SB logo
en's Homepage

Is Verificationism True?

It seems very likely to me that in its most basic sense, verificationism must be true. Whilst it certainly needs lots of work to make it more specific and precise, since many of the objections brought up about it seem perfectly fair, at least in principle the system must hold, since it is trying to explain some very basic facts about what we mean by meaning. One of the simplest things that Ayer seemed to be trying to point out in Language, Truth and Logic was the idea, about which you wouldn't expect any dispute (though some appears to have existed) that when someone makes a statement about something they are suggesting not just that something is true or false, but also that it makes some sort of difference. Ayer points out that when someone says something, if they mean anything by it, they must mean that something is in some way different because of it, since otherwise the statement would be one made in complete isolation, without being about anything in the world, and, he suggests, this only amounts to making some noise, though more sophisticated noise that just babbling.

In many senses of the word this is already how we use the word 'meaning'. When someone says 'But what does that actually mean?' they are most likely asking what effect something will actually have on the world - they aren't asking what you are trying to say, but how they should change their view of the world, or their actions in the future, if what you say is true. Ayer merely tries to make it clear that in fact, in every case, we use meaning in this way.

There often seems to be some confusion when people hear that he was suggesting that what something means is equivalent to how we find out whether it's true or not, but this seems to be a misunderstanding of what he was saying. When we say 'The sky is blue' we do not mean by that 'Look up!' (and obviously you wouldn't shout 'The sky is blue!' to advise someone to step out the way of a falling object) - what Ayer was saying is that to say 'The sky is blue' is the same as saying 'If you were to look up you would have the sensations of the expanse of visible atmosphere being the colour you have come to know as blue'. Ayer suggests that equivalent sentences such as 'The sky is bluorange' mean nothing simply by virtue of the fact that you couldn't know how to decide whether this was the case or not. If the speaker went on to say ' bluorange I of course mean partly blue and partly orange' then we would now know how we could decide whether or not the sky was in fact bluorange - through a specific empirical test - and the sentence would be significant.

Berlin complains that statements such as 'There are mountains on the other side of the moon' should not be considered insignificant purely for the fact that when he was writing it was not possible to determine whether this was so or not. But this is not what Ayer means by saying verifiable - what he means is that there is some way of verifying it, and that we can describe what that way is. It should have been possible for Berlin to say how we would go about proving that there were mountains on the other side of the moon, and, as proved by events since his writing, it has been possible to show that this is indeed the case. What Ayer means is that there is some possible way of verifying it, not that we have to go about verifying everything. All he is saying is that it often doesn't make a difference whether something is the case or not, and if it does, then obviously this would have to have some effect on the world, and one that is detectable.

It might, equivalently, be interesting to consider the proposition that 'When Berlin was writing his works, there were no mountains on the other side of the moon'. If Berlin had considered this question, then, most likely, he would have found it impossible to verify whether it was true or not. If he were not able to suggest a way to find out at that time that there had been mountains on the other side of the moon, then in the case of this particular statement, the evidence might all have gone by the time it were possible to test the theory. It might be the case (though it seems unlikely) that the other side of the moon was flat when Berlin was writing, and only became mountainous between that time and those when space shuttles allowed us to see for ourselves. Whilst there might remain evidence (such as floating debris) that the other side of the moon had been smooth, Berlin, again, could probably have predicted this, but if he couldn't then it would render the statement unverifiable. More importantly, however, if it was never possible to discover that the other side of the moon had been smooth, then it is fairly obvious this could not have an effect on anything. If the tides were affected, for example, then this in itself could be the observable indication that some change had occurred, but if no change is detectable, that must go hand in hand with the idea that any difference is entirely insignificant and inconsequential.

A more interesting objection that Berlin brings up is that of a statement containing a disjunction, about a certain premise. He wonders what verificationists would have thought of a statement in the form 'If I stay then such-and-such, or else something else'. He points out, probably quite rightly, that in this case it isn't really possible to verify both things - it's not just that it would be difficult, but that there is a logical impossibility involved in investigating what would happen if you stayed and what would happen if you went, since you couldn't do both at once. To make the claim that this would mean statements about one of the options is not significant is clearly unfair - this would mean that you would never be able to weigh up the pros and cons of staying or going, on the basis of the fact that statements about it would be meaningless.

