A revised version was published in the book Lines of Thought: Rethinking Philosophical Assumptions (Cambridge Scholars Publishing, 2014)
THE ULTIMATE ANSWER TO THE SCEPTICS TOGETHER WITH THE ULTIMATE PROOF OF THE EXTERNAL WORLD
Philosophy unravels the knots in our thinking; hence its results must be simple, but its activity is as complicated as the knots that it unravels.
The so-called argument from ignorance concerning the existence of the external world is one of the most puzzling sceptical arguments ever conceived. To understand how it works, we must first present some examples of general sceptical hypotheses about the external world. Here are a few:
1. I am dreaming the external world.
2. I am hallucinating an external world.
3. I am a soul deceived by a malign genie who makes me believe that I live in this world, which actually does not exist (the Cartesian version).
4. I am a brain in a vat, connected to a supercomputer that makes me believe I live in a real world, whereas what I am really experiencing is merely a virtual reality program running on the supercomputer (the main contemporary version).
Typical of such sceptical hypotheses is that it is at least logically possible that they are true, since it seems impossible to conclusively refute them. Indeed, it seems that we are not even able to know that they are false.
Now, consider the trivial statement ‘I have two hands’, and the sceptical hypothesis ‘I am a brain in a vat’. Using them, we can already conceive the following instance of the argument from ignorance about the external world:
1. If I know that I have two hands, then I know I am not a brain in a vat.
2. I don’t know that I am not a brain in a vat.
3. Therefore: I do not know whether I have two hands (1, 2 MT).
Indeed, if I cannot know that I am not a brain in a vat, how can I know that I really have two hands?
Since the statement and the sceptical hypothesis can change, let p be any statement about the external world, including even trivial ones like ‘I have two hands’, ‘This table exists’, ‘There are stones’…, and let K be the operator for knowledge, and h any sceptical hypothesis. The general form of the argument can be represented in the following modus tollens:
1 Kp → K~h
3 ~Kp (1, 2 MT)
At first glance, this argument seems compelling: Since I cannot know that I am not a brain in a vat, it seems that I cannot know the reality of anything belonging to the external world.
Nonetheless, the argument from ignorance seems less convincing when we see that we can apply a modus ponens, constructing a converse anti-sceptical argument that could be called an argument from knowledge concerning the external world. This converse argument has the following logical form:
2 Kp → K~h
3 K~h (1, 2 MP)
And here is an instantiation:
1. I know that I have two hands.
2. If I know that I have two hands, then I know that I am not a brain in a vat.
3. Therefore: I know that I am not a brain in a vat. (1, 2 MP)
Both arguments seem equally compelling. Which is the right one? The sceptical or the anti-sceptical argument?
In this paper, I give a negative answer to both these alternatives, for I think that both arguments are equivocal. I intend to justify this claim with the help of a long, complex argument involving three steps: in the first step, I will show that in all their stages both arguments involve attributions of external reality that can be made explicit. In the second, I will analyze the concept of external or objective reality, showing that it has at least two different senses, one belonging to everyday life, another belonging to the very unusual contexts of sceptical hypotheses and artificial realities. Finally, in the third step I will show that the attributions (or disattributions) of external reality implicitly shift from one of these senses to the other in the course of both arguments, which can be easily seen after these attributions/disattributions are fully explicit. These tacit shifts in sense make both arguments equivocal and therefore fallacious.
Making explicit the attributions of external reality
in the arguments
It is something that Kant has already noted: that the category of reality (Dasein/Nichtdasein) is employed in all categorical judgments. We also need to see that in each stage of the arguments from ignorance and knowledge, attributions (or disattributions) of external, objective or concrete reality are being made, although typically in an implicit way. Indeed, all the stages of the arguments must include implicit attributions/disattributions of external reality. To begin with, consider the argument from ignorance. What the argument claims is that since we lack knowledge of ~h, we are unable to know the reality (or existence) of the external world. Because of this lack of knowledge of the reality of the external world, we are unable to know the reality of any state of affairs belonging to it. Hence, we are unable to attribute external reality to anything stated by p.
In both arguments, what is meant by Kp is that I know the external reality or existence of the fact represented by p. Therefore, the conclusion ~Kp of the argument from ignorance amounts to the same thing as the conclusion that I do not know that I can attribute external reality or existence to the content of p. Although always given, such attributions of external reality involved in any trivial statement p remain more or less implicit in everyday language. For example, when p is the statement ‘This piece of chalk is real (or exists)’, the attribution of (external) reality is indeed explicit, since the issue is existence. However, when p is ‘I am holding a piece of chalk’, the attribution of reality remains implicit, although the statement can be restated more explicitly as ‘I am really holding an externally real piece of chalk’. Attributions of external reality usually remain implicit for a simple reason: Since they are always present in discourses, it is superfluous to spell them out. (This case is comparable to that of statements: these are so commonplace that one does not need to make explicit the illocutionary act of stating by saying ‘I state p’.)
One way to uncover the attributions of external reality implicitly made in knowledge claims concerning the external world is to deny that we know the external reality of what is asserted in the utterances. It is a strange thing to say, ‘In asserting that this is a piece of chalk, I do not know whether it is real or not’. It is no less absurd to say, ‘I know that I am holding a piece of chalk, but I do not know whether or not it is (externally) real, and I also do not know whether I am really holding anything’. It is therefore clear enough that since Kp is a statement about the external world, an attribution of external reality to its factual correlate must always be at least tacitly involved.
Once we have grasped this, argument (I) instantiating (A) can be restated in a way that makes explicit the assumptions concerning external reality:
1. If I know that I have two (externally) real hands, then I know that I am not a(n) (externally) real brain in a real vat.
2. I do not know whether I am not a(n) (externally) real brain in a real vat.
3. Therefore: I do not know whether I have two (externally) real hands. (1, 2 MT)
However, argument (II) instantiating (B) can also take a form that makes explicit the attributions of external reality:
1. I know that I have two (externally) real hands.
2. If I know that I have two (externally) real hands, then I know that I am not a(n) (externally) real brain in a real vat.
3. Therefore: I know that I am not a(n) (externally) real brain in a real vat. (1, 2 MP)
These arguments only make explicit what is already assumed in (I) and (II), namely their concern with external reality. As promised, below, I show that the attributions of external reality in both arguments have a different meaning in the premises than in the conclusions, which makes them equivocal. However, in order to achieve this end we need to make a sufficiently detailed preliminary examination of the meanings of expressions such as ‘external reality’ or ‘existence’.
Carnap’s semantic distinction and its limitations
In the search for an analysis of the concept of external or objective reality or existence in its possible association with scepticism concerning the external world, we may ask whether Rudolf Carnap’s famous distinction between external and internal questions of existence or reality could be of some help. His distinction applies to all domains of knowledge, but its application to what he calls the ‘world of things’ (the external world) is what interests us here. The usual questions about the reality or existence of the external world are what he calls internal questions. In this case, he states:
…to recognize something as a real thing or event means to succeed in incorporating it into the system of things at a particular space-time position so that it fits together with the other things recognized as real, according with the rules of the framework.
Therefore, when we ask ourselves whether the Statue of Liberty really exists, or whether there really is a Santa Claus, we are stating internal questions, the first of which is very successful in locating something among other things belonging to the external world, while the second is not.
As Carnap also notes, philosophers may ask about the existence of the world of things (the thing-world), about the reality of the external world as a whole. However, for him it would be misleading to ask whether our external-world-as-a-whole really exists, understanding this as an internal question of existence. This would be to pose a metaphysical question that is unverifiable and consequently senseless – for an internal question can be answered, and therefore stated, only when it is about things related to one another within the system – never when it is about the system as a whole. For Carnap, a question about the existence of the world of things would only make sense when understood as an external question which concerns merely our decision to use a linguistic framework for the world of things (the thing-language for the thing-world). This acceptance, however, is not the result of a cognitive decision, but rather of a pragmatic one, based on factors like the expedience, fruitfulness and efficacy of the framework.
Although Carnap’s distinction does not seem to lack a difference, it has its own flaws, already pointed out by philosophers such as Barry Stroud and P. F. Strawson. These theorists have convincingly argued that the problem of the reality of the external world as a whole cannot be reduced to the mere status of a non-cognitive linguistic decision: We are in some way inescapably led to accept the existence of the external world.
In order to understand that there is more to the question than the assumptions in Carnap’s distinction, consider the following statement, showing a pervasive ambiguity in our attributions of reality to the external world:
(1) I know that the (external) world is real
This statement is ambiguous, for it could mean
(1a) I know that our external world has reality.
but sometimes it seems that it can also mean that I know that this is the external world, or
(1b) I know that our world (contrary to sceptical hypotheses) is the ultimately real one.
The difference between (1a) and (1b) becomes clearer when we consider what truth-value we would assign to each statement. Consider the negation of (1a):
(~1a) I do not know whether our external world has reality.
Surely, most of us would agree that (1a) is true, whereas (~1a) is false. I know, and we all know, that our external world has reality in the sense that it contains reality, that it is full of reality or existence. This truth can only be denied metaphorically, as when a poet imagines fog-shrouded London as an ‘unreal city’.
But statement (1b) is a different case, for it seems to be false. To see this more clearly, consider its negation:
(~1b) I do not know whether our world (contrary to any sceptical hypothesis) is the ultimately real one.
