Ordinary language is all right.
One could divide humanity into two classes:
those who master a metaphor, and those who hold by a formula.
Those with a bent for both are too few, they do not comprise a class.
'When a person has suffered very much and been very exhausted by his own sensibility, he sees that one must live day by day, forget very much, and finally clear away as much life as continues to arrive.'
'We're Russians', the woman says to the barista. 'Oh', the barista says, interestedly, 'have you been here a—' and you can nearly hear the expression on the Russian's face, flat, factual, as she finishes the barista's sentence: 'Long time'. —No point to having an attitude one way or the other.
What more would recognizing the role of 'believing in' entail? It's instructive to consult ordinary language again. Recall some of the contrasts from the case of God: believing in him, believing that he exists, proof of his existence as against lived belief. In the case of the world, hardly anyone asks whether we believe in it (though our acquaintance with reality is often questioned), and it takes something like a philosopher's constructions—'belief in the existence of the world', or better, 'belief in the existence of the external world'—to make disbelief sound abstruse rather than mad. The analogous contrast to philosophical proof of existence as against lived belief in it is not quite that the one misses the other, which is readily lacking (as in the case of God), but that it's hard to believe in the need for the one because the other seems to take care of itself, in the ordinary course of living. And if we would only with some oddity talk of not believing in the world, it is a sad fact that we not only sometimes do not believe in life, but that when this is so, the words to say so fail us. In the shadow of this fact, disbelief is not yet mad, but sick, and the long distances between needless proof, a matter-of-fact existence in the world, and a troubled one are cause for everyone's despair. Since one tradition of complaint against the philosopher's proof of existence (that practice dispels a doubt at best theoretical, that the vita activa has no use for this particular deliverance of the vita contemplativa) contends basically that philosophy does not believe (without proof) that the world exists because it does not believe in (practical) life, this suggests a convergence: the counterpart to 'believing in God' in the case of the world's existence is belief in life, itself just as much (as the melancholic's position shows) susceptible to disappearance, contraction, and expansion as a lived belief in God is, and just as liable to be distant from (or its absence concealed by) explicit affirmations of belief in the world's existence.
The philosopher and the melancholic are both as it were set apart from others. The philosopher steps aside, the melancholic falls away from them. But both show how not believing in life means not sharing in it, and this suggests a way in which belief that the world exists can be conceived as a sort of dogma of ordinary life, for the philosopher to be questioned and hopefully replaced by a sort of orthodoxy, and for the melancholic, simply not believed in, not in his life. 'Belief in', just as much as 'belief that', can be 'what we believe (in)'—just as the religious case already shows—and to lack it or not to affirm it would seem accordingly to have implications for one's community with others.
But world and life are not one. Perceiving the affinities between this pattern and the more clear-cut religious case does not itself show the way to make the case for lived 'belief in' the existence of the world as against philosophical proofs of existence that one may affirm but can't believe in, because, as it were, the terms are too far apart. We talk about believing in life, not believing in life; we talk about knowledge of the world's existence, and about proof for it. But it's not obvious how to infuse those proofs with our belief in life, not obvious how or that they don't believe in life (that's just life's complaint against philosophy, and philosophy has its answer: when you stop and think about this, it says, you'll see). Ordinary language seems to permit this discrepancy, and does not offer straightforward guidance as to how to remove it.
(from an old draft to an old draft)
What does the interlocutor of §28 say the person to whom one gives a definition doesn't, might not, know? He 'weiß ja dann nicht, was man mit »zwei« benennen will; er wird annehmen, daß du diese Gruppe von Nüssen »zwei« nennst!'. Literally, the stress on 'what' just refers to various possibilities, a name, a color, a number. It would be clearer to say, the name of the nuts, the color of the nuts, the number of the nuts, in which form the question would be, which of those 'things' the one defining the word meant. But by stressing 'what' the interlocutor emphasizes the essence of the error he has in mind. Not to know 'what one wants to call "two"' is as it were to have no clue which of the things that could be meant, one means at all. To assume one means this, this particular pair of objects right here, being pointed at, is to see nothing else that could be meant about them than their name, nothing else than that the sound "two" is a name. 'Man meint, das Lernen der Sprache bestehe darin, daß man Gegenstände benennt' (§26). The interlocutor is hung up not on the ambiguity of ostension but on a literalization of it. When Wittgenstein goes on, perhaps counterintuitively, to confirm the ambiguity of ostensive definitions ('die hinweisende Definition kann in jedem Fall so und anders gedeutet werden') or even to assert its greater extent than that proposed by the interlocutor, he does it by asserting a possibly less-literal sense for what the interlocutor supposes to be the most-literal acts of ostension, those involving names (of nuts or people). Thus where the interlocutor appears drawn to a picture of ostensive definition in which one 'form' of pointing, and its associated type of word, is primary, with the potential for misunderstanding arising from the other forms' 'literal' resemblance to it, Wittgenstein maintains that no one of those forms need be primary. Still, they all 'look the same'—so is that the basis for the ineluctable possibility of ambiguity?
In §27, Wittgenstein introduces a 'language-game in its own right', asking for and explaining names ostensively (in connection with language-games  and ). His remarks there suggest a focus on proper names and common nouns (block, slab, etc.), but §28 naturally makes the transition to other types of word—without, however, marking the likelihood that a 'name' might be asked for, and given, not just in the context of the ostensive naming game alone, but in connection with some other game, some other use of words which may thus require words of its own to be played.
'Wie hast du es gemacht?' (§33): compare to 'Wie weiß er aber, wo und wie er das Wort ›rot‹ nachschlagen soll und was er mit dem Wort ›fünf‹ anzufangen hat?' (§1). 'How do you know?', 'how does he know?'.