Wittgenstein’s use of the word ‘phenomenology’ to describe his own work in Philosophical Remarks and The Big Typescript has occasioned much puzzlement and confusion. This paper seeks to shed light on what Wittgenstein meant by the word through a close analysis of key passages in those two works. I argue against both the view of Nicholas Gier that Wittgenstein held ‘grammatical’ phenomenological remarks to be synthetic a priori and that expressed by Moritz Schlick that Wittgenstein held grammar to be tautological. (...) I also take issue with the claim by Merrill and Jaakko Hintikka that Tractatus Logico-Philosophicus ought to be regarded as a work of phenomenology. Wittgenstein’s adoption of a variant of Husserlian phenomenology, I argue, lasted no more than a few months, from February to October 1929, during which he conceived of phenomenology as something that sought to establish what one might call ‘analytic a posteriori’ truths through something like the Husserlian method of ‘bracketing’. By the end of 1929, I claim, he had abandoned that in favour of the view expressed in the slogan ‘phenomenology is grammar’, where ‘grammar’ is not confined to the tautological, but is rather intimately linked to Wittgenstein’s notion of ‘the understanding that consists in seeing connections’. (shrink)