Statement (~1b) seems to be true, because we feel that we do not have the epistemic resources to reject the logical possibility that, unbeknownst to us, some sceptical hypothesis about the external world is true and that its reality is only virtual.
In contrast to (1b), (1a) cannot be shown to be false, even if it is true that I am a brain in a vat or a soul deceived by a malign genie. Even if a sceptical hypothesis were true, I would still be justified in thinking that the world I am experiencing is a perfectly real one, and not something like, for example, the limited world I experience when watching a movie or when asleep in the faint world of my dreams.
Our question now is: What kind of relationship could be found between Carnap’s distinction and the present distinction between two cases of attributions of reality? Initially, we have an uneasy feeling that sense (1a) of the statement ‘The (external) world is real’ concerns an internal question of reality, while sense (1b) concerns an external question of reality. However, since both statements concern the whole world of things, it is clear that, following Carnap, the attribution of reality in both of them should answer an external question of reality that could be established as the result of pragmatic decisions... If this is the case, however, why do we distinguish (1a) from (1b), regarding the first as true and the second as false? Why are we so prone to attribute cognitive status to (1a), but not to (1b)? In sequence of this paper we will delineate an analysis of the concept of external reality or existence that makes it possible to answer these questions.
Introducing a new semantic distinction
My own strategy for analyzing different kinds of attributions of external reality draws on Wittgenstein’s later semantic reflections and methodological approaches. I will assume two very plausible semantic hypotheses derived from his writings. The first is that the meaning of an expression can be approximated from how it is used, and that a difference in the way of using or employing an expression corresponds to a difference in its sense or meaning. A closer examination of our linguistic praxis shows that there are many nuances of meaning that we seldom notice. This lack of awareness, even if it is not the only source of philosophical problems (as Wittgenstein sometimes seems to think), is at least a relevant source of philosophical misunderstandings, particularly in the case of what seem to be non-substantive riddles such as the sceptical one.
The second insight is his view that the criterial rules for the application of an expression are constitutive of its meaning. According to Wittgenstein, without criteria for its application one expression is devoid of meaning, and when we change these criteria, we change its meaning (its form of application, its usage or ways of use).
The first semantic insight is related to the second by the fact that when we speak about ways of using words, we are speaking about rules (or combinations of rules) determining the spatio-temporally located, episodic uses of expressions. Criterial rules (or combinations of rules) can be seen as rules that cognitively contribute to these episodic uses of expressions. Consequently, as a way to explain the meaning of conceptual expressions, we can perform a criterial analysis, making explicit the criteria for applying the expressions.
Let us apply the first insight to what seems intuitively right in Carnap’s distinction between internal and external questions. It is clear that in the typical case when we ask or answer an internal question about existence, we are using the concept of external existence or reality in a way that differs from the way we use the concept of external existence or reality to ask or answer an external question. Consequently, we are applying the concept of external existence or reality in different senses. Thus, when we ask an internal question, such as whether the Statue of Liberty really exists, we are using the expression ‘really exists’ in a different sense than when we ask whether the external world as a whole really exists in itself and not, for example, as a virtual reality created by a supercomputer…
I will call the sense or kind of attribution of external reality usually linked with an internal question an inherent one, and I will call the sense or kind of attribution of external reality that can be linked with an external question – concerning the reality of the world beyond the sceptical hypothesis – an adherent one.
There are some linguistic clues that confirm this distinction. First, a linguistic characteristic of the inherent sense is that we can replace the words ‘real’ or ‘exists’ with the word ‘actual’. Moreover, instead of saying that something is inherently real or exists inherently, we can in this sense also say that it possesses reality, that it has reality, that it is full of reality: the piece of chalk that I am holding is actual, it has or possesses reality. On the other hand, something that is not real in the inherent sense, such as an imaginary piece of chalk, is called non-actual, or non-existent, something that does not have reality or is empty of reality. However, the same contrast does not apply to the concept of adherent reality. When we attribute adherent reality to something (my hand, the world in which it is located…), we can say that it is actual, that it possesses reality. However, when we negate the attribution of adherent reality, we cannot negate actuality or the possession of reality. We cannot affirm that a world that is adherently non-real (like the world of the brain in a vat) and everything in it (like my illusory hand and the illusory chalk) lacks these properties. Such a world, although adherently unreal, is often conceived as perfectly actual. Moreover, we usually can say that this world possesses (inherent) reality, that it is fully real in its experience. Indeed, it seems that lack of actuality and lack of possession of reality are typically properties of lack of inherent reality, not of lack of adherent reality.
Although we have found some linguistic clues confirming our proposed distinction between inherent and adherent senses or attributions of reality, the distinction remains wanting. However, this distinction can be sharpened and more profoundly analyzed when it is considered in terms of criteria. In what follows, I will make a criterial analysis of conceptual expressions attributing external reality – expressions like ‘external reality (or existence)’ or ‘objective or concrete reality (or existence)’ – in order to distinguish more adequately the inherent from the adherent senses. I will begin with the search for criteria for inherent attributions of reality.
Inherent senses of our attributions of reality
Let us consider, in searching for criteria, the conceptual expressions used for the attribution of external or objective reality or existence in the supposedly inherent sense. The genetically primary use of these expressions seems to be to ask whether or not things in the external world around us really exist, since they are the first objects of our acquaintance. According to our understanding of Wittgenstein, we are entitled to suppose that the inherent sense of the conceptual expressions used for the attribution of external reality to the things around us is constituted by criterial rules for this attribution. Such rules would show us that only by the satisfaction of certain criteria of external reality would we be able to apply expressions in an inherent sense, such as ‘is externally real’, ‘exists objectively’, ‘is actual’ or ‘has concrete existence’.
Can we find such criteria? My claim is that criteria do indeed exist for the inherent sense of attributions of reality. More than this, many influential thinkers have addressed them. Thus, according to the representationalist Locke, our opinions about material objects are justified by properties associated with our ideas of the senses, e.g., their involuntary character, their orderly and coherent agreement, reflecting law-governance, and other people’s awareness of them. According to the immaterialist Berkeley, ideas formed by the imagination are faint, indistinct and also entirely dependent on the will, while ideas perceived by the senses are vivid and clear and have no dependence on our will. As well for Hume, impressions are seen as perceptions of real things when they enter the soul with the ‘most force and violence’, unlike the ‘faint images’ that we have in thinking and reasoning. For Kant, the conformity to law (Gesetzmässigkeit) of all the objects of experience is what defines the formal aspect of nature. For J. S. Mill, the external (material) world consists of continuous or warranted possibilities of sensation, following from one another in conformity with laws. According to Gottlob Frege, the main criterion of objectivity is interpersonal accessibility, followed by independence of the will, while the main criterion of reality is spatial and/or temporal location. Hence, for him the realm of objective reality is made up of those things that are interpersonally accessible and spatially and/or temporally given. In a well-known paper, G. E. Moore summarizes the properties of external reality, stating that the real is something independent of the mind that is verifiable by others, continuously connected with other things, and in this way has certain causes, effects and accompaniments (I would say ‘displaying regularities’) with the highest degree of reality. Finally, Sigmund Freud suggested that a new-born child is driven by the pleasure principle, constantly seeking immediate gratification and unable to distinguish between the external and internal worlds. Only gradually does the child learn that the external world, unlike its own inner fantasy world, does not obey its will, and this forces it to learn to postpone the gratification of its drives. In this way, it comes to replace the pleasure principle with a different one, namely, the reality principle.
Indeed, starting in childhood, we gradually learn to distinguish external reality from appearances by means of criteria such as the greatest vividness of sensation, independence of the will and interpersonal accessibility, and it seems to be a conceptual truth that world-states experienced without any of these properties should be said to be unreal, non-real or non-existent in the usual sense. Although it has already been argued that criteria like these are of no use, since none of them is sufficient, we still have the alternative of combining them on the grounds that taken as a whole they are strong enough to be conclusive. Doing this in a non-systematic way, we can say that the things around us – using the word ‘thing’ in the widest sense, in order to include objects, properties, conditions, states of affairs, events, processes, etc. – are real when:
1. the sensory experience of them has the greatest intensity,
2. they remain independent of our will,
3. they are interpersonally checkable by anyone in the right position, and their experience is usually co-sensorial (although this latter requirement can vary from case to case),
4. they display regularities (external things obey regularities such as those imposed by natural laws, social norms, etc.).
These are what we may call the primary or standard criteria of external reality. My contention is that we continually apply these criteria jointly in order to be certain of the reality of the things around us. These four criteria together are what must usually be satisfied for a justified application of predicates like ‘…is externally real’ and ‘…exists objectively’ in their primary inherent sense, namely, attributing reality to things belonging to the external world surrounding us. Moreover, taken together these criteria form what we could call a definitional criterion in the sense that once they are given, they warrant the attribution of inherent reality. In other words, the satisfaction of all these criteria is a sufficient condition for the attribution of external reality.