The solution, however, is to claim that Berlin is trying to claim that the complex statements '(A -> B) & (~A -> C)' and '(A -> B) v (~A -> C)' are one and the same. It should be clear that the first statement isn't going to mean the same as the second, and it is, in fact, the second that we are discussing. It is possible to say 'If I stay then I'll see my friend, and if I go I'll get some work done', but the 'and' in this sentence is not, in fact, having the simple conjunctive use that it appears to have - clearly what is actually meant is 'If I stay then I'll see my friend, or if I go I'll get some work done'. People are not seriously suggesting that there is a truth value for the statement 'If I stay then...' and 'If I don't stay then...' at one time, but instead consider the instances of stay and not-stay separately. Clearly in this case it would be possible to verify each instance - we can explain how to verify what would happen if we stayed (we'd stay), and we can explain how to verify what would happen if we didn't (we'd go) - sure, we can't do them both at the same time, but this isn't required - we can certainly verify one or the other.

Quine suggested that there are major problems for verification in the form of various statements that have significance but it isn't possible to verify, giving the example of the expansion of the universe. He states that we would receive the same data if the universe were uniformly expanding constantly as if it was staying the same size, and yet he seems to be missing the point that this is exactly the sort of situation that verificationists are discussing. It the universe is expanding at a constant uniform rate then yes, it wouldn't be possible from inside the universe to observe this fact , but clearly in this case it would not be significant that the universe were expanding, and to say so would be uttering a nonsense, if it makes no difference and cannot make any difference.

Again, it seems a little strange to argue with a principle that begins so analytically, and attempts to discuss a priori what we mean by meaning with examples. Clearly if verificationism is right and by saying that a statement has meaning is to say that it makes a difference whether the statement is true or false then clearly it makes no sense to bring up an example and say 'There, that makes sense', unless the example is clearly one where we would be very hard pushed in attempting to claim that contrary to how it obviously looks the statement is meaningless (i.e. a statement which we would very much hold to have a truth value). Quine's example doesn't work, and the very reason that it is interesting - the theory of relativity laid out by Einstein, which assures us that we couldn't ever detect constant motion - tells us that a statement like 'The universe is constantly, uniformly expanding' is insignificant; Einstein himself explained that to say 'Sitting in a chair I am at rest' is no more correct than to say 'Sitting in a chair I am moving at x-million miles' simply because it takes into account the movement around the sun - each statement fails to explain what the movement is taken relative to. So Quine seems to have missed exactly what verificationists are saying - in realising that it would never be possible to prove conclusively that the universe is expanding, since the expansion would have no detectable effects, he should accept that the question of whether it is or not cannot be significant.

Ayer brings up another interesting example - that of the distant past. Whilst things that happened in the past no doubt had an effect on the present, the further back you go the less detectable this effect is likely to be. It seems ridiculous to say that there were far fewer events that had a major impact on the world even a thousand years ago, and surely recent events that are considered to have shaped world history such as the fall of the Berlin Wall might be considered quite minor opposite events that to contemporaries a couple of millennia ago would have seemed very important and influential. Yet it is not possible to decide conclusively on the basis of the current status of the world what exactly happened in the distant past. It appears that verificationists would have to deny that many statements about the distant past are significant, if it is not possible to decide exactly what happened. This would, however, seem perfectly fair. It is not being denied that one thing or another actually happened, what is merely being denied is that making a statement about such things could have any point, and whether any significant statement could be made about past events, whose truth or falsehood would make no difference. Clearly there are events in the distant past about which significant statements can be made - scientists debate about the origins of the Earth, for example - but it should be clear that these are things where people feel there is in existence evidence (however hard it might be to find) which would point to what happened - measurement of the changes in the speed of expansion of the universe are key to the understanding of the origins of the universe, and whilst it might require a long wait, physicists explain that a vast amount of information would be divulged simply by seeing whether the universe continues expanding forever, stops expanding and becomes relatively stable, or collapses back in on itself.

Overall, therefore, it seems likely that the Verification Principle should hold, at least with slight modifications and refinements. It succeeds merely by defining the word meaning in a way that I think is true to how we actually view it, and is useful in exposing statements that appear to mean something, but which closer inspection should makes us realise are in fact meaningless. There are many problems and objections laid against verificationism, yet none of them seem to tackle directly its definition, and the ultimate resource that verificationism could always fall back on is the claim that whilst the verificationists have found a way to debate important questions of philosophy, even if they were to concede truth and falsehood to the statements made by the metaphysicians - their statements simply don't make a difference.