Here are three principles which seem to remain constant between the Tractatus and the Investigations:
1. You can’t assess (language/thought) from a point of view outside (language/thought); anything you might (say/think) about (language/thought) can only be (said/thought) within (language/thought).
2. (Ordinary language/the propositions of natural science) are fine as they are, and do not stand in need of philosophical foundation or explanation.
3. Linguistic units only have meaning in the context of their significant use.
Here is a big difference:
4T: Significant uses present a picture of the facts.
4I: Significant uses connect words with actions. (“Meaning is use.”)
One magnificent illusion of the Tractatus, if it is that, is its claim that it has no doctrine of sense. Whatever sort of structures we find in our language, whatever pictures of the facts our language presents, are also logical pictures; and logical pictures are the only ones that matter for logic – that the proposition is true if things are the way it says they are (and not otherwise).
But the contrast between 4T and 4I, assuming there is no reconciliation, attacks that way of looking at things head on. If the viewpoint of 4I is tenable, then viewing significant use through the lens of ‘presenting a picture of the facts’ is in fact a doctrine of sense; and a highly distorting one, insofar as (a) there are many things we do with language besides presenting pictures of the facts and (b) when we do use language to present pictures of the facts, it is almost always the case that the facts are presented in particular ways for particular purposes, gaining their significance from the way they are woven into our lives and activities.
From the point of view of the Investigations the Tractatus is trying to build a house with nails and lumber alone, forgetting not only all the other tools you need, but even that you almost always need a hammer to pound the nails.
A few quick interpretative points about these principles:
(A) These are all in some sense debatable theses. I think Wittgenstein can be read as offering good arguments for all of them, albeit perhaps not final and completely decisive ones.
(B) On the other hand, they are also all fundamental theses, and people are very unclear in general about how to debate fundamental theses. Deductive proof is in a sense off the table, since such a proof would undermine fundamentality, or else establish co-fundamentality with other theses which now would need the same kind of treatment. You are pretty much restricted to (a) showing people who believe that they are denying those theses that they run into some sort of absurdity or contradiction in trying to do so and (b) showing how the fundamental theses provide a satisfying organization of our thought and action (that is, holistic justification – alternately known by the titles “inference to the best explanation” and “affirming the consequent” – but it sometimes effects a kind of persuasion even so.)
Am I missing other options here?
Wittgenstein, quite aware of the problems with defending fundamental theses, deals with the difficulty in two ways. To use a quasi-Tarskian language for momentary convenience, in the Tractatus he primarily moves to the meta-level to illustrate structure at the object level, and then shows how his own move to the meta-level was illicit and in fact it self a form of the very nonsense he was trying to expose, so that his own claims and terminology are put under a certain kind of erasure (‘kicking away the ladder’). His approach in the Investigations, by contrast, is to start out with a (mostly tacit) acquiescence to ordinary use, and then to either (a) cut off attempted philosophical ascent to the meta-level at its base or (b) show how flimsy and parochial the roots and/or applications of such ascents are when they are presented as serious theories of and/or reforms of our everyday forms of understanding.
Philosophers sometimes feel when they read the Investigations that Wittgenstein is somehow ‘playing coy’ or ‘doesn’t have a real argument for his view,’ but it is in some ways more plausible to view Wittgenstein’s rhetorical performance in the later work as manifesting a certain kind of integrity, or honor. That is, since there can be no argument of the standard type for a fundamental thesis, he will not waste your time trying to offer you one; rather he will show you how competing fundamental theses fail, and then will put his view into practice and show you what it leads to.
(C) “So are Early Wittgenstein and Later Wittgenstein saying the same thing, or different things?” The overall views in the early and later work are starkly different. There are principles in common and formal points of contact on certain issues, but the overall conception of language and thought they present are not the same. If anyone ever inferred from the fact that principles of the form (1) – (3) are accepted early and late, or that the early and late work are both critical of something they call ‘philosophy’, etc., that there was a continuity of overall viewpoint, they were incorrect to do so. (I’m not sure anyone ever did so infer, but even so.)
I want to say: one should not, in general, infer from the fact that two philosophical systems accept the same principle that they are in anything more than superficial formal agreement with one another concerning that very principle. Even if the systems are constructed by the same philosopher.
(I am thinking here not only of Early and Later Wittgenstein, but also of Later Wittgenstein, Heidegger, and Dewey, specifically in relation to 4I and 4T above. It is true that something like the contrast between these principles figures importantly in all of their work; that that contrast is important to how they set their own work off from that of other thinkers of their time; and I would even agree that they are all reacting to something like the same “Crisis of Western Thought” that occurred after the turn of the last century and that there is a kind of significant negative agreement between them that contributes to the motivating spirit of all of their work. Nonetheless I do not see much in common between their overall views.)
(D) In my last post I said that I had some sympathy for the idea that something like descriptive adequacy, assessing the truth of propositions and the propositional structure of reality, was at least part of the foundation of language and thought.
The argument for that side of the issue, advanced against a Wittgensteinian, might go something like this: “Look! I fully grant that there are such forms of language as asking questions, giving orders, praying, and so on. And I admit that there may be operators (think of the developments in modal logic since Wittgenstein’s time!) which cannot be represented in what one might call ‘the descriptive logic of factual assertion’. But surely, anything we do at all that we characterize as meaningful involves some sort of attitudes towards something – the world or our own thoughts and feelings or both. And these attitudes themselves can’t be characterized without at least potentially descriptively adequate beliefs about that world and/or our self. So while what the early Wittgenstein calls ‘the propositions of natural science’ may not be exhaustive of all the resources of our language, they nonetheless are an indispensible part of every use. Even to go pick up a slab when someone says ‘slab’ involves representing something in my environment as the thing I must pick up in order to satisfy the order, right?”
Expressed more generally we might say that action springs from belief and desire, where beliefs are representational of the world. If that is right, it surely seems that the best case scenario for an Investigations-style interpretation of our language would involve the inclusion of a more Tractatus-style interpretation of it as an at least quasi-autonomous functioning subsystem.
I think the later Wittgenstein has a powerful defense against the type of move I indicated in the last two paragraphs. I’ll try to sketch it in another post soon.