In this paper I will reconstruct and explain verificationism. Subsequently, I will show how the acceptance of verificationism leads to solipsism, the view that one can have knowledge only of their own mental life, by pointing to a necessary difference in meaning between first-person and third-person mental predicates.
The main thrust of verificationism is this. A statement is meaningful only if it is logically verifiable. The only statements that are logically verifiable or knowable are those which reduce to some description of the given. The given is the domain of all that is knowable; it is roughly your perceptual experience at some particular point in time. The given should not be confused with the terms ‘the internal world’ and ‘the external world,’ both of which are meaningless for the verificationist. This is because propositions like ‘there is an external world,’ will turn out to be not logically verifiable (for the truth or falsity of the fact makes absolutely no change in our perceptual experience). All difference in the given is detectable. Because the given is what is presented to you in perceptual experience, there can be nothing in the domain of the given that is undetectable.
Features in the given are describable with atomic words or atomic sentences. Atomic words, like green, pain,1 and so on, can only be known by ‘pointing’ to some feature of our perceptual experience. They cannot be understood in terms of other words. The word’s meaning is established by the agreement of the reactions of others. That is, the use of the word occupies the same relational-role in the given as it is experienced by each of us. For the verificationist, the question of whether the phenomenal quality of his green-experience is identical to the phenomenal quality of my experience, is meaningless. This is because fact is not logically verifiable.
All propositions ultimately reduce to some sequence of atomic words, whose meaning directly describes the given. ‘There is a deer by the bush,’ reduces to ‘there is a brown spot with such-and-such features by that green spot arranged in so-and-so way.’ To see this, suppose that a proposition’s meaning is something over and above its determining some state of affairs in our perceptual experience. If this additional meaning is expressible, then it would be a (complex) proposition (and so nothing over and above an atomic description of some feature of our perceptual experience). But if the meaning is not expressible, then it cannot mean anything, for that which expresses nothing means nothing. So the truth or falsity of a proposition must correspond to a difference in the given in order to be meaningful.
It follows from this that the meaning of a proposition is identical with its verification in the given, for we verify ‘there is a deer by the bush,’ by observing a familiar arrangement of brown situated by another familiar arrangement of green, and perhaps some audible rustle — for this is what is present in the given. So if we cannot conceive of some verification in the given of the fact, then the fact means nothing. So, a proposition is meaningful only insofar as it is logically verifiable. A meaningful statement says that under certain conditions, certain data appear. (For such a statement to be verified re vera, is for there to be consistent agreement in the reactions of a sufficient number of persons to a given stimulus — an agreement that under certain conditions, certain data appear. In this way, hallucinations and illusions will not be verifiable. For example, the meaning of ‘pain’ is grounded in the fact that many people react the same way to a painful stimulus and would describe or talk about the experience using the same or similar words.)
But we should be reluctant to accept verificationism. For if we do, then we fall into solipsism — I can never know that you have a mental life like mine, that the given confronts you as it does me. Here’s how.
To know when to ascribe mental predicates like, ‘is in pain,’ to others, I must know what it means to be in pain. (For if I do not know what pain means, then I could not even have any thoughts about pain.) On verificationism, the meaning of ‘is in pain’ is identical to its method of verification in the given — I see her wince, say ‘ouch,’ or recoil in such-and-such a fashion, for these are the data in the given associated with the fact.2
But how do I know to ascribe ‘is in pain’ to myself — what is the method of verification in first-person ascription? Schlick argues that mental propositions like, ‘I am in pain,’ are verified in just the same way as propositions like, ‘that Sirius has a satellite.’ (99) That I was in pain yesterday can be verified by examining others’ statements about my behavior yesterday, or happening upon a note that I had written reporting my pain.
But this cannot be my method of verification for asserting (in the first-person), ‘I am in pain.’ For I know that I am in pain only when I experience the sensation of pain. If I am not experiencing the sensation of pain, then I cannot know that I am in pain (and so will not ascribe ‘is in pain’ to myself). For suppose I observe my limbs flailing in the pain-characteristic way, but am experiencing no sensation of pain. I would not come to ascribe ‘is in pain’ to myself; more likely, I would just puzzle over why my limbs were moving such a way. Someone else may ascribe ‘is in pain,’ to me on the basis of my behavior — and while such an ascription seems reasonable, I know that it is not true. I may even correct them, saying ‘I am not in pain,’ and, if they take my word, that should count as verification for them of the fact I am not in pain.3
Or, I may suppress all my pain behaviors, this does not entail that I do not experience the sensation of pain and so know that I am in pain. Nor is it after I wince that I know my stomach is in pain, rather the wince is must be a reaction to the sensation. I do not verify that I am in pain by observing my behavior or pain report, rather I verify it by ‘looking inward’ as it were, and finding no sensation of pain. The takeaway is that the methods of verification for ‘X is in pain,’ are necessarily distinct between the first and third-person ascriptions. Self-ascription is on the basis of private sensation; other-ascription is on the basis of the others’ observable behavior or report.
