The analysis in the first part of this enquiry established that it is not, or at best only partly, possible to map sentences modified by sentence adverbs, such as ‘Unfortunately, it is raining’, on to the logical structure of sentences representable by the notation S A P, such as ‘Caroline regrets [(that) it is raining].’ It should be obvious why this is the case. Truth conditions do not yield meaning. Donald Davidson has, throughout his professional life as a philosopher, been an advocate of the thesis that truth conditions indeed are able to yield a complete theory of meaning for sentences1, while the meaning of the sentence was supposed to somehow depend upon its parts. But this account is at best problematic, and at worst wrong. For instance, the meaning of words changes with their use in different contexts. Frege’s distinction between sense and reference offers no satisfying solution to this problem. Rather, the contextual use of a word apparently determines the meaning of a word in a sentence, the meaning of which depends, in turn, upon a social or textual context. The referents, or meanings, of sentences are not, as Frege thought, truth values; and even if they were, it would not follow that all sentences mean – in the sense of refer to – the same thing, namely, truth, for truth is not an independent entity but an abstract idea.2
Furthermore, it does not suffice for someone to know what has to be the case for a sentence to be true to know its meaning, as Davidson was convinced. Even as a competent speaker of a natural language, I can know what has to be the case for a particular sentence to be true without knowing what the sentence means. Being a competent speaker does not necessarily entail knowing every word a natural language may contain; thus there can be sentences a competent speaker of such a language does not understand because they do not know some of the words the sentence comprises, although they know the concepts and thus know what has to be the case for the sentence to be true.
Let us now return to Wittgenstein.
In §569 of the Philosophical Investigations we read:
Language is an instrument. Its concepts are instruments. Now perhaps one thinks that it can make no great difference which concepts we employ. As, after all, it is possible to do physics in feet and inches as well as metres and centimetres; the difference is merely one of convenience. But even this is not true if, for instance, calculations in some system of measurement demand more time and trouble than it is possible for us to give them.3
Yes, language is an instrument, a tool, as it were, for communicating what we think in the broadest sense. We need not communicate everything, to be sure, if we abstain from the naive construal of thinking as an inner monologue. Thinking is not always, albeit sometimes, equivalent to silently talking to oneself. Most of the processes in our brains are not even accessible to consciousness.
Indeed, Wittgenstein’s contention that ‘Whereof one cannot speak, thereof one must be silent’ is ambiguous. There are things whereof we cannot speak because they are not possible yet, not because we could not speak of them in principle. People could not speak of rockets and spaceships in antiquity and the middle ages, not because it would not have been possible to do so in principle, but because, on the one hand, there were neither rockets nor spaceships, and, on the other hand, they lacked both the technological and the physical knowledge of today. Science as we know it did not exist, and their worldview was mostly based upon metaphysical speculation. Therefore, they could not speak of these things, although they were obviously possible in principle, for otherwise we should not be able to use and talk about them now.
On the other hand, if we construe thinking as broadly all the processes in our brains which are concerned with interpreting and reacting to incoming information, it is true that there are many things, indeed, the most part of our thinking, whereof we cannot speak. In the overwhelming majority of all cases, we remain oblivious to how we know or why we do something. It often seems to us that we do know, to be sure, but this is usually owing to nothing more than the notorious phenomenon of postrationalization. It is easy to rationalize events, that is to say, to explain them in a rational manner, after they have happened. In other words, we construct a rational explanation by inventing reasons and possible causal chains.
1. Confer, among many: Davidson, Donald: ‘Truth and Meaning’; originally published in Synthèse, 17; Dordrecht-Holland: D. Reidel Publishing Company, 1967, pp. 304–323; reprinted in Davidson, Donald: Inquiries into Truth and Interpretation; Second Edition; Oxford: Oxford University Press, 2001; pp. 17–36.
See also: Davidson, Donald: ‘A Coherence Theory of Truth and Knowledge’; reprinted in: Reading Rorty; Edited by Alan Malachowski; Cambridge: Blackwell, 1990. Reprinted in Davidson, Donald: Subjective, Intersubjective, Objective; Oxford: Oxford University Press, 2001/2002; pp. 137–153.
2. I shall not discuss these aspects in further detail here. For further information, see, for instance, the respective entries in Stanford Encyclopedia of Philosophy: http://plato.stanford.edu/entries/frege/#FreLan (Frege’s Philosophy of Language), http://plato.stanford.edu/entries/davidson/#Tru (Donald Davidson on Meaning and Truth), and Internet Encyclopedia of Philosophy: http://www.iep.utm.edu/frege/#H4 (Frege’s Theory of Sense and Reference), http://www.iep.utm.edu/dav-lang/ (Donald Davidson’s Philosophy of Language).
3. Wittgenstein, Ludwig: Philosophical Investigations; The German Text With a Revised English Translation; Third Edition; Translated by G. E. M. Anscombe; Oxford: Blackwell Publishing, 2001; p. 128e. (Emphases in the original.)