Any purely logical explication of mentalistic concepts (to which conceivability certainly belongs) is vulnerable to the objection of leaving something out, namely the observer. This applies, in particular, to the idea of the space of possible worlds. Various proposed remedies like Quine’s concept of centered worlds may only hide the main issue.

Without going into the burden of arguments, I would like just to mention my conviction that the participatory approach is more appropriate if not necessary. If I cannot subtract my own self from the world entirely I shouldn’t try to do it at all. That means that this or that conceivable alternative world cannot be kept completely apart from my own world, which I consider (at least methodologically) to be real. Instead of implanting parts of the real world into another, completely new hypothetical world, I would rather look around in my real world as if certain selected states-of-affairs were tentatively "replaced" by some counterfactual states-of-affairs. (In a special case I could include even myself, at least to a certain extent, in that counterfactual domain. For instance, while sitting at my desk, I can easily imagine myself walking outdoors and talking to non-existent people.)

In the participatory approach I would not feel the urge to fill the coherence gap with new inventions. The gap would be filled in a natural way by the states-of-affairs of the real world and I would resolve potential incompatibilities only in the case when they might threaten the purpose of the construction. Any conceivable world, conceived by me, would be always inhabited by me, its conceiver. Consequently, it would be different, at least as to the conceiver’s identity, from all of somebody else’s conceivable worlds.

Under this view the idea of the zombie world would raise a serious problem. I can imagine all other people being zombies, indeed, but I cannot imagine myself being a zombie. Granted, I can be sometimes in a coma, but while being in a coma I cannot imagine anything, including myself being in the coma.

As a matter of fact, we all live in conceivable worlds. We construct them all the time without any dread of contradictions or inconsistencies, most often without noticing them at all. Indeed, even the most orthodox metaphysical realist would admit that certain aspects of the world of our daily experience are adapted by our (individual or collective) minds even if we behave as if they were independent aspects of the actual world. Beginning with linguistically-driven perception, through visual distortions and various illusions, and ending with hallucinations and dreams, all these are obvious cases of mixing our constructed worlds with reality.

In a way human perception has something in common with using thought experiments: a certain situation is comprehended with the help of an assortment of hypotheses, some of them constructed ad hoc, others chosen from a certain body of established facts, some plausible, others rather daring. Every particular hypothesis is not always open to verification and therefore the "danger" of contradictions has to be ignored or tacitly admitted. In fact, perception is a continuing dynamical process open to novel situations of various degrees of unexpectedness, from entering a room for the first time to the exploration of worlds behind the horizon of everyday experience (the horizon of vision, of scales, of familiarity, of the suitability for linguistic description, etc.)

We are prepared to change such hypotheses, sometimes more sometimes less, similarly as we are prepared to change our "default" expectations when entering a friend’s house for the first time.

People sometimes construct rather sophisticated conceivable worlds for certain specific purposes. Thus a defense lawyer, by utilizing gaps in testimonies, invents an alternative story of the crime in order to weaken the charge against his defendant. Analogously a sinner fabricates his (or her) story of complete innocence, and may do this so thoroughly that he (she) starts sincerely believing it.

I maintain that it is the fundamental nature of the human phenomenal world that it permits, perhaps even invites us, to graft into it "locally" our expectations, imaginations, conceptions, beliefs, hypotheses, fantasies, petty lies and contradictions without impairing drastically the overall coherence of the world.

Let us turn to a possibility of actual "realization" of conceivable worlds, namely the empirical technique called virtual reality (VR). This technique has not yet been given the attention by philosophers that it would deserve.

Let us assume, for the sake of this study, an ideal case of unlimited quality (precision, speed, coordination) of VR technology. It is not necessary to deal with technical details in order to pose some more philosophically relevant questions. For instance, the question about the nature of a cybernaut’s experience when he is immersed in the virtual world while keeping his real body. Brenda Laurel comments,

There are even more general issues. Is there any threshold between being merely a spectator and being an actual participant? What is the nature of the real world insofar as it can be (should be?) distinguished from the virtual world?

Let us call the virtual world the sum of everything that forms the individual environment of the cybernaut, everything that he can perceive and experience while immersed in the VR setting. We (from the outside!) can distinguish three components of such a virtual world which differ in their relationship to the real (not computer-generated) world. The first component is genuinely real; it comprises things, states-of-affairs, and features in which the virtual world overlaps with the real world: besides the cybernaut himself (his brain, body, bodily positions, movements) there is also his datasuit, and whatever happens not to be a product of computer simulation (for instance the floor or a real object foisted into the cybernaut’s hand). The second component is realistically virtual—things, states-of-affairs, and features that are computer-generated with the aim to be as similar to the actual reality as possible, e.g. the positions and movements cybernaut’s hands (as projected into his goggles), some chosen properties of virtual objects (so that they seem to behave realistically, e.g. occlusion, perspective, maybe gravity). The third component is unrealistically virtual—things, states-of-affairs, and features that are computer-generated to deliberately differ from reality: imaginary fairy-tale landscapes, things without gravity, flying pigs, and possibly enhanced cybernaut’s capabilities (to lift heavy objects, to fly, etc.).

Of course, distinguishing between these three components in concrete cases may not be easy and for the cybernaut expectedly impossible. Their relative proportion may significantly vary from case to case and also within the same virtual world; all three may be present in a single entity (e.g. when an image on the screen with realistic shape but unrealistic color or weight is associated with a genuine object in cybernaut’s hands). I think that the most enjoyable experience may be when a very little bit of unrealism is added to a very trustworthy environment (this is, in fact, the sleight-of-the-hand trick of magicians).