Recall that on verificationism the meaning of a statement says that under certain conditions, certain data appear. So perhaps Schlick would say that the first-person meaning of pain is that, under such-and-such conditions, I met with certain data, viz. pain sensations. And while this is reasonable, it is not the sort of thing that can be verified by anyone else. For it is logically impossible that I can have access to another person’s experience (for my experience is necessarily mine), and by the same token, it is logically impossible that another person can verify that I am in pain via accessing my sensation of pain in my perceptual experience.
Suppose I wish to have been Archimedes, running naked down the streets of Syracuse, crying ‘Eureka!’ I enter an ‘experience-machine,’ and have the experience of running through Syracuse. But this is my experience, not Archimedes’; the referent of my ‘I’ thoughts is me, not Archimedes. For if the referent of ‘I’ were Archimedes, then I would simply be Archimedes and could have no thoughts of my non-Archimedean self to which I could associate my experience-machine experiences. The takeaway is this. If you can genuinely access another person’s experience, then you are simply that person, and if you were to ‘have the experiences of another person’ then that is your experience of what-it-is-like to be someone else and you have not genuinely accessed their experience. Whence the logical impossibility. So I can only verify that ‘I am in pain’ through my experience or sensation of pain (which we have shown is necessarily private), and I can verify that ‘she is in pain’ only through observing characteristic behaviors in my perceptual experience (for if she has any sensations like mine, they must be private to her).
Because the meaning of a proposition is identical with its verification in the given, it cannot be the case that ‘I am in pain’ means the same thing as ‘she is in pain.’ To say ‘she is in pain’ is to say that ‘she is moving her limbs in such-and-such a way’ or ‘if we scanned her brain, we would observe the firing of c-fibers.’ Whereas ‘I am in pain’ is to say that ‘I have a sensation of pain.’ From this it follows that I can never say, ‘she is in pain,’ and mean something like, ‘she is experiencing a sensation of pain,’ for this would not be logically verifiable (from my perspective). When Suzie says, ‘I am in pain,’ what she means is something different from what I mean when I agree, ‘Suzie is in pain.’ When she uses the first-person, she is referring to her sensation, but when I use the third-person, I can only refer to her behavior or report. But if she were mute and suppressing her behavior, she is either experiencing the sensation of pain or not, and while Suzie can verify it for herself, it is impossible that I could ever verify it. Let’s say I take ‘is in pain’ to mean only the third-person sense and take ‘S’ to refer to what I know to be my perceptual experience of the sensation of pain, then I can never know if any other person experiences (some degree of) S when I ascribe ‘is in pain’ to them.
I cannot verify that another person has a sensation of pain in the way that I do. This means that I cannot verify that mental predicates, like ‘is in pain,’ apply to others as I must apply them to myself. But if this is so, then I cannot know whether or not someone is experiencing some private mental sensation (e.g. of pain) as I do. So all I can know is my own perceptual experience and mental states. And this matches the characterization of solipsism that we have provided.
The verificationist may reply the following way. Because our question regarding the ascription of ‘S’ to others is not logically verifiable, it has no meaning. It isn’t false that others have mental lives, nor is it true — we simply do not know what it would mean. All it means to have a mental life is to have a certain neurophysiological structure and behave in a certain way.
This reply, however, is ineffective. If I know my S, then I know that there is a unique kind of mental experience ascribable to me. But if the question of other persons’ mental life is meaningless, then my own mental life is incommunicable, for when I tell someone ‘I am in pain,’ I will mean one thing but be saying another.4 Anything else that I might ascribe a mental life to must mean something other than how I mean it when I ascribe it myself, for the question of others’ having perceptual experiences and sensations is meaningless.
Anything else that I might ascribe a mental life to must mean something other than how I mean it when I ascribe it myself. On the verificationist view, mental life just means some kind of behavior pattern. But when we ask the solipsistic question, we do not mean to ask about behaviors, we mean to ask whether other people have mental lives in the sense that I do, whether pain hurts for them as it does for me. This is outside the given, and unknowable. And this certainly is solipsism.
- For such a description of pain can only amount to something like, ‘pain hurts,’ ‘pain is the opposite of pleasure,’ or ‘pain is what makes you recoil.’ The first is a tautology, the second is almost as trivial, the third overbroad and not necessary, and none of them convey any nontrivial knowledge about what pain actually is to the person who has never experienced it. ↩
- Keep in mind. Schlick maintains that the meaning of a proposition is communicated only if it is verifiable. ↩
- Even so, for them to verify that I am not in pain required a characteristic behavior/report from me, whereas my knowledge came from my sensation. ↩
- Ironically, this creates a situation where, far from worrying about the reality of the mental lives of others, we end up worrying how we could ever communicate to others that we have a mental life. ↩