In order to illustrate this, let us suppose that I hold up a piece of chalk and say, ‘The piece of chalk I am holding is real’, or simply, ‘I am holding a piece of chalk’. Insofar as criteria (1) to (4) are satisfied, I am entitled to think that the piece of chalk isn’t just a figment of my imagination, but rather something externally real, objectively existent in the inherent sense. Indeed, in order to be true, the claim ‘The piece of chalk I am holding is (externally) real’ must satisfy criterion (1), because the vividness of the sensations is maximal, unlike those in a dream. It must satisfy criterion (2), because the chalk is independent of your will or mine (I cannot make it disappear like a mental image). It must satisfy criterion (3), because its experience can be the object of interpersonal verification, that is, current or previous experience convinces me that it can be recognized as the same by any other knowing subject (usually we cannot share a hallucination – synchronized collective hallucinations aren’t easily conceivable). Moreover, usually it is co-sensorially experienceable: Not only can I see the piece of chalk, but I can also touch and smell it. However, this sub-condition cannot be generalized to all cases: a rainbow cannot be co-sensorially experienceable in any way and I never saw a hammer bird, though I have already been disturbed by several. Finally, a real piece of chalk must also satisfy criterion (4), because it must have the regularities of objects that comply with natural laws: the chalk can be used to write on a slate, and one can break a piece of chalk. If dropped, a piece of chalk falls to the floor, while imaginary chalk might conceivably float in the air, etc. Finally, our experience of things must have some endurance in time in order to show regularity (when we awake, for example, we need a minimum of time in order to recognize our surroundings).
We attribute reality to things in the world, because we assume that they satisfy these criteria, either directly to our senses (as in the case of opaque medium-sized dry goods), or indirectly by means of their association with things that can be directly experienced (as in the case of the subatomic entities posited by physics, like quarks, neutrinos or electromagnetic waves). If all these criteria are satisfied, the piece of chalk must of necessity be seen as inherently real. In this way, we incorporate things among other things in a spatio-temporal system of things according to the rules of the thing-language (in this case the criterial rules), as Carnap requires for his internal questions of existence.
Indirect satisfaction of the criteria of reality
We ascribe reality to many things in our environment, although they in no way meet the criteria of inherent reality in the default form. This is the case with the very small entities discovered by scientific research, such as bacteria, viruses and DNA in biology, molecules in chemistry, and entities like the subatomic particles and force fields studied by modern physics. How is this possible? We can clarify the problem with an example. Suppose we observe a trail in a cloud chamber and conclude that this trail was caused by a positron. Certainly, the trail meets the standard criteria of reality: it has maximum perceptual intensity, it is independent of the will, it is interpersonally checkable, etc. We say that the trail is a real trail, and we interpret it as caused by the passage of a positron through the chamber. (Based on previously verified theories of particle physics and cloud formation, etc., we believe that charged particles moving through the supersaturated gas in the chamber create ions, and water vapor condenses around these ions, creating the visible trails.) The positron isn’t directly verifiable, however, because it is not itself visible and does not seem to satisfy anything like the standard criteria of reality. Even so, we are willing to say that in order to be the cause of the trail, the positron must exist objectively, it must be real. How is this possible?
The answer brings us to the process of semantic extension, already mentioned by Aristotle in his account of focal non-homonymity. I will use as an example his famous account of the semantic extensions of the word ‘health’. Supposedly, the predicate ‘healthy’ was primarily and originally assigned only to human beings and animals, and later extended to food and exercise. These also came to be called ‘healthy’ in practice, inasmuch as they are causal determinants of the health of living organisms. Applied to such cases, the principle of semantic extension states: If A causes the property F of B, then we may be entitled to extend the assignment of F to A. In other words, if food and exercise have the property F of promoting health in human beings, then we are entitled to extend the assignment of F to food and exercise, even if the sense is not really the same, since it means ‘promoting health in humans and animals’.
Something similar can be said about the concept of external reality in order to explain the extension of this concept to things that cannot be experienced. In the external world, the property of the real existence of a given effect is in some way causally related to the source of this effect. Now, my claim is that the above-clarified principle of semantic extension also applies to the case in which F is an attribution of external reality. This suggests that if effects are considered real, their causes may also be said to be real, even if they cannot be directly experienced and thus cannot be subjected to our standard criteria of reality. In other words, if certain things meet the standard criteria for external reality, so that we can attribute inherent reality to them, then we can also assign the inherent sense of reality to their causes, even though we cannot say that they satisfy the criteria. As a terminological matter, I propose we say that in this case the standard criteria of reality are indirectly satisfied, meaning by this that we are allowed to say that something has inherent external reality when its effects satisfy the standard criteria of reality. Therefore, we can say that the cloud chamber was traversed by a real positron: it was real in the extended sense of being a cause of effects that meet the standard criteria for the inherent sense of the word ‘reality’.
It is important to note that the converse semantic principle is also true: If A is caused by the property F of B, then we may be entitled to extend the attribution of F to A. Thus, we could say, for example, that physical strength is healthy, although we are applying the word to an effect of health in human beings. Applied to the concept of reality, this principle supports the idea that if causes are real, their effects can also be said to be real. Or: if causes satisfy the standard criteria for the inherent sense of reality, then we can say that their effects indirectly also satisfy them. Thus, for example, if the movement of an iron bar magnet, which is real, produces the movement of electrons in a copper wire, the inferred electromagnetic energy generated by this motion can be considered real.
Again, considering the same example, we can note that the generated energy, in turn, also has an effect that we can measure with a galvanometer. The movements of the galvanometer’s pointer are real according to the standard criteria of reality. However, in applying these criteria we see that we have two ways to extend the concept of reality to electromagnetic energy: we attribute reality to it based on a semantic double transfer from the effect to the cause and from the cause to the effect. Therefore, we conclude that we can regard things as indirectly satisfying the standard criteria of inherent reality at least if they are contained in a network of causes and effects that do satisfy those criteria.
I think that these considerations explain our inclination and even our semantic entitlement to attribute reality to things that we cannot experience with our unaided senses, as well as to say that we can experience them indirectly. Moreover, I believe that this also inclines us to furnish the appropriate rationale for scientific realism, a view that I share (anti-realists could try to reject such an extension of criteria fulfillment as misappropriating the concept of reality).
We have considered methods for extending the concept of reality to the realm of things that we could say we experience only indirectly. Now we need to give a justification for this extension. We want to know why we are entitled to apply the concept of reality to things that are causally or constitutively related to things that are directly experienced. In my view, the reason is that in the inherent sense of reality the boundaries between things that satisfy the criteria directly and those that satisfy them only indirectly are not definitive and unchangeable. They are contingent and possibly changeable, since to a considerable extent they depend on the contingent nature of the subject’s sensory organs. To highlight this point, imagine that our senses were different than they are. Suppose that we were aliens with visual organs capable of distinguishing microscopically small details, so that we could actually see bacteria with the unaided eye. Or suppose that our brain were equipped with sensors so that we could ‘see’ electromagnetic fields, or that we were so constituted that we had a kind of internal Geiger counter with which to sense the presence of dangerously high levels of radiation... In such cases, the domain of application of our standard criteria of external reality would be wider. This domain would extend to many things we currently regard as real because they appear to satisfy secondary criteria of reality. This merely circumstantial character of what should be regarded as the standard way to satisfy the criteria for the inherent sense of reality provides a rationale for the view that we are entitled to extend our attributions of reality to things that we cannot experience directly with our unaided senses. It shows that the difference between the primary and secondary satisfaction of reality criteria is largely contingent and arbitrary. Consequently, there is no good reason why we should not extend the domain of the real beyond what we can directly experience with our unaided senses.
Proof of the external world
We have seen that our attribution of reality can be extended to postulated entities that cannot be directly experienced. However, if you have no prejudice against induction, you will agree that there are much more commonplace spatio-temporal extensions of these attributions. As discussed above, we apply our concept of external reality originally and primarily to (A) those objects, properties, conditions, states of affairs, events, processes … that surround us and presently satisfy (directly or indirectly ) our standard criteria of external reality (e.g., respectively this computer screen and the electrical energy that illuminates it). But this is not all. We also apply the concept of external reality to (B): many other things that are not immediately experienceable, but which we have good inductive reasons to suppose would, under suitable conditions, satisfy (directly or not) our standard criteria, and consequently can also be regarded as real. This is the case with (B1): everything we have ever experienced that at the present moment is too remote or inaccessible for us to (directly or indirectly) experience. It is also the case with (B2): many things that certainly must exist (we are assured), even though they have never been – and probably never will be – experienced by anyone (what could be called the openness of the world). We believe this because in all our past we have been constantly experiencing new things that we would never have expected to ever experience. And certainly this is also the case with (B3): many things that we know to satisfy the criteria only via testimony. Now, my claim is that beginning with the successive experience of (A), namely, things around us that directly or indirectly satisfy the standard criteria of inherent reality, and based on (B1), (B2) and (B3), we can inductively infer that there is presently a whole world of things around and beyond us that satisfy these criteria of reality. They satisfy them in the sense of being confirmed as able to satisfy them if they could be experienced under the appropriate circumstances, although they are not presently being experienced. This is why there is nothing wrong with using concept-words like ‘real’ or ‘exists’ to claim that our external world as a whole is objectively real or exists. It is real insofar as by ‘our external world’ we mean something like ‘the sum of all the things that we think to satisfy our standard criteria of external reality’. This is an extension of the inherent sense of our attributions of external reality, which Carnap neglected to consider when he proposed his distinction between internal and external questions of existence.
