Here's one I like from the Blue Book:
"I shall in the future again and again draw your attention to what I call call language games. These are ways of using signs simpler than those in which we use the signs of our highly complicated everyday language. which a child begins to make use of words. The study of language games is the study of primitive forms of language or primitive languages. If we want to study the problems of truth and falsehood, of the agreement and disaggregation of propositions with reality, of the nature of assertion, assumption, and question, we shall with great advantage look at primitive forms of language in which these forms of thinking appear without the confusing background of highly complicated processes of thought. " (Wittgenstein, The Blue Book, 17)
This description makes it seem that Wittgenstein is proposing a research program into various aspects of language. In fact, I must note the affinity between this way of talking about language, that is, with recourse to "forms of thinking" (from this view, language must answer to thought – language is the vehicle of expression for thought), and contemporary discipline on linguistic processes present within the cognitive and linguistic sciences. Notice also that how the use of the term "language games" is coordinated with the notion of simpler, more basic and / or elementary forms of language. Later on, Wittgenstein rejects the idea that any language game is really more or less basic than other language games.
Lastly, I was surprised at Wittgenstein's attitude towards the kinds of standard protocols characterized by a Cartesian or Lockean epistemology. Was Wittgenstein still in the clutch of traditional western epistemology here? Does he intend, or did he intend, to really use 'language games' – here as more primitive forms of language – as a way to answer questions about the nature of truth and falsehood? Would we say that he's exercising a sort of representational theory of meaning, and then a correspondence theory of truth? I'm not so sure, but it certainly bought to make you aware of something: even in this "later" period, Wittgenstein may not have totally grasped the implications of seeing language and meaning as an activity.
Now let's look at a description of language games present in the later work, The Brown Book:
"Systems of communication as for instance 1, 2, 3, 4, 5), we shall call" language games ". altogether consistent with most contemporary theories of cognitive development wherebe infants learn the meaning of words via being able to represent what the word (s) denote] and here they even have the entertaining character of games. which we describe as incomplete parts of a language, but as languages complete in themselves, as complete systems of human communication. To keep this point of view in mind, it very often is useful to imagine such a simple language to be the entire system of communication of a tribe in a primitive state of society. "Think of primitive arithmetics of such tribes." (Wittgenstein, Brown Book, 81)
Well, clearly this description lends itself to a very different kind of interpretation. Here, language games are not conceivable as more or less simple versions of 'real' "everyday" language. On the contrary, language games are considered as complete systems of language. Wittgenstein needs to be careful here, though. It is quite easy to take "language games are complete systems of language" as meaning that a particular language game is complete in the sense of operating on quite independent syntactic and semantic rules such that the 'meaning' of an expression within that particular game need not be interpreted further. Wittgenstein should not be read here as saying that, in all cases for p within language game S, p need not be further interpreted by speakers of language game S. The possibility to stay, even with a "complete system of language" for any p to be 'given a new sense' / interpretation.
To sum up: the completeness of system of language does not guarantee that any expression within that system is immune to interpretation.
But to go back to my main point, here clearly Wittgenstein's intentions are quite different as compared to his somewhat epistemic intentions implied by his description of language games in the Blue Books. If you're interested in Wittgenstein's development from the Blue to the Brown Books, consult Rush Rhees' forward to Ludwig Wittgenstein's The Blue and Brown Books of which Rush is translator. Rhees takes a different approach to this kind of discussion, but it is quite enlightening and highly recommended none the less.
If you're interested in Wittgenstein-related products online, check these out.