Wittgenstein’s grammatical observations: Collecting not experiences, but sentences

Wittgenstein’s grammatical observations: Collecting not experiences, but sentences

Gabriel Guedes Silva (UFMG)

Wittgenstein’s Philosophical Remarks (hereafter PR) is filled with arguments based on radical imaginary scenarios which admittedly are not limited to describing physically possible situations. While these arguments often resemble thought experiments, Wittgenstein explicitly rejects this term, stating: “What Mach calls a thought experiment is, of course, not an experiment at all. At bottom, it is a grammatical investigation” (PR §1; Ms 107, 284; see also PR §66).

I argue that a better understanding of the role of these arguments — similar in some respects to what Ernst Mach calls thought experiments — can be gained by considering Wittgenstein’s following way of describing philosophical activity, which seems novel when it stands in contrast to the conception in the Tractatus Logico-Philosophicus (hereafter TLP). 

On the same day Wittgenstein criticizes Mach’s use of the term Gedankenexperiment (6 February 1930), he also writes: “In philosophy, we continually collect a body of sentences without worrying about their truth or falsity” (Ms 107, 286; my translation). Wittgenstein seems to value this description, as he further develops it in his discussion of pain ownership (cf. Ms 107, 285-6), and as it recurs in Philosophical Remarks (cf. PR §60) and the Big Typescript (cf. BT, §550).

My aim is to propose two working hypotheses to shed light on Wittgenstein’s description of philosophy as a collection of meaningful propositions. The first, and more straightforward, is that Wittgenstein is responding to Mach’s Knowledge and Error. Mach opens Chapter XI, “On Thought Experiments,” by claiming that “Man collects experiences by observing changes in his surroundings.” By contrast, I argue that Wittgenstein emphasizes that his phenomenological investigations do not aim to collect experiences but rather to collect sentences and build a comprehensive grammar with them. In fact, phenomenological possibilities are not determined by any kind of experiment (see, for example, Ms 107, 1).

Beyond reacting to Mach, Wittgenstein’s description of philosophy as the collection of meaningful propositions also invites comparison with Moore’s method in A Defense of Common Sense (1925). Moore summarizes his article as an attempt to set out key respects in which his philosophical position differs from that of other philosophers. His method is to present a long list of propositions, each of which he claims to know with certainty to be true. Wittgenstein, in contrast, stresses that his philosophy is not concerned with truth: “I am not collecting true sentences but meaningful ones & therefore this observation is not a psychological one” (Ms 107, 285–6).

I contend that Wittgenstein had strong reasons to distinguish his approach from Moore’s. One such reason is that Wittgenstein now understands himself as working in theory of knowledge (cf. Some Remarks on Logical Form, p. 163; PR §57, 60), a branch of philosophy that held less prominence in the TLP (cf. TLP, 4.1121) and that he associated with Russell’s and Moore’s work (cf. TLP, 5.541).

Finally, I seek to connect Wittgenstein’s use of arguments resembling imaginary scenarios and his novel description of philosophy as a collection of meaningful sentences, to what can be seen as two major shifts between the TLP and the philosophy presented in the PR: the abandonment of the pursuit of a more perspicuous notation and the acceptance that there is more than one fundamental mode of projection. These changes motivate Wittgenstein to approach the question of sense in a more piecemeal and grammatically situated way.