Accepting this kind of reasoning, I will call all attributions of reality going beyond the present experience of our surroundings, including indirect experiences of things, extended inherent senses (uses, ways of application) of our conceptual expressions for external reality or existence. This contrasts with what we might call the primary inherent sense, which is given by the direct satisfaction of the standard criteria of reality and is genetically prior. When the inherent sense of the word ‘reality’ is extended to the maximal limits of its applicability, we can apply it to the external world as a whole.
Now, let us use the expression ‘satisfaction of the criteria of external reality’ in a wide sense, which includes not only the direct satisfaction of the standard criteria of external reality, but also the indirect satisfaction of these criteria by things that cannot be experienced with the unaided senses. Having in mind only the inherent sense of external reality or existence, we can formulate an approximate reconstruction of the reasoning that leads us to the commonsensical conclusion that our external world as a whole really exists, that it is real in the inherent sense:
1. Many things that are being experienced right now satisfy (as I have stated, directly or indirectly…) the criteria of external reality (our bodies, external things around us, visible or not…).
2. Most of the many things we have experienced in the past have been experienced more than once and have satisfied the criteria of external reality each time they were experienced.
3. (inductively, from 2) There are many things that have been experienced in the past and, although they are not being experienced now, (are still able to) satisfy the criteria of external reality.
4. We are constantly experiencing many new things around us that satisfy the criteria of external reality.
5. (inductively, from 4) There must be many things never experienced that satisfy the criteria of external reality.
6. Testimony constitutes a reliable way to achieve knowledge.
7. There is much testimony for things that satisfy the criteria of external reality.
8. (deductively, from 6 and 7) There must be many things that we have not personally experienced that satisfy the criteria of external reality, this being known via testimony.
9. (deductively, from 1, 3, 5 and 8) There is a sum total of things, some of which are (A) presently experienced things satisfying our criteria of external reality; some of which are (B1) things not presently being experienced that we know satisfy our criteria of external reality, since they have satisfied these criteria in the past; some of which are (B2) things that are still unknown, but are able to satisfy our criteria of external reality, because we are constantly experiencing new things that satisfy these criteria; and some of which are (B3) things not experienced that satisfy the criteria of external reality via testimony.
10. What we mean by the idea of our external world as a whole is the sum total of things, encompassing those that are (A), (B1), (B2), and (B3).
11. (deductively, from 9 and 10) Our external world as a whole satisfies the criteria of external reality.
12. Whatever satisfies the criteria of external reality is (inherently) real.
13. (deductively, from 11 and 12) Consequently, our external world as a whole is (inherently) real.
In my view, this is the true proof of the external world’s existence, the proof whose absence Kant once called the scandal of philosophy. Even if only intended as a rough approximation, this argument is plausible enough for our purposes. It is able to capture what is meant by the plain man when he makes statements that appear philosophically naïve, such as, ‘It is obvious that the external world (as a whole) exists’, or ‘Only philosophers and madman would doubt the reality of our external world’. What is claimed with these statements is that we have a good inferential basis to believe that the whole world, as a sum of its constituent parts (presently experienced, already experienced, still not experienced by us, and claimed by others to have been experienced by them), satisfies our standard criteria of external reality, and therefore actually does exist. Indeed, even Stone Age peoples may have engaged in such reasoning in a non-reflexive way, since they could already judge, we surmise, that the world is real. However, an extremely comprehensive reasoning like this is like the glasses that we use without seen them; it would remain forever implicit, unless philosophical analysis discloses it in some detail.
The inherent senses of the concept of external reality are familiar and non-problematic. The existence of an extended inherent sense of reality can explain why we believe in the truth of statement (1a), affirming that we know that our external world as a whole has reality, for (1a) is the assertion of the inherent sense’s widest extension. It also explains why we somehow feel that (1a) should answer an internal and not an external question of existence. Moreover, the existence of the extended inherent attribution of reality shows that Carnap was mistaken when he concluded that one cannot meaningfully pose something like an internal question about the reality of our world as a whole, for this question would be unverifiable and metaphysical. He believed this because with the internal question he had in mind only the primary inherent sense of our attributions of external reality, along with some relatively modest extensions of it, without foreseeing the possibility of its inductive generalization to what we mean with words like ‘world’ or ‘universe’. However, it seems clear now that, based on a tacit inductive procedure which begins with the satisfaction of the standard criteria of reality, we arrive at justifiable knowledge that the whole external world referred to by us is real in what I will show to be the important sense of the word.
Adherent attributions of reality
Let us suppose that while under the influence of a hallucinogenic drug for several hours I experience a perfect illusion of Ancient China in the time of Kublai Khan. After the effects wear off, I can say to myself: ‘This was a world of my imagination, not the real one’, for I have good reason to think so. However, in this case I am not disattributing reality in the inherent sense, because all the standard criteria of reality have been sufficiently satisfied. In the case of this hallucinatory world, I suggest, I would be disattributing reality in the already named adherent sense of the word, a sense that can also be considered in sceptical contexts.
What are the criteria for this adherent kind of attribution of reality? We can explore this point using sceptical thought-experiments. For instance, I can imagine that one morning I might wake up in a completely unfamiliar environment, in a different body, surrounded by strange, alien creatures. They explain to me that until this time-point I was not living in the real world. During my entire previous life, I was only a ‘brain in a vat’, monitored by a supercomputer that simulated my external reality. They say that this is a conventional pedagogical procedure intended to foster mental diversity on the planet Omega, where during its early development each new brain lives in a dream world created by a computer program specially designed for it. In my case, this just happened to be the program ‘Philosophy Lecturer on Earth’. Now, they further explain, my brain has been implanted in a real body, and I will henceforth have to live in the truly real world. Since all my further experiences turn out to be completely consistent with their explanations, I gradually come to accept that they are telling the truth. I now believe that the world I previously experienced was not the real world, but only a virtual reality.
It is important to see that I can find criteria that would lead me to this adherent disattribution of reality in the adherent sense. However, they have nothing to do with any straightforward application of the standard criteria for any inherent sense of external reality! For the highest intensity of experience, the independence of the will, the co-sensoriality, the possibility of interpersonal access, etc. were all already available to me when I was living on the Earth as a brain in a vat, just as much as they now are on the planet Omega. I can even say that my world, when I was just a brain in a vat, was actual, it had reality in the primary and extended inherent senses, neither more nor less so than the world presently accessible to me in the here and now.
We have already thought about criterial configurations suggesting that the past was adherently unreal. Still, it is just as easy to imagine the criteria working in the present or in the future. We can easily imagine criteria with which to decide whether or not the world in which one will someday live or is currently living is the (adherently) real one. Let us suppose that under the criminal code of planet Omega, a convicted murderer is not executed, but instead condemned to spend the remainder of his life as a brain in a vat. After sentencing, a convicted murderer is anesthetized and his brain is removed and immersed in a vat containing fluids that keep it alive and functioning. There he can continue to live a perfectly ‘normal’ criminal life, even if disturbingly aware that he is experiencing only a virtual world where the only reality is one simulated by a supercomputer. He will live in a world that is perfectly real (actual) in the inherent sense, but which he knows is and will remain virtual, non-real in the adherent sense. (This should not seem so strange: during a nightmare we can often semiconsciously, based on our self-awareness and memories, reassure ourselves that we are only dreaming…). As well here, we may have criteria for the adherent non-reality of one world relative to another, insofar as we can be aware of the constructed, artificial nature of one of the realities. It is in this relative sense that we can know that one of these worlds is not real.
It may seem that attributions/disattributions of adherent reality concern only the world as a whole. However, the concept of adherent reality can be applied equally well to specific parts of the world. Suppose, considering a common experience with virtual reality, that a person uses a special high-tech glove to close her hand around a holographic projection of a teacup. Inherent criteria like intensity, co-sensoriality, or even interpersonal access might be satisfied. In this way, the holographic projection acquires some inherent reality... However, the fact that the person knows that this is an experiment being conducted under the particular conditions surrounding the cup serves as a criterion to convince her that the teacup she is holding is adherently unreal relative to the external world as we know it. In this way, the adherent reality of the fact referred by a statement p (e.g. ‘I have two hands’) can also be separately contested.
Criterial configurations for adherent reality
More precisely, how could we achieve the knowledge that some things are adherently unreal? How can I know, for example, that the real world is the one on the planet Omega, and not the earlier one I knew on the Earth? My reasoning could be formulated as follows:
1. All of my most recent experiences are of the new inherent reality of the planet Omega (my new body, the bizarre creatures among whom I now live, the incredibly advanced technology…)
2. I still remember my experiences of the inherent reality of the very different Earth world.
3. I know that the standard inherent criteria of external reality can be satisfied in a fictitious environment, although in most cases only partially (e.g.: co-sensoriality and interpersonal experience without much intensity in dreams, or intensity and regularity without co-sensoriality in experiments that create an imaginary reality).
4. I have by now heard and understood very reasonable explanations for the change, compatible with the hypothesis of a fictitious world (my earlier experiences were generated by a supercomputer to serve pedagogical aims, etc.)
5. These explanations are supported by evidence (I have seen the supercomputer and other brains in vats…)
6. (From 1 to 5) I conclude that the Earth’s world, contrasted with Omega’s world, is not the real one in the adherent sense, since its inherent reality was a manufactured by-product of the relatively adherent reality of Omega’s world.
