Anthony Flew, probably the most popular and vocal atheist of the previous generation, once submitted the challenge of the positivists to religious language using the following parable:
Once upon a time two explorers came upon a clearing in the jungle. In the claring were growing many flowers and many weeds. One explorer says, “Some gardener must tend this plot.” The other disagrees, “There is no gardener.” So they set up a barbed-wire fense. They electrify it. They patrol with bloodhounds. (For they remember how H.G. Wells’s ‘The Invisible Man’ could be both smelth and touched though he could not be seen.) But no shrieks ever suggest that some intruder has received a shock. No movements of the wire ever betray an invisible climber. The bloodhounds never give cry. Yet still the Believer is not convinced. “But there is a gardener, invisible, intangible, insensible to electric shocks, a gardener who has no scent and makes no sound, a gardener who comes secretly to look after the garden which he loves.” At last the sceptic despairs, “but what remains of your original assertion? Just how does what you call an invisible, intangible, eternally elusive gardener differ from an imaginary gardener or even from no gardener at all?”
~ Dan R. Stiver, The Philosophy of Religious Language, p. 48
Here, Anthony Flew’s parable is obviously taking a parody of the teleological argument (it is ironic and profoundly significant that it was precisely the teleological argument, more than any other, which convinced Flew of Theism late in life). This challenge is intended to demonstrate to the Believer that her faith is either not meaningful at all, or else it is in principle verifiable. If it is not meaningful at all, she should recognize that she is allowing the misuse of language to confuse nonsense into sense. If it is meaningful, then how does she propose to verify her religious claims? The presumption here on Flew’s part is that no such conclusive verification exists for religious beliefs, which is, I take it, mistaken. Not only is there empirical evidence for miracles, but also for foundational Christian beliefs such as the resurrection. The problem is that if one accepts the positivist’s constraint on semantics, then even verifying events like the resurrection does nothing to make the Trinity meaningful. My intuition tells me that there is something fundamentally wrong with the very theory of semantics which Flew was/is entertaining.
If the principle behind this theory of semantics is articulated naively then it seems self-defeating. For instance, the principle could be articulated as follows: “if some indicative sentence is not an analytic truth, then it must be empirically verified to be possibly true”. The principle response of critics is easy to anticipate: the principle by its own standards cannot possibly be true, since it is neither an analytic truth, and nor is it verifiable empirically.
However, more deeply, this constraint on semantics takes for granted that language cannot extend into metaphysics. For example, if two scientists have two different theories which make all the same predictions and seem to be in principle impossible to distinguish by reason of empirical evidence, then the positivist would want to collapse both of these into the same proposition. The Metaphysical realist, however, has no reason to be satisfied or impressed with that. If it is cognitively meaningful to make a distinction between the two theories then it seems the distinction is not illusory but actual and metaphysical. I think many articles of faith, articulated in religious language, are metaphysical rather than strictly empirical in nature, and it is by reason of this wider parameter for our semantic range that we can understand the sense of doctrines like the Trinity or the incarnation.
In any case, Theism is empirically verifiable, and it is difficult to see how even a strict logical empiricist today escapes Theism (even if they accepted it merely as a feature of an empirical model, rather than a metaphysical postulate). Theism, and even a good portion of religious language, can be in principle verified eschatologically. More significantly they can be verified by communal and personal religious experience.
Perhaps Cardinal Avery Dulles’ suggestion that we must adopt a nuanced symbolic epistemology in order to counteract the reductionistic approach to semantics by suggesting a ‘restorative criticism’ which allows us to avoid reducing the scope of our language and meaningful predication, is right after all. This also allows language to be used beyond the arena of mundane predication (thus art can, properly speaking, actually ‘say’ something, and participate in language).
However, as I said, I think the principle objection which the religious believer, along with the metaphysical realist, will want to put to the positivist here is precisely that their restriction of semantics to radical empiricism is just not acceptable – those of us who take metaphysical distinctions to be cognitively meaningful (even self-evidently) haven’t got any good reason to think such things meaningless.