Our perception of reality is very much dependent on human collective experience. When other people coherently talk about, and/or interact with something unfamiliar or invisible to me, it is easier for me to consider it as real. There is no technical problem to immerse many cybernauts into properly coordinated virtual environments. Each of them would see the others’ virtual bodies in their realistic positions and with realistic behavior, and they could (realistically) communicate about experiences. Yet each of them would "live" in his own individual virtual world generated by his "personalized" software (whose coordination would be a task of an underlying operating system).

I propose to treat the case of VR in a similar theoretical way as we did with the thought experiments. In particular, we might consider the two problems mentioned above, namely the coherence gap problem and the spectator problem, in connection with VR.

A rather challenging issue in the context of virtual reality is its coherence. Consider the (real) world of the designer of VR. Among many other real things, it includes the computer with the VR program. The program is made to generate a conceivable world, namely the virtual world for the cybernaut. While the computer and program as such are anchored in the inherently coherent reality, the generated conceivable world is necessarily limited due to the limited knowledge of the programmer. Thus there exists an analogy of coherence gap mentioned earlier.

A cybernaut, on the other hand, immersed in the virtual world may find it difficult to encounter anything suspicious to him (we may assume that the virtual world is generated by a sufficiently sophisticated program and that the cybernaut is exposed to it for such a long time that he would take his world for real). Of course, he may also be as suspicious about the correctness of his perception as any of us can be suspicious about our perception of the real world.

In spite of all that, we, as detached theoreticians, can contemplate a difference on behalf of the cybernaut. Let us imagine that the cybernaut is performing an unlimited series of experiments in his virtual world aimed to dig deeply into the depth of the functioning of his world. No program could be designed to generate prompt coherent outcomes for all possible experiments and thus sooner or later the cybernaut has to encounter the coherence gap (maybe without even noticing anything).

Virtual reality techniques render an interesting testing ground also for the spectator problem. Unlike thought experiments that arrange a certain new situation to be inspected so to say "from outside", the virtual reality experiments literally drag the cybernaut into the virtual world. He may act in it and his actions induce adequate responses, at least when the realistically virtual component is prepared for that. Thus the cybernaut is more a participant than a spectator. However, his very participation may be played with in various ways.

Jeffrey Shaw of Center for Art and Media in Karlsruhe produced some time ago a striking installation. He placed a revolving platform with a TV monitor and a chair on it into the center of a room. The viewer sits in the chair and by leaning to the left or right he can turn the whole platform in any direction he decides. On the monitor he can see the same room, the same platform with the monitor and chair, and the same movement, as viewed from a vista point somewhere behind his back. The image of the background is moving correspondingly with the movement of the platform so that the viewer can see it in the same way as he sees it in reality. Except one crucial difference: the chair is empty.

How fascinating it may be to find yourself in a world where you are a participant—but absent! For us, in the age of science, it is perhaps a normal and important thing called objectivity. An astrophysicist studying the evolution of galaxies, a geologist studying continental drift, or an archaeologist studying the migration of cultures knows this very well. Granted, stories they tell and words they use depend very much on what they and we consider interesting and meaningful but this does not change the fact that we are not personae (have no roles) in these stories. This view of a scientist as a spectator rather than participant has become somewhat eroded nowadays in various areas, whether in modern physics, psychology, sociology, or art theory.

If we admit the spectator view in astronomy, geology, or archaeology, and at the same time invite a participatory view in sociology or psychology (physics is in dispute about this) then we face a question whether we can talk at all about a unified scientific world-view. Let us leave this question open—I am afraid that a premature answer may lead to narrowing the scope of the term ‘science’ rather than to some grand unification.

The aim of this paper was to touch upon certain issues of conceivability and conceivable worlds in a preliminary and tentative way rather than to make categorical claims or put forth specific proposals. A deeper analysis should follow, namely of what we actually do (and what we think we do) when conceiving hypothetical worlds partly similar and partly dissimilar to the world of our natural experience. Or alternatively, what actually makes us so confident about the reality and coherence of the world that we call "real". Philosophical phenomenology may turn out to be a useful tool in this direction of study.

I have used the term ‘conceivable’ in a rather wide and nontechnical sense. I prefer it to the term ‘possible’ for at least two reasons, first, not to induce a direct connection with various explications of the term ‘possible’ in philosophical logic, and second, to remind that conceivability has some relation to the mind of a person, the conceiver. I believe that our comparison of thought experiments in physics or in philosophy on the one side with "real" experiences in a virtual environment on the other side may help for further insights into the relation between intended and imposed modifications of reality.

There are other variants of conceivability besides those discussed in this essay. Further analysis might take into consideration various cases of consciously uncontrolled human experiences (illusions, dreams, hallucinations, etc.), as well as the essential human directedness towards the future with openness to a spectrum of alternative future situations.

What applies to a single person’s constructions of conceivable worlds can be mutatis mutandis applied also to collective constructions, as I already tried to hint at certain places above. Even the currently dominant world-view based on the combination of various scientific conceptions with "naive" views of reality, which therefore cannot avoid various incompatibilities, inconsistencies, and gaps in coherence can be analyzed in an analogous way as one can analyze thought experiments or virtual reality.