What this argument shows is that I reached my conclusion that my previous Earth world was not the real one based on a criterial configuration that has nothing to do with the standard criteria of inherent reality. The adherent criteria of external reality (supposing it is still reasonable to speak of criteria here) are of a coherential nature. In the case above, they form a broad criterial configuration that consists in the coherence of the new information with my awareness of my new and old experiences and knowledge.
Moreover, when we consider the other cases in which an adherent reality is found, we discern a similar coherence model with a variable criterial configuration. Consider my virtual experience in Kublai Khan’s China: since I remember taking a hallucinogenic drug, and since my world before and after the experience remains the same, I can infer that the world I temporarily experienced was a mere drug-induced fantasy. Consider also the experience of virtual reality with the holographic projection of a teacup. Because of the available information and the context of the experience, the subject of the experience knows that the teacup, although inherently having some properties of real things, is adherently unreal. The important thing is above all that in each of these cases (that of the brain in a vat, the intermediate facsimile world of Kublai Khan’s China, and the particular virtual reality of the holographic teacup ), what is experienced is only adherently unreal relative to what we comparatively accept as an adherently real world.
We now see that the criteria used to identify an adherent reality must be utterly different from the standard criteria of reality used to identify inherent reality. However, there is one link between adherent and inherent criteria of reality. This is that the criteria of reality in the adherent sense are used to decide between two conflicting (inherent) realities that already satisfy the inherent criteria of reality, distinguishing one of them as an illusory by-product of the other.
A possible objection
Opposing to the sceptical hypothesis, one could object that such criterial knowledge that the external world or even parts of it are adherently real or unreal is rather feeble. It could be, for example, that the new world of Omega in our first example was only another figment of someone’s imagination, just as the first one… with a different program, called ‘Being awakened from life as a brain in a vat’, was run in place of the old ‘Philosophy Lecturer on Earth’ program. It is also conceivable that my past life up until this event was lived in the real world, and my brain was removed from my head by aliens and then placed in a vat of life-sustaining fluids, where the new program ‘Being awakened from a life as a brain in a vat’ was run... In this case, I would be doubly fooled in my beliefs about both my present and my past life. It is even conceivable that in the case of the experiment with virtual reality the subject was actually a brain in a vat and the supposed holographic projection of a teacup was in fact the only real thing she had access to, which she held using a robotic arm operated from outside her fictional world! In conclusion: it seems that the proposed relative and variable criterial configurations for adherent reality are like shifting sand; diverging from the cases of the standard criteria of inherent reality, which are indefeasible when in fact given, the criteria for adherent reality are always defeasible, even when in fact given. This would mean that in the end of the day we cannot really know whether any particular world is the real one...
Our answer is that this objection demands an unnecessarily absolute sense of adherent reality, whereas in fact we have only a relative and casual sense of this reality. In order to qualify this answer, we need to distinguish between two senses of the adherent attributions of reality that at first glance seem to be possible:
(a) a relative sense (considered in our examples of counterfeit realities).
(b) a non-relative or ultimate or absolute sense.
Below, I will show that sense (a) is workable, that there is indeed a relative sense of adherent reality, while (b) is in fact an illusory sense of adherent reality and should be dropped.
We have already analyzed several examples in which the relative sense (a) of adherent reality was used. The notion of a relative sense of a word is well known. The sense of a word is relative when it is arrived at only by contrast within a context. The word ‘small’, for example, has a relative sense: a baby elephant is small relative to an adult elephant, but large relative to a mouse. In its adherent sense the word ‘real’ functions in a somewhat similar way. Our attributions/disattributions of reality make sense only against a contextual background, such as that offered by the absence of sceptical evidence: we attribute adherent reality in its absence, otherwise we disattribute it. In this case, the attributions/disattributions of reality are relative in the sense that they are based on contextual evidence that is always defeasible, for it can always be challenged by new evidence and information, as we have seen. I maintain that, within its narrow limits, this relative sense of our adherent attributions/disattributions of reality is legitimate, since we can conceive of coherence criterial reasons for them, as we did in the last section. But these criterial reasons always have a comparative character subordinated to the identification of some kind of fictional scenario.
Consider now (b), the postulated non-relative or ultimate or absolute sense of our adherent kinds of reality attribution. This should answer the question of whether our world is ultimately the real one, that is, real in an indefeasible sense, real beyond any possibility of doubt arising from sceptical hypotheses. It should be clear that such criteria for ultimate reality cannot be as such available to us, and we cannot conceive of any.
Thus, we reach the conclusion that although there is a double answer to the objection that we cannot really know if a world or any part of the world is adherently real: we can know that the world or some part of it is adherently non-real (or real) in the relative sense of (a), measured against a contextual scenario of non-counterfeit (or counterfeit) realities, but not in the postulated absolute sense of (b), independently of any context, for we lack criteria for this absolute sense. We conclude that the allegation that we cannot know whether things are adherently real arises from the assimilation of sense (a) into a postulated sense (b).
Reflection on this issue shows why we believe in the falsehood of statement (1b) from our earlier discussion of Carnap’s view. This statement asserts that I know (without any sceptical contrast) that our world is the ultimately real one. We cannot possibly know that much, because we lack criteria for knowing that the external world is ultimately real or that any sceptical hypotheses are definitely false. For, supposing that we have evidence for a sceptical hypothesis, it is always possible that this evidence could be refuted by new informational evidence. Still, nothing excludes the possibility that the new conclusion would likewise be refuted by further evidence, and so on. From this we may conclude that we are also unable to know that parts of our external world are ultimately or absolutely real in the adherent sense – a conclusion that extends to the fact referred to by any statement p. Indeed, we cannot even know whether our hands are ultimately real.
Because of his views, Carnap would be compelled to say that (1b) is true. He would have to say that in fact we attribute ultimate reality in the adherent sense to our external world as the result of a posit, a pragmatic decision based on grounds like the expedience, fruitfulness and efficiency of a linguistic framework. However, as Barry Stroud has convincingly argued, not only does there seem to be no valid alternative to the thing-language, it also does not seem to make any sense to claim that the external world’s existence depends on our decision to adopt a particular linguistic framework. In my view, Carnap reached this conclusion because he confused the extended inherent attribution of reality to the thing-world, which is non-relative, with the relative adherent attribution of reality, taking the second of these to be the first. The result is the attempted fusion of both concepts in the non-relative positing of the reality of a world as something pragmatically justified.
The conclusion is that we are actually unable to know the ultimate adherent reality or unreality of an external world. However, we should not feel threatened by this conclusion, since it turns out to be harmless when we reflect that in this case we have no criterion for knowledge, and expressions without criteria are devoid of meaning. If we accept this, then the statement:
The external world is ultimately real (in the adherent sense).
is as devoid of sense as the statements: ‘The whole world (with everything in it) doubled in size last night (in an undetectable manner)’ or ‘My brother died the day after tomorrow’, which being unverifiable are like the loose wheel in the machine, to use Wittgenstein’s metaphor. Moreover, the following statement is true:
I do not know whether the external world is (ultimately) real.
Indeed, I do not know this. However, this statement has the same nature as the following ones:
I do not know whether the whole world doubled in size last night in an undetectable manner.
I do not know whether my brother died the day after tomorrow.
I do not know whether colourless green ideas sleep furiously.
I do not know whether Saturday is in bed.
All these five statements are all equally true because they are about senseless things and senseless things cannot be known; but even if true, these statements are empty enough to be totally trivial and harmless. The point to be noted is that our world matters to us because of its quality of being inherently real (intense to the greatest degree, independent of the will, public, uniformly structured, enduring, etc.) and not because it is an adherently real world in the ultimate sense. We will never be in the position to know this, and this kind of knowledge is at bottom senseless.
How can the question of whether our external world is adherently real in a non-relative, ultimate or absolute sense not seem devoid of meaning? My take is that all we have here is an illusory impression of meaning that arises from confusing the adherent and the extended inherent senses of our attributions of reality to the whole world. In the extended inherent sense, it is perfectly correct to say that we know that our external world is real, and this can be interpreted as an attribution of adherent reality. However, we are tempted to fuse the relative sense of the adherent attribution of reality to our world – only possible against the background of accepting or refuting fictional contexts like those considered in sceptical hypotheses – with the extended inherent attribution of reality to our world. We may suppose that in this way we can arrive at an ultimate sense of these attributions that is free of criteria, and this is the issue at stake when we ask whether our external world (or some part of it) is the real one in the adherent sense. That is: it is our lack of awareness of these fine semantic distinctions what leads us to see the absolute or ultimate ‘sense’ of attributions of reality as though it were something palpable, whereas it is really only a semantic fata morgana.
Recapitulating what we have learned so far
The result of our analysis is that we have found certain uses or senses of the expression ‘external reality’ that must be distinguished:
(A) The inherent senses or attributions of external reality, based on the standard criteria of reality (1) to (5). These inherent senses form a scale that begins with the primary inherent sense and continues with various degrees of extended adherent sense, culminating in its application to the whole world. These senses can be genetically specified as follows:
(A1) The original inherent sense of the concept-words for external reality. This sense is constituted by the application of criterial rules for the satisfaction of the primary or standard criteria when we acknowledge the reality of things around us (e.g. ‘My hands, this tree, that animal, are real’). We human beings always begin with this default acknowledgment in order to extend our knowledge of the external world.
