Is the familiarity of recognition a matter of seeing something as something? If A shows B a stick that he then pulls apart to reveal as a cap and pencil, B might say, "Oh, this is a pencil," recognizing the object as a pencil. If A simply shows B a pencil and says, "what's this?" and B replies, "oh, this is a pencil," we are tempted to say that B recognized the object as a pencil just as he did in the first case. But in the second case there was no moment of recognizing the object as a pencil, because there was never any question of what it was.
When we use the same word in two different expressions—e.g. "looking for a word in my memory" versus "looking for my friend in the park"—we are tempted to say that there must be some similarity between the two situations described. Whether we can identify the similarity or not, we think it must exist, of the same word would not apply to both situations. This is like saying that someone must understand an order before carrying it out. This is not the case, for sometimes people carry out orders without any distinctive act of understanding.
Even the word "similar" is not always used in the same way. For instance, the similarity between a pale blue and a dark blue is different from the similarity between a mental strain and a physical strain. The similarity we identify between two things often depends on the context.
Suppose we ask someone to arrange the vowels in order from lightest to darkest, and he writes "i, e, a, o, u." It does not follow that the person saw some similarity between each letter and a color, or even had any colors in mind when he arranged the vowels. There is not one paradigmatic use of "darker" to which we must compare all unorthodox uses of "darker." We can talk about a deep sound, a deep sadness, or a deep well without comparing these different uses with one another or with some ultimate definition of "deep."
Suppose we teach someone the rule "Add one," by building the series 1, 2, 3, 4, etc., all the way to 85. Suppose we urge him to go on, and after he reaches 100 he writes 102, 104, 106, etc., claiming this is what he thought we meant by the rule "Add 1." If we reply, "I meant for you to write 101 after 100," we cannot mean that "101 follows 100" was somehow in our minds when we laid down the rule "Add 1." Nor is an act of insight or intuition necessary at every stage in order to follow a rule correctly. Wittgenstein explains, "We need have no reason to follow the rule as we do. The chain of reasons has an end" (Wittgenstein, 143). There is not necessarily an insight or mental act that should tell us to follow 100 with 101. This is just what we would normally do.
Wittgenstein's discussion of "darker" has an interesting reflection on the nature of metaphor. We say a usage is metaphorical if it suggests something other than the literal usage of the word. In these passages, Wittgenstein challenges the idea that we even have a clear sense of what the literal usage is. When we talk about a "deep sound," a "deep well," or a "deep sorrow," which of these uses is literal and which is metaphorical? Historically, we could point out that the "deep" of "deep well" is older than the other uses of "deep," but we don't think of a deep well when we refer to deep sounds or deep sorrows, nor are we even aware of using words figuratively.