(A2) The extended inherent sense of concept-words for external reality. This sense inductively extends the application of standard criteria of reality to what is not being experienced presently or primarily. In this sense, we can say that things indirectly experienced like atoms are real, we can say that many external things we have experienced in the past are real, etc. If we consider all the things that can be assumed to be real in the original and extended senses of our concept-words for external reality, we arrive at the concept of the reality of the world as a whole, as shown in our proof of the external world. The plain man is appealing to this sense when he claims: ‘Of course the external world is real; were it not real, it would not be our external world’. (When this sense is intended, it also affirmatively answers something like an internal question of existence applied to the whole thing-world, pace Carnap.)
(B) The adherent senses of concept-words for external reality. These senses can arise when we need to distinguish the pseudo-reality of dreams, hallucinations, artificial reality and the worlds considered in sceptical scenarios. In the first cases, we can say that parts of the world w1 are unreal, because they are products of kinds of virtual reality produced by w1 itself, while in the last case we can say that a world w1 is unreal because it is a by-product of a world w2. The important point about the adherent senses of the concept-word ‘reality’ is that they are relative to the contrasting contexts given by sceptical hypotheses, hallucinations, dreams and other ways of producing counterfeit realities. The insistence on believing that there is an adherent sense of reality in the ultimate or absolute, non-relative sense, leads us astray. Considered in this way, the meaningfulness of questions concerning ultimate adherent reality is an illusion.
We are now prepared to advance our objections to the sceptical argument. In what follows, I will show that the skeptic is able to infer that we do not know the adherent reality of any trivial proposition p in an absolute sense. We have already seen that although this is true, it is trivial, and the skeptic wants something more. What he wishes to prove is that we paradoxically do not know any world or fact in the world in its inherent reality, namely, in its actuality, its concreteness, its independence, its publicness. However, his attempt is doomed to failure, since the word ‘reality’ is used in its adherent sense in the premises and in its inherent sense in the conclusion, configuring an equivocal and therefore fallacious argument.
Refuting the Argument from Ignorance
Now we come to the last stage of our argument, which consists in applying the semantic distinctions between different kinds of attributions of reality to the sceptical and anti-sceptical arguments. Consider first the expanded form of the argument from ignorance about the external world:
1. If I know that I have two real hands, then I know that I am
not in reality a brain in a vat.
2. I don’t know whether I am not in reality a brain in a vat.
3. Therefore: I don’t know whether I have two real hands. (1, 2 MT)
At first glance, this more explicit form of the argument is also flawless. However, it is not difficult to show that in this form – and consequently as well in its original form – the argument cannot stand up to scrutiny. We can show this by specifying the kinds of reality attributions it makes. This gives us two direct ways of interpreting attributions of reality that seem to make sense, a weaker and a stronger one. I will argue to show that it is the second that is properly intended by skeptics.
(a) Under the weak interpretation of the attributions of reality in the argument, the skeptic tries to convince us that we cannot know that all the knowledge we think we have of things belonging to the external world is not part of a universal illusion. In this case, all the reality attributions in the argument would be understood as belonging to the adherent sense, as shown by the following formulation:
1. If I know that I have two (adherently) real hands, then I know that I am not in (adherent) reality a brain in a vat.
2. I do not know whether I am in (adherent) reality a brain in a vat.
3. Therefore: I do not know whether I have two (adherently) real hands.
Premises 1 and 2 are contextually understood as involving adherent attributions of reality, which makes them true, producing what I think is a sound argument (which would not be the case if the attributions of reality they make were all inherent). Indeed, it is true that ultimately I cannot know that I am not a brain in a vat in a non-relative adherent sense of reality, and consequently, I do not know whether anything in the world is ultimately real in the adherent sense, including that I have two adherently real hands. Nonetheless, the conclusion that we cannot know the ultimate adherent reality of those things is quite innocuous and comes as no surprise. As we have already seen, to deny knowledge of the ultimate adherent reality of our world (and consequently of its parts) amounts to the same thing as to deny that we could notice that the whole world doubled in size last night, if this was not in any way detectable, or to deny that I can know that my brother died the day after tomorrow... Lacking criteria for their truth, these statements are devoid of sense, and to deny that we have knowledge of what they state is to say something true, namely, to reject knowledge of something that lacks factual content. Hence, the impression that we are losing something important would be false, arising from a confusion with the special case of relative attributions of adherent reality, which can only be made against the improbable background of a sceptical scenario that has been shown to be true. Moreover, in this weak interpretation, our knowledge of the inherent reality of things and states of affairs constituting the world, of their concreteness, which is important for our lives, remains unchallenged. I believe that healthy sceptics would approve this. However, this paltry conclusion could not satisfy the usual full-blooded skeptic that defends the argument from ignorance.
(b) It is clear that the skeptic is trying to express a stronger understanding of the argument. Thus, as in case I-b, the skeptic understands the first two premises as sentences concerning adherent reality, since this is the natural way to understand them as they are formulated in a sceptical context, understanding by ‘natural way’ our tendency to read sentences and even arguments in ways that seem to make them true. But he understands the third sentence, the conclusion, as concerning the inherent reality of the referents of p. This is also a natural and non-trivial way of understanding this sentence, since p is usually asserted in ordinary and non-sceptical contexts. As a result, the instantiation of the argument from ignorance will be as follows:
1. If I know that I have two (adherently) real hands, then I know
that I am not in (adherent) reality a brain in a(n) (adherently)
2. I don’t know whether I am not in (adherent) reality a brain in
a(n) (adherently) real vat.
3. Therefore: I don’t know whether I have two (inherently) real
hands. (1, 2 MT)
This is the most persuasive understanding of (I-a), not only for the skeptic, but for all of us, since we are also naturally inclined to interpret arguments in such a way that they lead to new, surprising, non-trivial conclusions.
However, the argument is obviously equivocal, because the kind of attribution of reality made in the premises is different from the kind of attribution of reality made in its conclusion. Because of this, although the premises are true, the conclusion is false. Here we are led to inadvertently shift from the understanding of the words ‘real’ and ‘reality’ in the adherent sense in the first and second premises to an understanding of the word ‘real’ in the inherent sense in the conclusion, which rules out the use of the modus tollens.
Taking advantage of our lack of awareness of the two senses of reality involved, the full-blooded skeptic is confusing us with an argument in which the suggested conclusion is that we cannot know the reality of p in the usual inherent sense, or, in our example, that I cannot know that I have two inherently real hands. He is suggesting that we do not know any state of affairs in our world as a concrete reality, that we cannot know our hands as having the maximal perceptual intensity, independence of the will, truly interpersonal access, etc., and that the plain man is wrong when he claims to know that he has two real hands or that our world really exists! The feeling of awkwardness produced by the argument from ignorance arises from this suggestion.
Finally, using ‘ir’ to indicate the inherent attribution of external reality in the statement and ‘ar’ to indicate the adherent attribution of external reality, the stronger form of the argument from ignorance can be expressed in symbolic shorthand as follows:
1 Kpar → K~har
3 ~Kpir (1, 2 MT)
The inescapable conclusion is that the most incisive form of the sceptical argument about the external world is either sound, but innocuous and trivial, or surprising and non-trivial, but invalid. In the last case it rests on a subtle fallacy of equivocation that falls apart when subjected to a sufficiently careful semantic analysis of what is involved in our ordinary use of words.
Refuting the argument from knowledge
Now we will turn to the argument from knowledge, according to which because we know p, we also know that h is false. Here too, we can find both a weak and a strong interpretation of the argument. The most natural and persuasive interpretation of the argument from knowledge is the stronger. According to this interpretation, the argument makes explicit kinds of attributions of reality that seem reasonable in context and imply the truth of the premises. The assumptions in argument (II), as intended by the ‘dogmatic’ anti-sceptical philosopher, can be made sufficiently explicit as follows:
1. I know that I have two (inherently) real hands.
2. If I know that I have two (adherently) real hands, then I know
that I am not in (adherent) reality a brain in a(n) (adherently)
3. Therefore: I know that I am not in (adherent) reality a brain in
a(n) (adherently) real vat. (1, 2 MP)
This understanding of the argument seems plausible, first because each premise is naturally understood as a true statement, secondly because its conclusion seems to be new, surprising and non-trivial. However, even though the two premises are true, I can’t apply the modus ponens to state the conclusion that I know that it is ultimately true that I am not a brain in a vat, since the attribution of reality in the first premise can only be inherent.
Here, too, we can represent the form of the argument symbolically in a way that shows that it is equivocal and thus fallacious:
2 Kpar → K~har
3 K~har (1, 2 MP)
Before concluding this section, I wish to comment on G. E. Moore’s attempt to prove the existence of the external world. This argument was conceived as a refutation of idealism and I believe it can be reasonably understood as a variant of the argument from knowledge interpreted in a weaker and acceptable form. Consider Moore’s own words:
I can prove now, for instance, that two human hands exist. How? By holding up my two hands and saying, as I make a certain gesture with the right hand, ‘Here is one hand’, and adding, as I make a certain gesture with the left, ‘And here is another’. And if, by doing this, I have proved ipso facto the existence of external things, you will all see that I can also do it now in a number of other ways: there is no need to multiply examples.
Since Moore explicitly states that he does not intend to refute the skeptics, but only to prove the real existence of the external world, we can summarize what he intends as:
1. I know that I have two inherently real hands.
2. If I know that I have two inherently real hands, I know that
there is an inherent reality around me (at least concerning my
3. Therefore: I know that there is an inherent reality around me.
(1, 2 MP)
There is nothing wrong with this argument, which hints in the direction of the proof of the external world already presented in this paper. Taken in this way, Moore’s argument has some force against the idealist, as he originally intended, insofar as the idealist (like the skeptic) is trying to build on the notion that our world is made of such stuff as dreams. However, this argument would be too weak if used to prove the falsity of any sceptical hypothesis, since these hypotheses are concerned with reality in the adherent sense.
I conclude that both the argument from knowledge and the argument from ignorance, strictly considered, are misguided attempts to prove something that cannot be proved. We can know neither so much nor so little. This is why the sceptical and the anti-sceptical problems about the reality of the external world can simply be dismissed, I believe, as subtle forms of linguistic-conceptual philosophical bewilderment.
Argument from Ignorance Concerning our Past Reality
An analogous argument from ignorance can be applied to limited forms of scepticism, such as that about the past. Consider Bertrand Russell’s remark that it is possible that our whole world, including ourselves with all our memories, was created only five minutes ago. Can we know that this hypothesis is false? Not with epistemological certainty. In this case, how can we know whether anything was the case before those five minutes? The argument from ignorance concerning the past can be instantiated as follows:
1. If I know that the French Revolution occurred in 1789, then I
know that the world existed before the year 2000.
2. I do not know whether the world existed before the year 2000.
3. Therefore: I do not know whether the French Revolution
occurred in 1789 (1, 2 MT).
Our answer to this sceptical argument resembles our previous one. We need to distinguish between two senses of our attributions of reality to a past occurrence. The first is an inherent sense, which depends on the insertion of the past occurrence in our historical world – a long succession of inter-related events. In this sense, I can claim to know that the French Revolution really did occur in 1789: there is enough documentary evidence available to prove the inherent reality of this past historical occurrence. What makes this claim possible is once again an extended application of our standard criteria for external reality, this time applied to the past. The criteria for the inherent reality of historical events are given by documentary and physical historical evidence. As such they are secondary criteria, functioning as indirect ways to warrant that the past was real in an extended inherent sense. This means that we are inductively assured that the past events would satisfy our standard inherent criteria of reality (like the greatest intensity of experience, independence of will, interpersonal access…) if observers like us could be located in the right time and place in the past. Indeed, to say that the French Revolution occurred in 1789 amounts to the same thing as to claim that the historical evidence shows that if observers like ourselves were living at the right time and place, the standard criteria for external reality regarding the events that constituted the French Revolution could have been satisfied in more direct ways.
However, there is also an adherent sense of real past existence. This sense is external to our historical world and could only arise in the context of a sceptical hypothesis about the past. Here again, we can conceive of both a relative and an absolute sense of external adherent reality. To give examples of the relative sense we can imagine some situation in which we could obtain sufficient evidence that our world didn’t exist before the year 2000, showing that all our memories before this date were illusory, and that all the relevant historical evidence before this year was counterfeit. For example: we and our memories and documents and our world and the galaxies and the whole expanding universe have been produced as an experiment in a macro-macrocosmic laboratory, as beginning in the year 2000. We would disattribute the adherent reality of the French Revolution based on the acceptance of this information.
However, outside such contrastive circumstances, that is, in an absolute adherent sense, it is trivially true that we cannot know that the world existed before the year 2000 or that there was a French Revolution in 1789, in the same way that we cannot know that the world has doubled its size during in the past night. In the adherent sense there is likewise no way to obtain knowledge of the ultimate reality of the past, and we could construct a weak skeptical argument to proof this triviality.
The clue to answering the strong understanding of the skeptical argument is the same here: the sceptical argument is an attempt to confuse us, shifting equivocally from our ignorance of the adherent reality of our past to ignorance of the inherent reality of historical events as a conclusion. What follows is the most reasonable interpretation of the argument (because it seems sound and non-trivial) reconstructed in a way that exposes its equivocal character:
1. If I know that the French Revolution really occurred in 1789
(in the adherent sense of reality), then I know that the world
was (adherently) real before the year 2000.
2. I do not know whether the world was (adherently) real before
the year 2000.
3. Therefore: I do not know whether the French Revolution
really occurred in 1789 (in the inherent sense of reality). (1, 2
To ensure the credibility of the argument, we selected two true premises dealing with adherent senses of reality. However, we cannot really apply the modus tollens to them in order to reach the conclusion, because the sense of our attribution of reality in the conclusion is different from the sense of this attribution in the premises, making the conclusion equivocal and consequently fallacious.
We could also have a weak understanding of the argument III. In this case we would read all its attributions of reality as adherent in a supposed absolute sense. In this case the argument’s premises would remain true and it would be sound and unequivocal. But although sound its conclusion would be totally trivial. For we do not need to do any logical reasoning to realize that in an absolute sense we cannot know that the French Revolution occurred in 1789.
Why the principle of closure seems threatened
Sometimes the problem we have dealt with is simplified to arrive at three statement-forms composing an inconsistent set:
(2) K(p → ~h),
This is paradoxical, because although each of the three statement-forms seems true (in the sense that its instantiation would render a true statement), one of them must be false. Thus, the skeptic denies (1), since he accepts ~K~h, which with K(p → ~h) entails ~Kp. The anti-skeptic denies (3), since he accepts Kp, which with K(p → ~h) entails K~h. There are also more neutral philosophers who reject link (2) in order to accept (1) and (3). These philosophers do this by rejecting the principle of closure under known entailment, which states that ‘Kp & K(p → q) ├ Kq’. Since the principle of closure is intuitive, and since K(p → q) seems to be the same as (Kp → Kq), so that the principle turns out to be ‘Kp & (Kp →Kq) ├ Kq’, which is obviously true, the rejection of this principle seems to be a rather high price to pay.
By subjecting this paradox to an analysis of its kinds of attributions of reality, we can conclude that the three statements can be seen true without impairing the principle of closure. The argument is the following: Based on our analysis, the first statement-form, Kp, is more naturally understood as attributing inherent reality to what p asserts, and can be seen as true. The third statement-form, ~K~h, is more naturally understood as concerning a lack of knowledge of the ultimate adherent reality of the world considered by h, and in this case it is seen as true. The second statement-form, K(p → ~h), could be made true if the reality of the antecedent and the consequent were both seen as adherent (both, Kp and K~h are seen as false, we don’t know as much, hence the implication is true; the same being the case with the pretense knowledge of the whole implication of senseless antecedents and consequents). So understood, the consistence of the set can be made clear by the following instantiation:
(1’) I know that I have two inherently real hands.
(2’) I know that if I have two adherently real hands, then I am not an adherently real brain in an adherently real vat.
(3’) I don’t know whether I am not an adherently real brain in an adherently real vat.
From (2’) and (3’) we can conclude by modus tollens (4’) ‘I don’t know whether I have two adherently real hands’, another trivial truth. Nonetheless, since this does not result in the denial of (1’), it seems clear that the set is not only composed of true statements: it is also a consistent set.
If our argument is essentially correct, then the contemporary forms of the sceptical and anti-sceptical arguments about our knowledge of the external world only seem to make sense. This is not due to a lack of logical syntax, but because this syntax is frivolously used, without enough semantic and pragmatic reflection.
 See Peter Unger, Ignorance: A Case for Scepticism (Oxford: Oxford University Press, 1975), chap. 1.
 If you are impressed by Hilary Putnam’s argument against the possibility that you are a brain in a vat, you can choose a different sceptical hypothesis, or imagine that you are a brain recently consigned to a vat (since philosophers usually hold that the argument can in this way be neutralized). Nonetheless, we should remember that Putnam’s argument is controversial. He argues that the hypothesis that we are brains in vats is self-refuting. If I am a brain in a vat, I cannot think that I am a brain in a vat, because I cannot have thoughts about brains, vats, water, trees… In order to have these thoughts, I need to have causal contact with these things. Thus, if I am a brain in a vat, the best I could possibly have would be thoughts referring to something like electrical patterns (‘objects in images’) generated by the computer program. In this case, he thinks, the reference to vats, water, trees, would be as illusory as the image of Churchill accidentally drawn by an ant walking across sand… lacking either an intention or an adequate cause. Since we are able to entertain the thought that we are brains in vats, we cannot be brains in vats.
The problem with Putnam’s argument is that it ignores the flexibility of language. Indeed, his argument unreasonably assumes that the representation accidentally generated in a brain, as a counterpart of electrical patterns in a computer, cannot have a referential function analogous to the representation generated in a brain by the experience of the real thing, since only the latter representation is properly caused by the real thing. However, why cannot the representation of a brain in a vat be referential in a similar way, although in fact misleadingly referring to something that exists only as electrical patterns in the computer? We can compare the brain image of a tree produced by the electrical patterns in the computer with the brain images we have of a tree in front of us. They are (i) qualitatively identical, as he admits. But there is not enough reason to think that they are not also: (ii) caused, the first by the electrical patterns (even without any relationship to what could make the statement ultimately true), the second by light reflected from a real tree, and that they are not also: (iii) both intended by the brain to represent what they seem to show, namely, trees, although in the first case equivocally. Furthermore, the comparison between the phenomena occurring in the brain in a vat and the case of the accidental drawing of the image of Churchill by an ant is utterly misleading, since not only does the ant have no intentions, it is not even being caused to copy anything. In my view, Putnam’s argument only convinces those who are already determined to be convinced. It rests upon a dogmatic application of his semantic externalism, the doctrine suggesting that meanings and thoughts are in some way suspended in the domain of external things – a doctrine already objected to in the present book. See Hilary Putnam, Reason, Truth and History (Cambridge: Cambridge University Press, 1981), chap. 1.
 Immanuel Kant, Kritik der reinen Vernunft (Frankfurt: Suhrkamp, 1968 (1787)), B 106.
 Rudolf Carnap, ‘Empiricism, Semantics and Ontology’, in his Meaning and Necessity: A Study in Semantics and Modal Logic (Chicago: University of Chicago Press, 1958 (1947)), p. 207.
 Rudolf Carnap, ‘Empiricism, Semantics and Ontology’, ibid., p. 214.
 Barry Stroud, The Significance of Philosophical Scepticism (Oxford: Oxford University Press, 1984), chap. 5.
 See P. F. Strawson’s comments on Stroud in ‘Scepticism, Naturalism and Transcendental Arguments’, published in his Scepticism and Naturalism: Some Varieties (New York: Columbia University Press, 1985), p. 7.
 T. S. Eliot, ‘Unreal city under the brown fog of a winter dawn’, in The Waste Land, ‘The Fire Sermon’, lines 200-210.
 Wittgenstein’s thesis seems stronger, since in many cases he identified meaning with use (Gebrauch). See his Philosophische Untersuchungen (Frankfurt: Suhrkamp, 1983), sec. 43. Later, however, he identified meaning with ways of using (Gebrauchsweisen) or ways of applying (Verwendungsweisen) words, and to be sure not with their episodic uses. See Über Gewissheit (Frankfurt: Suhrkamp, 1983), sec. 61. This allows us to identify these ways with semantic rules (or combinations of rules) determining the episodic uses. Examples of such rules would be the so-called criterial rules, which will be further considered in this text.
 This is why Wittgenstein has much to say about scepticism, although I do not agree very much with his views about it, and I do not adopt them here.
 For Wittgenstein, criteria ‘give our words their common meaning’ (The Blue Book, Oxford: Basil Blackwell 1958, p. 57). His doctrine about criteria is scattered throughout his manuscripts. Important passages are found in The Blue Book, pp. 24-25, in his Philosophische Untersuchungen, sec. 354, in Zettel (Frankfurt: Suhrkamp, 1984), sec. 438, in Wittgenstein’s Lectures – Cambridge 1932-35 (New York: Prometheus, 1979), p. 28. It is worth noting that the thesis that criteria are constitutive of meaning would have no point if we had in mind only objectively given criteria. However, by criteria he also means the conditions intrinsically required in the criterial rules (or combinations of rules, which is one way we could understand his doctrine of meaning as a calculus). These should be conceptual and verificational semantic rules – and not just the objectively given criteria that might or might not satisfy these rules, making them applicable or not. For an investigation of criterial rules and their semantic role, see G. P. Baker, ‘Criteria: A New Foundation for Semantics’, in Ludwig Wittgenstein: Critical Assessments, (ed.) Stuart Shanker (London: Croom Helm, 1986), vol. 2, pp. 194-225. See also the last chapter of P. M. S. Hacker, Insight and Illusion: Themes in the Philosophy of Wittgenstein (Oxford: Thoemmes Press, 2nd ed., 1986).
 John Locke, Essay Concerning Human Understanding, (ed.) P. H. Nidditch (Oxford: Oxford University Press, 1975), Book IV, chap. 11.
 George Berkeley, Three Dialogues Between Hylas and Philonous (1713), Complete Works, (eds.) A. A. Luce & T. E. Jessop (London: Thomas Nelson and Sons, 1948-57), III, p. 235.
 David Hume, A Treatise of Human Nature, Book I, Section 1.
 Immanuel Kant, Prolegomena, § 16.
 J. S. Mill, An Examination of Sir William Hamilton’s Philosophy (London: Longmans, Green & Co., 1889), chap. XI.
 Gottlob Frege, ‘Der Gedanke: eine logische Untersuchung’, originally published in Beiträge zur Philosophie des deutschen Idealismus, I, 2 (1918), 58-77.
 G. E. Moore, ‘The Meaning of Real’, in his Some Main Problems of Philosophy (London: George Allen & Unwin, 1953).
 Sigmund Freud, ‘Formulations on the two Principles of Mental Functioning’, in The Standard Edition of the Complete Psychological Works of Sigmund Freud, (ed.) James Strachey (London: Hogarth Press, 1958), Vol. 12.
 This is how Laurence BonJour objects to the criteria of reality proposed by Locke. See BonJour’s book, Epistemology: Classic Problems and Contemporary Responses (Lanham, MD: Rowman & Littlefield, 2002), pp. 130-135.
 If they are really given or only supposed to be given in an otiose question, since their adequate experience is what defines their being given. The concept of criteria in a definitional sense (primary criteria) is explained by Wittgenstein in The Blue Book, p. 24.
 As Aristotle explained: ‘There are many senses in which a thing may be said to “be,” but they are related to one central point, one definite kind of thing, and are not homonymous. Everything which is said to be “healthy” is related to health, one thing in the sense that it preserves health, another in the sense that it produces it, another in the sense that it is a symptom of health, another because it is capable of it’. Metaphysics, 1003a 33-37, in J. Barnes (ed.): The Complete Works of Aristotle (Princeton, NJ: Princeton University Press, 1984), Vol. II, p. 1584.
 In the present world, we need an extended concept of testimony: it must include not only what others state and have written about, but also what we can know through electronic means, etc. I will use testimony here in this wide sense, so that, for example, a hidden camera can offer testimony of an event.
 A precise and detailed reconstruction of the ways in which we obtain knowledge of external reality would require an empirical study of how the concept of external reality is gradually learned, etc. Since my purpose here is to make a general defense of philosophical realism and to answer skeptics, my suggested reconstruction must suffice.
 Irwin Copi, Introduction to Logic (New York: Collier-Macmillan, 1972), p. 93.
 Barry Stroud, The Significance of Philosophical Scepticism, chap. 5.
 More plausibly, David Deutsch has argued for the greater simplicity of the hypothesis that our world is the ultimately real one. According to him, sceptical hypotheses require a world that is actually far more complicated than the idea that our world is the real one, for they are parasitic on the latter. This would require more complexity than we find in the external world as we know it. Consequently, the sceptical hypothesis is not as simple and parsimonious as what he calls the commonsense view and for this reason should be rejected. See David Deutsch, The Fabric of Reality (New York: Penguin Books, 1995).
In my view simplicity is surely an epistemic virtue insofar as it is theoretical, as when we make comparisons between scientific theories with the same scope. However, it does not seem to be an epistemic virtue with regard to the factual simplicity of concrete states of affairs known and unknown to us. Compare the claim that there is only one egg in a basket with the claim that there are twelve eggs. Just because the first alternative is factually simpler does not make it more probable than the second. Hence, since Deutsch’s simplicity is factual rather than theoretical, it does not appear to count as a reason for the truth of the claim that the reality of our external world is the ultimate one. We just don’t know what is outside there.
 This sentence sounds sufficiently intuitive to me. Nonetheless, I believe that it could be further analyzed as the following conditional: ‘In a system where I could know that I had two adherently real hands, I would probably also know that I was not, in the system’s adherent reality, a brain in a vat, but rather an adherently real human being possessing an adherently real body…’
 The first premisse is true because the antecedent and the consequent of its implication are false.
 In fact, religions have often contrasted our corrupt, sinful world with the real one, which exists beyond the senses. Much of the philosophical tradition, from Plato to Plotinus and from Berkeley to Hegel has tended to treat the external world as less real than it really is. There are social-psychological reasons for this attitude, which cultural critics from Nietzsche to Freud have attempted to clarify.
 It would be hasty to conflate my proposal with current forms of contextualism. I do not base my considerations on the different strengths of knowledge (see Keith DeRose, ‘Solving the Sceptical Problem’, Philosophical Review, 104 (1995), 1-52), nor do I maintain that the context changes the ‘angle’ of scrutiny. See Michael Williams, Unnatural Doubts: Epistemological Realism and the Basis of Scepticism (Princeton, NJ: Princeton University Press, 1996). My approach is instead to change the focus from the concept of knowledge to the concept of external reality, investigating its (contextually relative) uses or senses in terms of application criteria. The result is much more precise, pressing the soft spot of the sceptical argument. Finally, I do not believe that my approach is incompatible with moderate foundationalism.
 G. E. Moore, ‘Proof of the External World’, in his Selected Writings, (ed.) Thomas Baldwin (London: Routledge, 1993 (1939)), pp. 165-6.
 See Bertrand Russell, Analysis of Mind (London: Routledge, 1989), lecture 9.
 Of course, we can also construct an anti-sceptical modus ponens counterpart of this argument, which would be correspondingly equivocal.
 Here again, the first premise if true because the antecedent and the consequent are false.
 The validity of the principle of epistemic closure has been challenged by anti-sceptical arguments, particularly those of Robert Nozick, in Philosophical Explanations (Oxford: Oxford University Press, 1981), pp. 240-5 and Fred Dretske, in ‘Epistemic Operators’, Journal of Philosophy 67 (1970), 1007-1023. Our treatment of the sceptical problem leaves this principle untouched.